649

Guidelines for life extension of existing HVDC systems

Working Group B4.54

February 2016 GUIDELINES FOR LIFE EXTENSION OF EXISTING HVDC SYSTEMS

WG B4.54

Members L.D. Recksiedler, Convenor (CA), Rajesh Suri, Secretary (IN),, Leena Abdul‐Latif (FR), Les Brand (AU), Phil Devine (UK), Malcolm Eccles (AU), Abhay Kumar (SE), Mikael O Persson (SE), Maurice Smith (SE), Stefan Frendrup Sörensen (DK), Marcio Szechtman (BR), Takehisa Sakai (JP), Rick Valiquette (CA), Andrew Williamson (ZA) Corresponding Members Hans Björklund (SE), John Chan (US), Richard Michaud (US), Predrag Milosevic (NZ), Van Nhi Nguyen (CA), Randy Wachal (CA)

Copyright © 2016

“Ownership of a CIGRE publication, whether in paper form or on electronic support only infers right of use for personal purposes. Unless explicitly agreed by CIGRE in writing, total or partial reproduction of the publication and/or transfer to a third party is prohibited other than for personal use by CIGRE Individual Members or for use within CIGRE Collective Member organisations. Circulation on any intranet or other company network is forbidden for all persons. As an exception, CIGRE Collective Members only are allowed to reproduce the publication”.

Disclaimer notice

“CIGRE gives no warranty or assurance about the contents of this publication, nor does it accept any responsibility, as to the accuracy or exhaustiveness of the information. All implied warranties and conditions are excluded to the maximum extent permitted by law”.

ISBN : 978-2-85873-352-1 Guidelines for Life Extension of Existing HVDC Systems Guidelines for Life Extension of Existing HVDC Systems WG B4-54

Table of Contents

DEFINITIONS ...... 5 EXECUTIVE SUMMARY ...... 6 1. Chapter 1 – General Procedure for Performing a Life Assessment ...... 9 1.1. Introduction ...... 9 1.2. Preparation ...... 9 1.3. Team ...... 9 1.4. Assessment Process ...... 10 1.5. Deliverable ...... 11 1.6. Life Assessment Timetable ...... 11 2. Chapter 2 – Thyristor Based HVDC Systems Performance Issues ...... 13 2.1. Survey of availability and reliability (over all HVDC systems in the world) ...... 13 2.2. Operating History ...... 14 2.3. Major equipment/system/sub-system failure/refurbishment summary ...... 14 2.4. Alternatives and justification ...... 15 2.5. Methods for assessing reliability, availability and maintainability of existing components 16 2.6. Basis for replacement/refurbishment of equipment ...... 17 2.7. Performance after replacement and refurbishment ...... 17 3. Chapter 3 – Life Assessment and Life Extension Measures of Equipment...... 19 3.1. DC Switchyard Equipment ...... 20 3.2. Valves ...... 32 3.3. Converter Transformers ...... 42 3.4. DC Control and Protection ...... 61 3.5. References: ...... 79 3.6. Valve Cooling ...... 80 3.7. Station Auxiliary Supplies ...... 84 3.8. Ground Electrodes and Electrode Lines (does not include sea electrodes) ...... 85 3.9. Reliability Centered Maintenance ...... 89 3.10. References ...... 89 4. Chapter 4 – Guideline for assessing Techno-Economic Life of Major Equipment ...... 91 4.1. Operational Issues – Maintenance Cost / Management and Availability of Spares ...... 91 5. Chapter 5 – Reccomendation for Specification of Refurbishing HVDC System...... 94 5.1. Introduction ...... 94 5.2. Main Components of a Converter Station: Guideline for the Specification ...... 94 5.3. Interfaces ...... 101

Page 3 Guidelines for Life Extension of Existing HVDC Systems

5.4. Maintainability including spares requirement ...... 102 5.5. Cost minimization ...... 103 5.6. Replacement time minimization ...... 103 5.7. Operation outage minimization ...... 104 6. Chapter 6 – Testing of Refurbish/Replacement Equipment ...... 106 6.1. Introduction ...... 106 6.2. High Equipment ...... 106 6.3. Low Voltage Equipment ...... 110 6.4. Auxiliaries ...... 110 6.5. Fire detection ...... 111 6.6. Tests of the connection with the existing equipment ...... 112 6.7. Example of commissioning procedures: IFA 2000 refurbishment ...... 112 7. Chapter 7 – Environmental Issues ...... 115 7.1. Insulating Oil ...... 115 7.2. Polychlorinated Biphenyl ...... 116 7.3. Sulfur Hexafluoride Gas ...... 117 7.4. Halon Gas ...... 117 7.5. Refrigerants ...... 118 7.6. Abestos ...... 119 7.7. Audible Noise ...... 119 7.8. Electromagnetic Effects ...... 120 7.9. Mitigation of Environmental Issues ...... 120 Chapter 8 – Regulatory Issues ...... 122 8.1. Renovation & Modernization ...... 123 8.2. Conclusion ...... 123 9. Chapter 9 – Techno Economics ...... 124 9.1. Financial analysis of refurbishment options ...... 124 Appendix A – Basslink Case Study ...... 127

Page 4 Guidelines for Life Extension of Existing HVDC Systems

DEFINITIONS Life Extension

Life extension requires a life assessment and the final result will be a life extension of the HVDC. However as a result of the life assessment it may be decided not to extend the life and this would be outside the scope of this document

Design Lifetime

The Design Life time of a component is the time during which the device is commercially available or is commercially viable in its original supplied form.

Operational Life

The Operational Life is the period of time over which a product will operate correctly, assuming that it has been maintained in accordance with its maintenance instructions.

Independent Expert

An Independent Expert is a person or person having an expert knowledge of a part or many parts of a HVDC link and may be a Consultant or not.

Maintenance spares

Within the system there will be components that are expected to wear out or have a limited lifetime, either in terms of operational (or storage) time or usage. These components, known as maintenance spares, need to be replaced at predictable and specified intervals.

Strategic spares

All components (including components that are maintenances spares) exhibit random failures that have to be treated in a statistical manner; every year it is expected that a small proportion of components to fail but cannot predict which component will fail or when it will need to be replaced. These are known as strategic spares.

One vendor spares

Some spares such as capacitors are available from many different suppliers and thus large quantities do not have to be stocked. Other spares are available only from the Original Equipment Manufacturer OEM and thus may or may not be available in about 15 years. For example, a HVDC control system has a life of approximately 15 years before a new generation is developed. If it is purchased in year 14, the spares may not be available for very long. Even if the OEM guarantees the availability, many of the parts from sub-suppliers may not be available any more. Thus the number of spares may influence the usable life of the equipment.

Resin Impregnated Paper (RIP)

The condenser paper in a bushing is impregnated with an epoxy resin compound.

Page 5

Guidelines for Life Extension of Existing HVDC Systems

EXECUTIVE SUMMARY In today's complex environment, energy players face growing demands to improve energy efficiency while reducing costs. Energy shortages and increased ecological awareness have resulted in great expectations for grid stability and reliability. Utilities and industries need to find eco-efficient solutions to maintain secure, safe and uninterrupted operations. A number of regulatory changes in the market have led to increased efforts by utilities and grid operators for optimized utilization of their existing networks with respect to technical and economic aspects. As the electric power transmission system ages, the topics of life assessment and life extension have become predominant concerns. At the same time, cost pressures have increased the desire to minimize maintenance. The goals of minimum maintenance and extended life are often diametrically opposed.

The concept of simple replacement of power equipment in the system, considering it as weak or a potential source of trouble, is no longer valid in the present scenario of financial constraints. Today the paradigm has changed and efforts are being directed to explore new approaches and techniques of monitoring, diagnosis, life assessment and condition evaluation, and possibility of extending the life of existing assets.

A major challenge for grid operators worldwide is to assure sufficient power with quality and reliability. In this regard High Voltage (HVDC) systems play a major role in bulk power transmission, system stability, integrating remote renewables and ride through of disturbances. Therefore HVDC systems represent an indispensable part of the electricity grid in the countries where they are installed.

HVDC has been in commercial use since 1954, and most of the systems are still in operation. However, the early mercury arc valve systems have been phased out and replaced by Thyristor Valves. This has extended the life of many of the early systems, but the Thyristor based systems are also approaching an age where the Thyristor Valves may require replacement or refurbishment. Operation and maintenance issues of these aging systems have become a challenge. The situation is further complicated by the fact that all of the HVDC systems are custom built by a relatively small number of Original Equipment Manufacturers (OEM).The HVDC manufacturers have supplied several different generations of equipment and these differences have to be considered in any life extension assessment.

Disclaimer: This Technical Brochure is about life extension of HVDC Converter Stations only. Upgrading the converter stations or operating them beyond its design specifications is out of the scope of this document. However for both of these it is highly recommended that the OEM is consulted as these are complex and a custom built installation and the normal design rules will likely not apply.

With the aging of the equipment, measures to extend the life of the equipment have to be considered by utilities and grid operators. Renovation, modernization and life extension of HVDC stations is usually one of the most cost effective options for maintaining continuity and reliability of the power supply to the consumers. These life extension measures have to be implemented with minimum impact on the HVDC system and the associated networks whilst maintaining an acceptable level of reliability and availability. If life extension is not economical, the systems may be disposed of in an environmentally acceptable way. Also, environmental issues need to be considered prior to a life extension project to avoid any inadvertent environmental damage.

The cost of outages to do a refurbishment must be considered as part of the overall cost. This may then dictate a Greenfield option where a new converter station can be built and only short switch overtime is required. An example of this is the Oklaunion Converter Station (CS) in the USA, where the outage costs tipped the scale towards a Greenfield versus a Brownfield option for refurbishment. The definition of the

Page 6

Guidelines for Life Extension of Existing HVDC Systems interfaces in the case of a Brownfield project is critical and more complicated than in a Greenfield project.

Most utilities are interested in better understanding and projecting service life of HVDC equipment to help manage risk; however, generic reliability data is inadequate for current decision support needs. It is important to establish industry-wide equipment performance databases to establish a broad-based repository of equipment performance data. With proper care and analysis, this data can provide information about the past performance of equipment groups and subgroups, and the factors that influence that performance. With enough data, projections can be made about future performance. Both past and future performance information can be useful for operations, maintenance, and asset management decisions.

However, for some components it is more difficult than most to determine the useful life and the actual end of life failure modes. The Thyristors themselves are an example, as they have been around for some 35 or more years and yet are showing little sign reaching end of life, except where some design or quality issues have been uncovered.

The life-extension may involve any of the following actions:

• Refurbishing the systems or subsystems • Selectively replacing aging components • Combination of the above In some cases life extension is not economically feasible and a Greenfield replacement may have to be considered.

The following steps need to be taken to arrive at a decision:

• Review the past performance of the major HVDC equipment and systems • Identify the future performance issues associated with the ageing of special components of the HVDC systems. There may be equipment that has not shown performance issues in the past but still may need an extension and should also be considered. • Determine economic life of various components in the converter station and for making replacement versus life extension decisions. The consideration of economic life will include capital cost, reliability and availability, cost of maintenance and the cost of outages and power losses. • The usable life of a refurbishment is likely in the average of 15 to 20 year range whereas a Greenfield option is likely 35 to 40 years and this needs to be factored into the evaluation but it is recognized that some components may have a different year range. One way of going about this activity could be to develop criteria, weightings and methodology for determining near-term action and forecasting the technical and financial effect due to system ageing. This should follow an approach based on condition replacement cost and importance of the equipment and components. Assessment of condition parameters could be in terms of equipment age, technology, service experience (e. g. after sales service quality, maintenance costs) and future performance, individual failure rates, and so on. A viable duration for the life extension should be determined and usually 15 to 20 years is achievable. Longer durations may be more difficult to assess with any degree of accuracy.

Evaluation of the possibility of extending the service life of electrical equipment is a techno-economic compromise which must lead to “run-refurbish-replace” decisions. Once the expected service life period has expired, refurbishment of such equipment falls within the life extension program. The investment at

Page 7

Guidelines for Life Extension of Existing HVDC Systems the initial stage is very capital intensive to the utility concerned, as the devices to be installed in the system for Residual Life Assessment (RLA) and condition evaluation purpose, are very costly. However, the decision to refurbish or to replace should be based on the study of comparable costs and benefits over the same potential life time of the asset.

Therefore, it can be concluded that the need for life extension and replacement of equipment in HVDC system arises due to:

• Arresting the deterioration in performance • Improving the availability, reliability, maintainability, efficiency and safety of the equipment • Regaining lost capacity • Extending the useful life beyond originally designed life of 35 to 40 years • Saving investment on new equipment • Not having availability of new spares due to obsolescence These CIGRE Working Group objectives help utilities as follows:

• design refurbishment strategies for their existing HVDC systems to extend equipment life, • evaluate O&M and reliability performance improvement strategies for their existing HVDC systems, • provide a guideline for determining economic life of various components in the converter station and for making replacement versus life extension decisions. The consideration of economic life should capital cost, reliability and availability, cost of maintenance and the cost of power losses. This technical brochure (TB) provides guidelines for the general procedure for performing life assessment (chapter 1). Following this, a more detailed description of performance issues of the thyristor based HVDC systems (chapter 2) is given and the life assessment measures of equipment (chapter 3) and guidelines for accessing the techno-economic life of equipment (chapter 4). Chapter 5 deals with the recommendation for specification of refurbishing HVDC system and chapter 6 follows with the testing of the refurbished and replaced equipment. Lastly, this brochure will outline environmental issues (chapter 7) and regulatory issues (chapter 8) involved in the life assessment and finalize with a financial analysis of the refurbishment options (chapter 9).

Page 8

Guidelines for Life Extension of Existing HVDC Systems

1. Chapter 1 – General Procedure for Performing a Life Assessment

1.1. Introduction Each assessment will be slightly different depending on the needs and expectations of the Owner as well as the information available. In many cases, only high level information will be available, in some cases detailed information will be available and in others only high level information is required by the Owner. In most cases, only a technical evaluation is required but in others a commercial justification may also be required. This guide is only intended to describe the technical procedure.

Only the HVDC equipment will be considered here as the AC equipment is outside the scope of this procedure and there are several sources of existing literature already covering this topic. This excludes AC and DC filters and Synchronous Condenser and SVC’s or Statcom’s

1.2. Preparation In preparation for performing the life assessment, it is critical to define the needs and expectations of the Owner as much as possible. The next step is to produce a written proposal detailing the scope of work, information available and to be provided by the Owner, deliverables and estimated cost. The proposal needs to be discussed with the Owner and modified as required and agreed upon. It is also critical to determine the amount of time that a life assessment should be good for. The average is 15 to 20 years whereas a replacement would have a life of approximately 35 to 40 years. A life assessment or refurbishment is normally referred to as a Brownfield where as a replacement is called a Greenfield.

1.3. Team A team needs to be established which may consist of the Owners’ staff, Suppliers’ staff and Independent Experts, or any combination thereof. The team may be different if only a piece of equipment is being evaluated or if the entire HVDC Stations is being considered. Formal communications should be through a leader on both the Independent Experts’ and Owners’ sides only. Formal communications is anything that affects price, schedule or quality. Informal communications are still encouraged.

In general, a specialist or expert should be employed for each of the areas but some areas are critical because of the importance, complexity, or cost:

• Converter Transformers and smoothing reactors– Critical because they are normally the highest cost item in a life assessment endeavor and are very complex. • Thyristor Valves – Critical as they are the next highest cost and a very specialized area. • HVDC Control and Protection – Critical as they are very complex even though not that costly • Cooling Systems - HVAC – not as critical but still needs proper evaluation • Auxiliary Systems – Not critical, it is usually a source of many operating problems • Other DC equipment – includes DCCT, VDR, disconnects and breakers and switches • Civil – Aging of structures’ and foundation’s – usually overlooked in assessments but should not be • Miscellaneous - Wiring, Fire protection, security buildings and ground grid. These items are usually overlooked in an assessment but should not be.

Page 9

Guidelines for Life Extension of Existing HVDC Systems

1.4. Assessment Process A kick off meeting with the team is recommended to obtain the information and discuss the format that it will be provided in, as well as any software required to open the flies and interpret the results. This should be provided in a standard format such as word, excel or pdf wherever possible.

Several visits to the station(s) are a must for discussions with the maintenance staff. Also discussion should take place with the operating staff to determine if the equipment is meeting their expectations. Also determine if there are any additional requirements that the existing equipment cannot provide and possible benefits.

The following should be part of the analysis:

• Operating problems or changes in the mode of operation • Maintenance records for the last 5 years • Any modifications performed and why • Any failures and failure reports • Original quality or design issues • Any equipment replaced and when • Spare or replacements parts or obsolescence for major or critical equipment. An example may be Thyristor failures per year, spares on hand and whether they are still available from a supplier. • Status of spares – questions to consider are whether they are usable? Have they been maintained? Have they been in service and removed because they were gassing such as the converter transformer? Have they never been in service? • Technical skills of staff to continue operating and maintaining the equipment. Is additional training required? • Normal life of each piece of equipment • Drill down to the smallest subsystem or components possible as it may be possible to replace only some components and not the larger equipment subsystem or system. This could save a lot of cost but requires that the detailed information be available. • Criticality of the HVDC link to the system and consequences if it is unreliable or out of service. • Risk assessment – is there a way to prevent or mitigate a risk. What would you do if the risk happened? • Some equipment may not have a history of problem and failures but consideration should be given that some of this will occur if the life assessment is long enough and should be considered as a contingency. • Replacement costs, wherever possible, should be obtained from a supplier. Where this is not possible estimated costs based on previous experience is required. • An implementation schedule will also likely be required. Wherever possible the schedule should be obtained from a supplier. Where this is not possible, a rough schedule based on previous experience is required. As equipment is costly it may be beneficial to use an Independent Expert to review an assessment provided by a Supplier.

Regular status update meetings between the Owner and the involved parties should be held to ensure everyone is on track and to disseminate any new information, if available.

Page 10

Guidelines for Life Extension of Existing HVDC Systems

1.5. Deliverable The deliverable is usually a report with recommendations and conclusions. A preliminary copy of the report should be reviewed with the Owner incorporating any comments or concerns. In some cases the Owner will request additional work in a specific area requiring additional analysis, modifying the report and another preliminary meeting.

Eventually a final report will be issued. The Owner will then combine this with the commercial analysis and make a final recommendation for Brownfield or Greenfield.

In the case that the Owner also wants a commercial analysis, it is necessary to understand these requirements and follow the standard way or format that the company produces the information and if the project is to be financed by a bank what is necessary for the bank to do its risk assessments.

1.6. Life Assessment Timetable

Note: Life Assessment should begin 5 years before the time indicated in the table below or if there are high failure rates or maintenance issues.

The reality is that there is no piece of equipment where a firm number is accurate. The idea of coming up with a number is that if this piece of equipment has not caused any major problems up to the point in time, a life assessment should be done to determine the remaining life, refurbishment and replacement of that piece of equipment, subsystem or system, or if replacement is done, then the entire HVDC project may be the best option. Manitoba Hydro has had analogue controls in-service for over 42 years and it could be 45 or even 50 years before they are replaced for a HVDC station with a design life of 35 years. But it has been assessed many times over the last 15 or so years. This is certainly not the average. Manitoba Hydro also has some air core AC and DC filter reactors in-service for 37 years with no failures; another converter station has had many AC filter air core reactor failures in a much shorter lifetime. Only major items are considered here and not subcomponents.

HVDC Equipment Lifetimes Note: Excludes design and production run quality issues HVDC Station Design life 35 to 40 years

Equipment Lifetime (Years) Comments

Converter Transformer 40 AC Bushings 25-30 25 OIP , 30 RIP DC Bushings 30-35 30 OIP, 35 RIP - DC bushings have more insulation Tapchanger 30 or 350 000 Operations, Seals and springs Coolers 25

Thyristor Valves 35 Thyristors 35 Valve Reactors 30 Tubing 25 Fiber Optics 35 Damping Capacitor 30 Damping Resistor 30

Page 11

Guidelines for Life Extension of Existing HVDC Systems

Electronic cards 25 - 30 Coolers 25

HVDC Controls ( analogue) 35 HVDC Controls ( digital ) 12 - 15 HMI 7 DC Smoothing Reactor (oil) 35 Bushing may have a shorter life DC Smoothing Reactor (air core) 35

Optical DCCT 30 Electronics may have a shorter life DC Voltage Divider 30

DC Surge Arrestors 35 DC Insulators 35

DC Wall Bushings 30-35 30 OIP, 35 RIP - DC bushings have more insulation

DC Switching Equipment 35

DC Buswork, structures 50 Ground Electrode 40 - 50 Normal Design Life Civil Work 50

Communications Systems 15

Page 12

Guidelines for Life Extension of Existing HVDC Systems

2. Chapter 2 – Thyristor Based HVDC Systems Performance Issues HVDC technology has experienced enormous growth worldwide in the past 50 or more years, and many HVDC systems are currently in operation. As these systems age, asset management questions of what to refurbish, what to replace and the scope of equipment repair and replacement is becoming increasingly important to extend the life of all HVDC links.

A HVDC System will likely have a Computerized Maintenance Management System (CMMS) which will contain more detailed records than that which is necessary for reporting to CIGRE or internally to high level management within an organization. Modern day CMMS also have asset management systems integrated into them to assist with Life Assessment decisions.

The following is some information about the collection of data from HVDC Schemes throughout the world from those HVDC schemes that have submitted the information. This information can be useful for benchmarking or for discussions with staff from HVDC schemes that are similar.

2.1. Survey of availability and reliability (over all HVDC systems in the world) CIGRE B4 collects operational performance and reliability data of all HVDC systems in commercial service in the form of annual reports. Such reports are prepared in accordance with a standardized protocol published in accordance with CIGRE publication 346 Protocol for reporting the operational performance of HVDC transmission systems: (note: please refer to the current version of the protocol as details may have changed since this publications).

Performance data includes reliability, availability and maintenance statistics.

Reliability data is confined to failures or events which result in loss or reduction of transfer capability. Statistics are categorized in order to indicate which type of equipment caused the reduction in transmission capacity.

Advisory Group B4.04 of CIGRE Study Committee B4 (HVDC and Power Electronics) summarizes the performance statistics for all reporting schemes every two years in a CIGRÉ paper entitled “A Survey of the Reliability of HVDC Systems Throughout the World.”

High levels of equipment reliability for individual HVDC links are also visible from the CIGRE performance reports. The CIGRÉ performance report also contains sections with the yearly number of forced outages and duration of forced outages (in equivalent outage hours) for each equipment category.

Convertor station equipment is classified into six major categories: AC and Auxiliary Equipment (AC-E), Valves (V), Control and Protection (C&P), DC Equipment (DC-E), Other (O) and Transmission Line or Cable (TL).

2.1.1 A.C. and Auxiliary Equipment AC-E This major category covers all ac main circuit equipment at the station (from the incoming ac connection to the external connecting clamp on the valve winding bushing of the convertor transformer). This category also covers low voltage auxiliary power, auxiliary valve cooling equipment and ac control and protection. It is subdivided into following subcategories:

• A.C. Filter and Shunt Bank AC-E.F

Types of components included in this subcategory would be capacitors, reactors, and resistors which are included in the ac filtering or shunt compensation of the converter station.

• A.C. Control and Protection AC-E.CP

Page 13

Guidelines for Life Extension of Existing HVDC Systems

Assigned to this subcategory are ac protection, ac controls, and ac current and voltage transformers. AC protection or controls could be for the main circuit equipment, for the auxiliary power equipment or for the valve cooling equipment.

• Converter Transformer AC-E.TX

The converter transformers and any equipment integral to the converter transformer such as tap changers, bushings or transformer cooling equipment is assigned to this subcategory.

• Synchronous Condenser (Compensator) AC-E.SC

The synchronous condenser (compensator) and anything integral or directly related to the synchronous machine such as its cooling system or exciter is included in this subcategory.

• Auxiliary Equipment & Auxiliary Power AC-E.AX

This subcategory includes auxiliary transformers, pumps, battery chargers, heat exchangers, cooling system process instrumentation, low voltage switchgear, motor control centers, fire protection and civil works.

• Other A.C. Switchyard Equipment AC-E.SW

This subcategory includes ac circuit breakers, disconnect switches, isolating switches or grounding switches.

The classification of other groups can be found in CIGRE publication 346 protocol for reporting the operational performance of HVDC transmission systems.

2.2. Operating History Records of equipment operating history are an integral part of equipment maintenance, and are required to verify the suitability of maintenance practices, in order to achieve the design life of the HVDC equipment.

Equipment operating history should contain basic equipment technical data, maintenance intervals, detailed records of maintenance activities completed during the scheduled outages, components replaced to keep the equipment operational, forced outages caused by the equipment, results of scheduled inspections and diagnostic tests and results of tests done after replacement of components.

The data collected is used to assemble equipment condition assessment reports which assist in identifying any requirements to increase and change maintenance, repair or replace equipment.

2.3. Major equipment/system/sub-system failure/refurbishment summary While Mercury Arc Valve (MAV) based HVDC links were designed to last at least 35 years with major overhauls, very few achieved this figure but Thyristor based HVDC links are designed to last at least 35- 40 years.

Not all converter stations will have a lifetime of 35 to 40 years. Therefore if running the equipment to failure is not an option due to expensive consequential equipment damage and long unplanned outages, midlife replacement and refurbishment of HVDC equipment is a valid option.

Page 14

Guidelines for Life Extension of Existing HVDC Systems

The major HVDC equipment requiring life assessment activities are: converter transformers, valves, valve electronics, controls and protection, valve cooling, AC filters, DC filters, smoothing reactors and circuit breakers.

For HVDC links constructed in the 1970’s (early Thyristor links) the following life assessment activities were undertaken:

• Valve controls and valve electronics upgrade: between 15 and 23 years in service after commissioning. (some analogue systems are still in-service) • Control and protection system upgrade: between 26 and 30 years in service after commissioning. • Replacement of MAVs with Thyristor Valves was normally done after 20-35 years of MAVs in service. The replacement of Thyristors was completed after 21 to 30 years of Thyristors in service. • Cooling systems have been upgraded or replaced after some 25 years of service • Replacement of oil filled Smoothing Reactors (with air core or oil) and oil filed Direct Current Current Transducers (DCCT’s) after some 35 or more years of service. • Refurbishment of converter transformers after about 30 years of service. For the HVDC links constructed in the 1980’s, the following life extension activities were undertaken:

• control and protection system upgrade: between 23 and 27 years in service • valve cooling system upgrade: after 24 years in service • valve upgrade (together with valve cooling) after 27 years in service.

2.4. Alternatives and justification In order not to degrade the performance of the HVDC link when some equipment is approaching its design life the following alternatives are available:

a) selective repair and refurbishment or replacement of HVDC equipment It is very important that equipment can be repaired as spare parts remain available, and the knowledge base (engineering and technical staff) is being retained.

Equipment replacement is required if spare parts are not available (OEM’s are no longer in business. Either the parts are phased out, discontinued, cannot be remade or they are reverse engineered locally), or the knowledge base is lost (maintenance personnel familiar with the equipment are retired).

Selective equipment replacement is an excellent method to achieve the design life of the HVDC link if other components that are not refurbished or replaced can last till the end of HVDC converter station extended life.

As this document is not covering any increased ratings, the rating of the converter station will remain the same.

b) complete replacement of HVDC converter stations Complete converter station replacement is required when the majority of the equipment is at the end of its design life; the HVDC link is still required for power transfer, or AC system performance improvement. This can ultimately be a combined economic or technical decision as to how extensive the complete replacement may be.

Page 15

Guidelines for Life Extension of Existing HVDC Systems

Complete replacement of HVDC converter stations is also an opportunity to increase the steady state power transfer capability, dynamic power transfer capacity of the link or where a lengthy outage to the existing converter stations is not acceptable.

In any case, actions for extending converter station life need to be addressed before HVDC link reliability and availability are impacted.

2.5. Methods for assessing reliability, availability and maintainability of existing components CIGRE’s Survey of the Reliability of HVDC Systems throughout the world enables HVDC link asset owners and operators to compare the performance of their own HVDC link against the performance of similar HVDC links in the world. It is recommended for owners seeking to access the reliability, availability and maintainability of components, to actively participate in CIGRE HVDC user groups.

The goal is that the accumulated data from several systems would establish a basis against which performance of individual HVDC links could be judged.

Performance of any HVDC system can be evaluated using data on energy availability (EA), energy utilisation (EU), forced energy unavailability (FEU), scheduled energy unavailability (SEU) and thyristor failure rates, as well as examining equipment categories causing forced outages or reduction of HVDC system capacity.

In the CIGRE Survey and statistics carried out by Advisory Group B4.04, the following definitions are used:

• Outage - The state in which the HVDC System is unavailable for operation at its maximum continuous capacity due to an event directly related to the converter station equipment or DC transmission line is referred to as an outage. • Scheduled Outage - An outage, which is either planned or which can be deferred until a suitable time, is called a scheduled outage. Scheduled outages can be planned well in advance, primarily for preventive maintenance purposes, such as annual maintenance programs. • Forced Outage - The state in which equipment is unavailable for normal operation but is not in the Scheduled Outage state is referred to as a Forced Outage. • Forced Outages can be caused by trips - sudden interruption in HVDC transmission by automatic protective action, manual emergency shutdown, or unexpected HVDC equipment problems that force immediate reduction in capacity of HVDC stations or system but do not cause or require a trip. • Energy Availability (EA) - A measure of the energy which could have been transmitted except for limitations of capacity due to outages is referred to as Energy Availability. • Energy Unavailability (EU) - A measure of the energy which could not have been transmitted due to outages is referred to as the Energy Unavailability. • Energy Utilization (U) - A factor giving a measure of the energy actually transmitted over the system. For example, comparing the performance of one HVDC link against similar pairs:

Scheduled Equipment Unavailability (SEU) has less significance than Forced Equipment Unavailability (FEU) in comparing different systems since scheduled outages may be taken during reduced system loading conditions or when some reduction in power transfer capability is acceptable. Discretionary outages for maintaining redundant equipment are also considered within the SEU category.

Page 16

Guidelines for Life Extension of Existing HVDC Systems

2.6. Basis for replacement/refurbishment of equipment HVDC converter station equipment (and subsystems) are complex and have varying design life times. Each piece of equipment, system or subsystem should be assigned a “normal” life time which, as it approaches, could trigger a life assessment

The criteria for the equipment replacement and refurbishment are related to the risks the asset owner is ready to take and potential lost revenue which is correlated to equipment performance.

For example, capacitors can be replaced after design life is exceeded. However they can also be replaced after the number of failures exceeds a percentage of installed capacitors per year (e.g. 2%). The latter option implies a number of filter bank trips or loss of redundancy (maintenance outage), which are the consequence of failed capacitor cans.

A conservative approach would be not to run the equipment beyond the manufacturers recommended design life. An assumption is that the spare parts and skilled maintenance personnel are still available.

The following conditions could require equipment replacement or refurbishment even before the design life is exceeded:

• Poor performance of equipment. An unacceptable number of HVDC trips caused by this equipment reducing HVDC availability, or long scheduled outages required to keep the equipment in a serviceable condition. • The type of equipment is not manufactured any more (for example circuit breaker) and there are no spare parts available. It is possible to postpone the replacement of the whole equipment fleet, say of circuit breakers, by replacing one or more circuit breakers, and using the parts from the units removed from service as a source of spares. In some cases the parts can be reverse engineered by the utility if it has the knowledge or by other firms such as tapchanger parts which specialize in this field • Engineering and maintenance staff retiring and the knowledge base of how to maintain some equipment is being lost and the supplier also cannot support maintenance of the equipment. • The results of equipment condition assessments showing poor or deteriorating equipment conditions (for example very low degree of polymerization paper inside converter transformers), could justify earlier replacement, even before equipment design life is exceeded. • Failures of the same type of equipment at other HVDC links, could justify unscheduled equipment condition assessment, and if required, early replacement • Manufacturer instructions to remove equipment from service due to production defect (e.g. use of unsuitable material for components during production) could result in early equipment refurbishment. • Under direction from an outside regulatory body (safety or environmental issues for example) • Technical obsolescence – older software versions are no longer supported by the OEM and the new software requires new hardware

• High Cost of Operations Maintenance and Administration. OP-EX stands for Operating Expense. COMA – Stands for Cost of operations, Maintenance and Administration

2.7. Performance after replacement and refurbishment Reliability performance data collected for CIGRE reporting purposes (data on energy availability, energy utilization, forced and scheduled outages and other data in accordance with the reporting protocol developed by the Advisory Group B4) can be used to evaluate success of equipment replacement.

Page 17

Guidelines for Life Extension of Existing HVDC Systems

Performance improvement should be visible by comparing HVDC reliability data and loss of redundancy data (number of forced outage events and the equivalent forced outage hours relevant to replaced equipment category) two years before replacement and two years after replacement.

However if the equipment was performing well and had enough spare parts (for example the control and protection system), but was replaced with the new model due to obsolescence, then good performance in the future will be the sign of successful equipment replacement. If equipment is replaced as result of condition assessments identifying poor or deteriorating equipment, prior to the equipment actually affecting performance indicators. There could be an expectation of a reduction in maintenance and that should be reflected in maintenance records and maintenance hours required.

Page 18

Guidelines for Life Extension of Existing HVDC Systems

3. Chapter 3 – Life Assessment and Life Extension Measures of Equipment Chapter 3 comprises the following sections, each addressing performance issues and technical life assessment, including:

• Remaining life (where feasible) • Refurbishment • Maintenance • Possible tests after a fault (converter transformer only) Maintenance is a critical part of achieving the ability to extend the life of the DC equipment. If proper maintenance is not performed life extension may not be possible or severely limited.

Most converter stations employ some kind of Computerized Maintenance Management System (CMMS). It can be part of a larger system, a standalone system or a home grown system. There is continuing pressure on reducing maintenance costs and outage times and some utilities have gone to Reliability Centered Maintenance (RCM) Systems and away from time based systems. RCM generally relies on doing maintenance based on levels of inspections, importance of equipment, is condition based and relies on appropriate and timely maintenance intervention. This has the effect of improving reliability and availability and reducing maintenance costs. This can be a hard sell to existing converter station staff that are used to a time based system.

During the warranty period the OEM’s maintenance requirements must be followed and well documented to prove that the maintenance has been completed. After this period condition based maintenance or RCM can be adopted.

It is very important to have good maintenance records to analyse the equipment performance with the Root Cause Analysis of any failures. Complete systems have been replaced by utilities because the root cause of the problem was not identified and thus the problem was still there after replacement. Trend analysis is also very important as a bad reading or information or statistical analysis can skew an analysis with very little data. Good records are also very important to assist in justifying any refurbishment or replacement as it can be readily shown how much improvement is possible and what the benefit could be in increased revenue.

Spares

The number of spares that are available especially for items such as the DC controls can impact the usable life of the equipment. Spares can be broken down in many ways, one way is as follows:

 RAM and performance guarantees spares – Spares required to meet the RAM, and performance guarantees requirement of the contract usually 2 years.  Initial spares- Spares for the first 10 to 15 years to get a record of the failure rates and then to order more. The supplier will usually guarantee availability for that period  Insurance spares – Spares to cater for infrequent failure ( e.g. DC wall bushing)  Consumable spares – Items to be consumed due to normal failure rates (e.g. fan, motors)  One supplier items – Some items are only available from the OEM, control cards are an example whereas AC filter capacitors are available from many vendors on relatively shorter notice, and thus stocking levels will likely be different. In most instances exact replacements should be available; however in certain instances a re-design of the bank may be required.

Page 19

Guidelines for Life Extension of Existing HVDC Systems

Some utilities are starting to specify that enough spare parts be supplied to last the 35 to 40 year life of the project. This could help ensure the viable life of 40 years of these types of devices but may be difficult to estimate the requirements.

3.1. DC Switchyard Equipment DC Switchyard equipment has both AC and DC imposed on it. Oil and paper is an effective insulator for AC voltages and the dielectric stresses in the insulation remain relatively constant over a fairly large temperature range. DC voltages must be constrained by paper and pressboard in what has been called a “DC Cocoon” by one supplier. The dielectric stresses for DC Voltages vary more with temperature and with different types of materials. Over the years this has caused problems even with established suppliers. Another issue is that DC ions and charges do not dissipate for many hours and this must be considered especially for a polarity reversal. This is a particular concern when designing oil filled smoothing reactors as well as converter transformers.

Laboratory testing has shown that shed profile and material types are important for HVDC equipment airside performance especially at higher HVDC voltages over 300 kV DC.

3.1.1 Oil Filled and Air Core Smoothing Reactors (SR) Description:

The first used Smoothing Reactors were oil filled and had relatively large inductance values to limit the fault current on the line side of the Smoothing Reactor to that which the Thyristor’s could still maintain full control and or avoid a DC Line resonance. As Thyristor capability advanced, this, in most cases, permitted the reduction in the inductance value. Advances in air core reactor technology have allowed their use as Smoothing Reactor, virtually replacing the oil filled reactor. The advantage being lower cost, simple design, low maintenance, no oil spill containment required and less risk of fire. Smoothing reactors perform a number of functions:

• Reduce the probability of valve commutation failures

• Prevent discontinuous current at low power levels

• Allow the valves to remain in full control for a fault on the line side of the SR

• Reduce front of wave DC line surge

• Reduce the DC harmonic voltages seen by the DC filters

• May be used to de-tune the DC transmission line resonances

Performance Issues:

Both oil filled and air core have faced failures and problems, but these issues are generally different.

The oil filled SR’s contain large amounts of paper insulation and at least one SR has failed due to inadequate drying of the oil after maintenance and refurbishment. DC bushings are another source of failures. Replacement DC bushings are not available in some instances if the original equipment manufacturer (OEM) is no longer in business and the replacement must be the exact same make and model. The DC barriers at the oil end of the bushing combined with the matching foils of the DC bushing capacitance core and the “special” low sodium porcelain make this problematic. This type of porcelain is no longer made by NGK and the one supplier ABB has done away from an HVDC oil end porcelain all together. Others have gone to RIP Bushings with or without SF6 gas insulation depending on the voltage. Another option is foam insulation instead of SF6 gas and the foam can contain SF6 bubbles or

Page 20 Guidelines for Life Extension of Existing HVDC Systems nitrogen bubbles. Oil spill containment and fire protection is now usually required, but because of this risk and cost, has accelerated the replacement of oil filled reactors with air core ones.

Air Core SR’s have an exterior coating of paint or RTV which protects the insulation from the sun and Ultraviolet (UV) rays. Cracks in this coating have allowed sun and moisture to get in causing failures. These coatings have to be re-applied or renewed approximately every 10 years depending on environmental, sun and pollution conditions. Some air core reactors have been lifted improperly during installation also causing failure later. When they are tested in the factory, Air core SR’s pass the noise test because there are no harmonics. However, in the field with harmonics present, they may or may not be acceptable. If they become too noisy then they are outfitted with noise barriers. Conversely, if they were not designed for this and the increased temperature causes shorter life and has resulted in failures and fires. Some air core SR’s have “Black Spots” on them but no failures have been reported to date. An addition of corona rings may eliminate the black spots.

Technical Life Assessment:

The average life of oil filled or air core Smoothing Reactor is 35 to 40 years but could be more or less depending on the issues faced by a particular HVDC system. The oldest in-service air core smoothing reactor as reported by Trench Canada was built in 1980. The DC bushings on the oil filled reactor may limit the life of the oil filled SR with oil filled bushings having a life of 25 years and RIP bushings 30 years. This number of 25 years was derived from statistics for AC bushings. Thus they may not be directly usable for DC Bushings which have a bit more insulation and appear to have a longer life. But this life assessment number merely suggests that an assessment would be prudent at that time. DC bushing failures were fairly common in the past and an analysis of the failures led to implementation of new test levels of 115% of the level that the SR is tested at. DC bushings have a shorter life than the SR itself and if replacement DC bushings are not available for any reason this becomes a major issue. The IEC/IEEE Standard 65700-19-03 (Issued July 10, 2014) is a joint Standard with IEC and IEEE on DC Bushings. It highlights that the DC bushings are not interchangeable, and while not required by the standard, it would be beneficial if the supplier provided sufficient information so any DC bushing manufacturer can supply the bushings. It is possible to replace the DC bushings, the DC barriers, DC bushing leads and field test but this is risky, very expensive and may still result in the SR replacement.

Loose blocking is another concern because of the large amount of paper involved in oil filled SR. If there is sufficient room, it is possible to place jacks in the support beam area and add insulation. The blocking should be checked by internal inspection every 20 years. The internal inspection can detect loose core laminations. For the remainder of the assessment, it is treated the same as a converter transformer (see section 3.3 below).

For air core SR’s, the support insulators are critical for mechanical support and dielectric support. Shed profile, such as long short or anti-fog, is important. After 25 to 30 years, two of the insulators should be removed from service and tested both mechanically and electrically. Failure of the grout between the metal flange and the porcelain is a common cause of problems. Deterioration of the outer coating of the SR itself such that the aluminium conductors are corroded would be a cause for immediate replacement. If a noise barrier is installed in the field the difference in the temperatures from the factory tests and field test would allow for a calculation of loss of life as every 6 to 8 oC is a half-life. This has been the problem of premature air core SR failures.

Page 21 Guidelines for Life Extension of Existing HVDC Systems

Refurbishment:

For an oil filled reactor it is likely that the DC bushings will have to be replaced as oil filled bushings have a 25 year life and RIP has a 30 year life (according to the DC bushing suppliers) whereas the SR itself has a 40 year life. DC bushings noted above has approximately 150 kV DC may have an oil end DC barrier requiring the DC bushings to be replaced with the exact same type, make and model. DGA sampling on all bushings (this must be done with care not to introduce contaminants or moisture into the bushing and some DC Bushings may require re-pressurizing with nitrogen) and dielectric tests at reduced test levels on at least two DC bushings will determine the end of life. The dielectric test level should be at least 10% above the protective level of the associated lightning arrestors. DC Bushings at or below 150 kV do not normally have the oil end DC barrier and can be supplied from a supplier other than the OEM. However, after this period of time the original supplied DC bushings of higher voltage may no longer be available. Then the oil filled SR will have to be replaced and it will likely be replaced with an air core SR. The DC smoothing reactors have large amounts of insulation, especially at higher voltages. This insulation gets squeezed or settles down over a period of 20 to 25 years. If there is sufficient room, consideration should be given to do a field repacking of the insulation with small jacks and oil impregnated pressboard to extend the life of the SR. There does not appear to be any economic benefit to replacing the winding on an oil filled SR as this cost will likely be as great or greater than that of a new air core SR. Most of the gasket materials and O-rings have a life of 40 years (nitrile) used in the SR’s and oil filled bushings will likely require replacement. Various sealants have been used successfully to seal oil leaks to defer the cost of replacing the gaskets or O-rings which can be very expensive when oil processing and outage costs are considered.

For an air core smoothing reactor the most critical area is the outer coating of paint or silicone RTV. This coating has to be refurbished every 10 years (more or less) to protect the SR insulation and applies to both outdoor and indoor installations as most indoor lights emit ultra-violet (UV) radiation. Owners of the air core SR’s may not be aware of this requirement. This coating keeps out UV radiation and moisture which if not refurbished will result in insulation failure, corrosion of the winding and require replacement.

Remaining Life:

The remaining life of a SR is dependent on many items such as years of service, operating conditions, design temperatures, DGA availability of spare/replacement parts, internal inspections of oil filled SR, test results and maintenance records. This usually requires the services of an Independent Expert in that area to determine what may also be required by the management in a company to lend credibility to any decision. Because outage of an SR is a pole outage, outage costs can be very expensive, so the amount of risk (if any) that a company is willing to tolerate is limited. While any decision to refurbish or replace also depends on the cost of outages and economics of a company, only the technical aspects will be considered here. The only economic consideration will be the fact that oil filled SR’s are more expensive than air core reactors both in capital cost and maintenance costs.

The DGA results are very important in making this decision but will be discussed under the Converter Transformer section. The other major factor is the condition of the DC bushings and availability of replacements. Oil filled SR’s have a risk of environment contamination and risk of fire which, if oil containment and fire protection is not provided on the existing equipment, may also assist to justify replacement rather than refurbishment.

Items to consider when determining whether or not to refurbish or replace

Oil filled SR’s and the DC bushings with the following problems are probably best replaced:

Page 22 Guidelines for Life Extension of Existing HVDC Systems

 bad capacitance measurement of power factor measurement, • or critical DGA of the oil, • or DC bushings are no longer available from the OEM and the DC bushing has an oil end DC barrier, • and is over 25 years old. When considering the replacement, it may be possible to reduce the milli-henry inductance of the SR making it more economical. If the DC Bushings are the problem, it may possible to refurbish the SR by replacing the DC leads, DC barriers and DC bushings with RIP and may be considered if the SR is less than 25 years old. However this is costly and you still have an aged SR. Usually these oil filled reactors do not have oil spill containment and fire deluge protection making it easier to justify replacement with air core units. Other problems that could cause immediate replacement is failure of the spare due to unknown causes and high DGA gassing results indicating temperature over 150 oC and paper degradation by CO2/CO ratio. See Converter Transformer DGA Results for further details.

For the air core reactors the deterioration of the outer protective coating, such that the insulation and aluminium conductors are visible, would be cause for immediate replacement. If the outer coating is only slightly cracked or the areas are splotchy; they can be refurbished by recoating, which extends the life of the air core reactor. Another concern is if the air core reactor has a noise shield and it was not required in the factory but was installed afterwards in the field. It is likely that the units will have a reduced life and require replacement sooner than the expected 40 years.

After 25 to 30 years two of the support insulators should be replaced with spares and tested dielectrically and mechanically, for remaining life. Depending on the number of insulators and the cost, it may be advisable just to replace them.

When replacing SR’s, consideration could be given to installing multiple reactors in series to form the design value but providing some installed redundancy to allow for removal of single reactors for future maintenance and/or repair. The allowable tolerance for sizing of reactors needs to be determined through appropriate study.

Remaining life 10 years or more:

If the units are less than 25 to 30 years of age, there will likely be a remaining life of 10 years or more.

For oil filled SR, it is likely that these units will survive 10 years or more:

 if the DC bushings are in good condition and/or spares are still available from the OEM,  the SR itself is not gassing, or at least not badly  an internal inspection shows it to be in generally good condition. For the air core reactors, if the outer coating is in good condition and has been regularly recoated throughout its life and the noise barrier was factory installed it is likely that these units will survive 10 years or more.

Further Assessment Required:

For all other conditions further assessment is required as the number of combinations get complicated. Additionally, company management may require an outside opinion for various reasons. The best option would be an independent analysis with assistance from the OEM supplier.

Page 23 Guidelines for Life Extension of Existing HVDC Systems

Maintenance:

The most important maintenance aspects of the oil filled reactors are regular Dissolved Gas Analysis (DGA) sampling and capacitance and Dissipation Factor (DF) tests on the DC bushings. Other tests such a resistance and Swept Frequency Response Analysis (SFRA) should be performed if the DGA results indicate some possible problems. An initial or prior SFRA test is necessary to have a “signature” to compare with. All the auxiliaries such as fans, pumps and instrumentation should be checked periodically.

A visual inspection of both types of SR is recommended, to look for oil leaks, broken bushing sheds and anything abnormal. Infrared and Corona scope tests are also recommended yearly for both types as well, looking for hot spots on the reactor and any associated bus work and bus connections. All insulators should be checked for cracking or damage and contamination.

For the air core SR keeping the air cooling vents clear of debris and blockage is most important as well as inspection of the outer coating to protect the insulation.

FIGURE 1 OIL FILLED SMOOTHING REACTOR

Page 24 Guidelines for Life Extension of Existing HVDC Systems

FIGURE 2 AIR CORE SMOOTHING REACTOR

3.1.2 Voltage Divider (VDR) Description:

The voltage divider can be either oil or gas filled device consisting of a non-inductive wound resistor with a capacitance in parallel. There is voltage measuring electronics at the base of the unit for sending the signal to the control room via “special” wire or fiber optics. The VDR for higher voltages may be made with two or more sections joined together and the outer housing can be composite or porcelain.

Performance Issues:

The VDR is used in the control of the HVDC system such that any problems with the VDR result in a system disturbance and is quickly noticed. Corona discharges on the exterior of the porcelain have resulted in erratic DC readings and disturbances. Coating with Silicone RTV should eliminate this problem. If more than one section is used there can be and have been problems with the electrical connection between the two sections. While this will again show up as a disturbance, this type of fault is difficult to ascertain as it is intermittent. An oil leak, if severe or unnoticed, can cause a unit to fail. Overall the VDR’s are very reliable. Newer units may have silicone rubber sheds.

Technical Life Assessment:

A VDR is considered to have a life of 40 or more years. Degradation of the capacitance will likely occur at the end of life, and because of the relative low cost, replacement should be considered. While these are normally sealed units, if there is a filling port, an oil sample may be taken for DGA analysis. Again care must be taken to prevent contamination of the oil with moisture or other. Any replacement decisions should consider the advantages of moving from oil to gas or solid insulation filled units.

When fiber optics is used to send the signals to the control room, the associated electronics will likely require repair or replacement prior to the end of life of the actual VDR. Replacement can be avoided by carrying appropriate spares or sourcing identical replacements as required.

Page 25 Guidelines for Life Extension of Existing HVDC Systems

Refurbishment:

In the case of an oil leak and it is not near its end of life, the unit can be refurbished. This is likely cost effective at higher voltages, but is less likely at lower voltages.

Maintenance: The following maintenance should be performed:

 Visual Inspection for oil leaks, tracking on the insulator, corrosion, broken sheds, etc. If gas filled, check for pressure, check and replacement of anti-moisture silica gel in the secondary terminal box.  On a less regular basis the voltage divider scaling and ratio should be checked by primary injection. The procedure should be provided by the OEM.  Corona scope. Initially after installation, after each maintenance and periodically thereafter the unit should be checked for corona especially during fog or light rain  Resistance and Capacitance measurements – During the periodic maintenance outages the resistance and capacitance should be measured.

3.1.3 DC Current Transducer or Transductor (DCCT) or DC Optical Current Transducer (DCOCT)

FIGURE 3 KRAEMER STYLE DCCT

Description:

DCCT‘s have evolved the most of any of the DC measuring devices. The most common ones will be described here but it is most likely others will be encountered. The first one was called the Kraemer type and required a voltage such as 600 Vac to chop up the current wave-form so it could be brought to ground potential via a transformer. The chopped up waveform then was rectified and scaled to represent the DC Current. This required a no-break AC supply usually from a motor generator (MG) set or a battery bank inverter. The next type of DCCT was the Hall Effect where the flux was measured in the DCCT core, which produced a voltage that scaled to represent the DC Current. An improved version of the Hall Effect is the Zero Flux DCCT. This used electronics to produce an equal but opposite flux to

Page 26 Guidelines for Life Extension of Existing HVDC Systems that produced by the Hall Effect. The measurement of the electronic current was scaled to represent the DC Current.

The next type is the DCOCT or Optical DCCT, which is a bit of a misnomer as it is not completely optical. The DC Current is measured by a DC shunt and the millivolt signal is measured by electronics in the head of the DCOCT. The signal is then transmitted via fiber optics to ground potential and to the HVDC controls. The electronics and fibers are duplicated for redundancy and are powered by another fiber optic and laser diode from ground potential. This type of DCOCT can be a free standing device supported by a porcelain insulator or in a fiber glass tube with silicone rubber sheds. It can also be an in- line device supported by the conductor itself and a small thin insulator string housing the fiber optics. The final type of DCOCT is a true fiber optic device and uses the Faraday principle. This is comprised of a fiber optic cable that is wrapped many times around the conductor. A polarized light is injected into the cable at ground potential and the change in the polar angle of the light is detected and measured upon its return. The amount of the angle change corresponds to the amount of DC current. The advantage of this type of DCOCT is that the fibers can be wrapped around the ground flange of a bushing so there is no expensive high voltage to deal with and corresponding reduction in cost. It is highly accurate even at low currents.

There are a number of lower voltages high current DCCT’s which are not discussed here.

The cost of these devices continues to come down and it is possible to have a combined device with a voltage divider, which can potentially further reduce costs if both are required.

Performance Issues:

The porcelain insulator has been a problem with tracking and flashovers. This has caused gassing in the oil and eventual failures of the devices themselves. The application of Silicone RTV has largely resolved this issue. Oil leaks are a continuing problem as the device ages and the paper insulation of transformers in the Kraemer style units show up with age as well. The Kraemer style DCCT’s also require the MG sets or battery bank inverter. However, with lack of spare parts for these devices as well; it soon becomes economical to replace the unit if over 25 to 30 years old once they show signs of problems.

Other problems have been with the electronics and the fiber optic connectors and are relatively easily resolved but the forced outages they cause can be significant.

Technical Life Assessment:

Once the Kraemer style DCCT starts showing signs of aging it is likely best to replace. This allows the removal of the MG sets or battery bank inverter, improving reliability (fewer parts) and reducing maintenance costs. The availability of spare parts and optical fiber (fiber connectors) will determine the technical life of the equipment. The fiber optic outer sheath will deteriorate with age as will the laser diodes. As long as the parts and fibers optic connectors are still available they can likely be kept in- service.

Refurbishment:

In most cases, it is not likely that the devices can be successfully refurbished and considering the cost may also not be economical. Exceptions may be the fiber optic cables and some electronic cards. Removal of the Silicone RTV from porcelain sheds by blasting with walnut shells or dry ice and replacement of the silicone RTV may be necessary in high pollution areas.

Maintenance:

Page 27 Guidelines for Life Extension of Existing HVDC Systems

Periodic testing of the fiber optic cables and laser diodes is necessary. DGA analysis of the oil filled devices should only be done if there is a problem suspected as contamination of the oil is a serious concern. Protecting the fiber optic cable from light can increase the life of the outer sheath in the control room, cable trays and in the cable termination boxes.

 Visual inspection.  Cleaning as necessary

FIGURE 4 DC OPTICAL CURRENT TRANSDUCER

3.1.4 DC Surge Arrestors Description:

DC Surge Arrestors are very similar to AC arrestors but contain about 20% more discs. However, they are a lot more expensive than an equivalent AC arrestor. The number, location and energy ratings vary depending on the insulation co-ordination study for that particular DC scheme. The older units were a porcelain housing and gapped silicon carbide discs. The newer ones are gapless zinc metal oxide and the housing may be made of fiberglass with silicone rubber sheds instead of porcelain.

Performance Issues:

Moisture ingress is one of the most important issues as a surge can then cause the unit to fail and it may fail catastrophically in spite of venting devices. A similar problem can occur in the counter and if it is located at ground level can be a safety hazard as it fails catastrophically.

Technical Life Assessment:

The DC surge arresters should last for 40 or more years. A lot will depend on the number of switching operations and the sealing against moisture. High voltage tests are done periodically on the gapped silicon carbide arrestors and if defective they are replaced. For the gapless metal oxide the leakage

Page 28 Guidelines for Life Extension of Existing HVDC Systems current is measured in-service or during a test and if out of range the unit can be replaced and tested. If several units of a similar type are replaced; consideration should be given to replacing them all. The life of silicone rubber sheds is currently not known as they have not been in-service for very long. However silicone RTV has been in-service for over 25 years without recoating, so it is expected the silicone rubber sheds will last at least that long and possibly longer.

Refurbishment:

DC arresters are not normally refurbished.

Maintenance:

• Visual inspection • Cleaning as necessary • Testing as outlined above

3.1.5 DC Support insulators and Bus work Description:

The DC support insulators may also be called post insulators and look very similar to AC support insulators. They support the DC bus work and other equipment such as the air core type SR. DC support insulators fabricated with a fiber glass core and silicone rubber sheds have shown improvements in voltage withstand, increased mechanical strength, less weight and less cost. They are new to the market but have been adopted as a replacement in one instance.

Performance Issues:

The DC support insulator is affected by more than just creepage and strike distance such as the case for AC support insulators. It is affected by shed profile, insulator material and voltage stress in kV/mm. This is not always appreciated even by the HVDC suppliers but has become better known since 800 kV HVDC was achieved. As an example a rectifier HVDC Converter Station built in the 1970’s had an increased creepage applied for the DC switchyard design. Yet in service they experienced an average of 14 DC side flashovers per year and likely as many at the Inverter Station. All the support insulators were coated with silicone RTV and no flashovers have been recorded in the 3 years since. A higher DC test voltage was obtained in laboratory tests with an alternating long short shed profile or an anti-fog shed profile than sheds that are uniform in length or helicoidal (spiral). Silicone rubber has performed better in the field and in tests it supports a higher voltage before flashover than porcelain. Corona scope measurements in the field also show an improvement.

Deterioration of the grout between the metal flange and the porcelain is an issue and can go unnoticed until a failure occurs or the bus work is removed and the metal flange separates from the porcelain at that time. Vibration and/or contaminants in the grout can weaken the support insulator and cause failures.

Green glass support insulators were produced with a contaminant in the glass. This contaminant grows with time and eventually shatters the sheds. This becomes a safety issue as well as a performance issue and will likely require their replacement. This problem has not appeared in the glass DC transmission line insulators.

Technical Life Assessment:

The life of a support insulator should be 40 years. Flashover performance of porcelain insulators is an issue in some HVDC schemes. Spraying on silicone RTV has solved the problem and can have a

Page 29 Guidelines for Life Extension of Existing HVDC Systems service life of over 25 years depending on pollution levels. Silicone RTV can then be removed by blasting with walnut shells or dry ice so as to not damage the porcelain. The insulator can then be re- coated with silicone RTV coating. It is also noted that silicone RTV has been applied on porcelain support insulators for a new 800 kV DC station in China.

After about 20 to 25 years, it is advisable to remove 2 or 3 representative support insulators and test them mechanically and electrically.

Refurbishment:

Apply an RTV coating on porcelain support insulators if DC side flashovers are a problem.

Maintenance:

• Visual inspection. • Cleaning as necessary

FIGURE 5 DC POST INSULATOR AND DCCT RED CAP DEVICE

3.1.6 DC Switches – See also AC Switchyard Description:

The DC switches are for the most part just AC equipment but a single phase complete with operating mechanism. The high speed switches are AC breakers which are interlocked such that the DC current must be below 50 Amperes before they can open. There are also some “specialty” devices called

Page 30 Guidelines for Life Extension of Existing HVDC Systems

Metallic Return Transfer Breaker (MRTB) and Ground Return Transfer Breaker (GRTB) which commutates the current into the DC line conductors or back into the ground electrode as required. The specialty devices comprise an inductance and capacitance which create an oscillatory circuit for a current zero. This allows the breaker to open at a current zero and transfer or commutate the current. The older converter stations and some of the new high voltage 800 kV have more than one valve group in a pole and require Bypass Switches (BPS) (AC breaker) and disconnects. The newer BPS is also a special SF6 breaker which must be able to commutate the current into the Thyristor Valve. The older BPS’s were air blast devices. There was also a bypass vacuum switch (BPVS) used in at least one HVDC link.

DC breakers are not covered by this TB

Page 31 Guidelines for Life Extension of Existing HVDC Systems

3.2. Valves Description:

The first HVDC Valves were mercury arc valves which started in 1954 with a link between the island of Gotland and Sweden. Virtually all mercury arc valves have been removed from service and replaced with Thyristor Valves. In the early 1970’s the HVDC valve design switched over to Thyristor valves. Many of the early schemes had lower power schemes, were air cooled and air insulated but as higher power levels were required the design shifted to air insulated and de-ionized water cooled starting in the late 1970’s. There was even one scheme, the Cahora Bassa project between Mozambique and South Africa, which was oil insulated and oil cooled. The South African end, called Apollo Converter Station, was converted to air insulated and de-ionized water cooled, while the Mozambique end called Songo Converter Station remains oil cooled at this time. Also for higher current levels, there were two or more matched (for forward voltage drop) Thyristors in parallel, but Thyristors developed quickly and soon only one Thyristor in parallel was required for most ratings. There were many Thyristor in series to achieve the voltage level required with a high of 280 devices for the early Thyristors to a low of 48 or less for more modern ones with most being in between. In about the year 2000, light triggered Thyristors became available, which reduces the amount of valve electronics that are normally required by the electrically triggered Thyristors.

Each valve hall has either six Thyristor Valves for 6 pulse operation or 12 Thyristor Valves for 12 pulse operation. They can be arranged in single unit, two in a unit called a Bi-Valve (double valves) or four in a unit called a Quadra-Valve (quadruple valves). The units can be floor mounted if there are no seismic concerns or hung from the ceiling if there are seismic concerns.

Each Thyristor Valve is broken down into “representative sections” which allows some of the factory tests to be done on a “representative section” rather than the full Thyristor Valves, reducing the cost of testing. As an example in the case of 280 Thyristors it is broken up into 20 sections, so 14 Thyristors per section. Included in each section is a Valve Reactor and two Thyristor Modules with 7 Thyristors in series and two in parallel in each module. The Thyristor is susceptible to excessive rate of change of voltage (dv/dt) and rate of change of current (di/dt), both must be kept below a stated value in order not to cause thyristor failures. The valve reactor is designed to limit the rate of change of current (di/dt) when the Thyristor Valve is turned on or off to a level below that which the Thyristor itself is capable of surviving. The valve reactor contains a number of iron cores which saturate at different current levels providing high impedance during turn on or turn off and yet low impedance when the Thyristor is fully conducting. For water cooled valves the valve reactor is also usually water cooled with the cooling circuit buried in the reactor winding. The valve reactors also have a damper winding or a resistor across it to damp out any resonances that may be associated with in the circuit.

Page 32 Guidelines for Life Extension of Existing HVDC Systems

FIGURE 6 50 MM THYRISTOR AND 100 MM THYRISTOR

FIGURE 7 A DISSECTED VIEW OF A 100 MM THYRISTOR

Across each Thyristor level is a damping circuit (also called snubber circuit) consisting of a non-inductive wound resistor and a capacitor. The capacitor was oil filled in the older valves but is more likely SF6 filled or dry type in newer valves to reduce the risk of fire. The snubber circuit is required to limit the rate of change of voltage (dv/dt) across a Thyristor to a level below that which the Thyristor itself is capable of surviving. There may be another damping circuit across each valve section or across the entire valve assembly. Also, there may be a Voltage Divider across each Thyristor level to measure the voltage across the Thyristor when it is not conducting which is another non-inductive wound resistor. If the Thyristor fails it is short circuited; this is then used to determine that a Thyristor has failed as there is no reverse voltage across it when the voltage reverses. Across each Thyristor Valve section may have a capacitor which provides transient and temporary voltage grading across the entire Thyristor Valve structure.

While there are a number of important parameters in choosing a Thyristor for a valve design, one very important parameter is Qrr or the stored charge when a Thyristor turns off. The stored charge for each Thyristor is measured in the factory and assigned to a band or range of stored charge. Each band is then

Page 33 Guidelines for Life Extension of Existing HVDC Systems assigned a color number or letter. Each Thyristor Valve has the same Qrr band and replacement Thyristors must have the same color as an example. In some cases a Thyristor from a neighboring band are allowed as well. For example if the required Thyristor is a B level Qrr it may be acceptable to use an A or C level but not a D level. In modern day Thyristors there is a large variation in Qrr that is acceptable and matching may no longer be required

The Thyristors and some of the components listed above are usually installed in a module. It may be possible to remove a module from service and replace the defective module with a good module or repairs completed in place without removing the module. If the module is fixed in place it usually becomes part of the valve structure. The advantage of the removable module was that it could be replaced quickly allowing the valve group to be quickly returned to service. The disadvantage was the increased cost and the risk of bad electrical and water connections which have to be reconnected each time. Most of the more modern designs now have the module fixed in place. Each Thyristor is triggered either electronically from some electronic cards or from an optical fiber connected directly to the Thyristor gating terminal. The electronic cards are called many different names by the different suppliers and have even changed names over time for the same supplier. They will be generically referred to in this document as Thyristor Control Unit (TCU). The TCU has optical receivers to convert the optical signal received from the Valve Base Electronics (VBE) via the optical fibers to an electrical signal to trigger the Thyristors. The function of the VBE is discussed in the DC Controls and Protection Section 4.5 below. The fiber optics can be glass fiber or “plastics” fibers. The TCU will also provide power to the electrical cards usually via the DC voltage across the Thyristor when it is not conducting and stored in a capacitor. The capacitor energy storage allows it to ride through AC system faults of typically 100 milliseconds but may be longer in the case of breaker failure.

FIGURE 8 NELSON RIVER BIPOLE II REACTOR MODULE

Page 34 Guidelines for Life Extension of Existing HVDC Systems

FIGURE 9 NELSON RIVER BIPOLE II THYRISTOR MODULE WITH TWO 50 MM THYRISTORS IN PARALLEL, A WATER COOLED RESISTOR AND CAPACITOR DAMPING CIRCUIT. THIS MODULE IS REMOVABLE FROM THE VALVE STRUCTURE.

FIGURE 10 LAMAR THYRISTOR MODULE WHICH IS PART OF THE VALVE STRUCTURE INCLUDES A SINGLE THYRISTOR IN SERIES AND A DAMPING CIRCUIT

In the case of a DC side fault in the converter station the amount of energy that the Thyristor can handle is limited. This requires the tripping of the AC breakers in a relative short period of time such as 2 to 3 cycles. To achieve this speed, a Thyristor or other solid state device direct tripping of the AC breakers, is provided instead of relays.

The control of the triggering point in the waveform determines if it is used in the rectifier mode or the inverter mode. In the rectifier mode the thyristor is triggered with an alpha between 12 and 18 degrees electrical whereas in the inverter mode it is triggered with an alpha of about 135 degrees electrical.

The Thyristor modules are stacked in tiers with post insulators between the tiers. There is corona shielding on the top and bottom and around each tier as required. There are usually fire barriers between

Page 35 Guidelines for Life Extension of Existing HVDC Systems tiers to prevent the “chimney effect” (which is for the fire to race upward) in the event of a fire as there have been a number of Thyristor Valve fires over the years. There may also be fire barriers in the tiers to separate components such as the oil filled capacitor.

FIGURE 11 NELSON RIVER POLE II VALVES

The fiber optics is installed in a channel with a cover on it making it difficult to inspect for defective fiber optics. The outer coating of the fiber optics for the early fibers will deteriorate with ultraviolet light. Many HVDC schemes operate with the lights out in the valve hall for that reason and to save energy.

Thyristor Valve Hall:

The Thyristor Valve hall is sized to provide adequate air gap clearance for the voltage that it is operating at. The valve hall will normally be air conditioned for the older water cooled Thyristor Valves and not air conditioned for the more modern ones. In the case of non-air conditioned valve halls, the temperature can reach 50 or 60 oC. This has an impact on personnel that may be working in the valve hall for maintenance and has a design impact on all the components in the valve halls such as wall bushings, bus-work, as they historically were designed for lower temperatures. Other components in the valve hall include converter transformer bushings which stick through the wall and are sometimes mistakenly called wall bushings. For the DC wall bushings, they normally have a DCCT at the flange or there may be a free standing DCCT to measure the DC current into or out of the valve hall. In some older projects, a capacitor connected across the valve group for transient voltage grading for multiple Valve Groups in a pole. The capacitor is usually placed inside the porcelain housing so it may not be obvious that it is a capacitor. There is also a DC valve arrestor across each Thyristor Valve which may or may not be part of the Thyristor Valve structure. The valve arrestor is part of the insulation co-ordination study for the converter station and the highest voltage valve DC arrestor may have a higher energy rating than the remaining ones in the valve hall.

There may be more than one valve group in series in a pole (4 in the case of Cahora Bassa) but one is typical for more modern schemes up to 500 kV DC. With the advent of 800 kV there are one or two

Page 36 Guidelines for Life Extension of Existing HVDC Systems valve groups per pole and it is anticipated this will be two valve groups per pole for 1100 kV DC which is currently under development.

The valve hall is usually equipped with fire detection systems but very rarely a fire suppression system. Fire suppression systems within the valve hall, if considered, should be manual action with at least two deluge valves in series of different manufacture. The system should not be pre-charged. The fire detection system can be regular smoke or ion detectors, air sampling systems and/or beam detectors. The air sampling systems, of which there are two different types, provide the quickest response. A thermal barrier during a fire at the ceiling can prevent regular smoke or ion detectors from operating properly. The high air flow rates of the air cooled valves will make fire detection even more challenging.

Thyristor Valve Cooling systems:

The Thyristor Valves generate heat from the losses associated with the forward voltage drop and load current through the Thyristors and valve reactors when they conduct and from the snubber circuits when they operate during turn on and turn off. Heat from the other components in the valve hall will make up the remaining heat loss but this is usually minimal.

In the case of water cooled valves approximately 5 to 10% of the heat is removed by the air cooling system and 90 to 95 % by the De-ionized Water (DIW) cooling system.

The cooling air is routed via channels through the valve structure and normally comes from outside air. In colder climates there will be some air recirculation and heaters. This is discussed in more detail in section 3.6 below. The air must be highly filtered to prevent dust accumulation and the valve hall is usually pressurized to further prevent dust infiltration.

In the case of air cooled valves, there are usually fans to direct the air through the valve structure to ensure adequate cooling throughout the valve and 100% of the valve is cooled by air.

The deionized water cooled Thyristors have a heat sink on each side of the thyristor “puck” which are tightly clamped at a high contact pressure and may include a grease or similar compound to prevent oxidation and enhance thermal conduction. The heat sinks must have a low thermal resistance. The snubber resistors are also likely water cooled as are the valve reactors modules. The valve reactor modules will have the cooling circuit imbedded in the main reactor winding.

De-ionized Water (DIW) is a very efficient heat transfer media and has become the norm for modern Thyristor Valves. It allows for a more compact design, higher power levels and is used in closed loop systems. The Thyristor heat sinks are fed from manifolds in parallel from Siemens and Alston Grid but in series in each module from ABB. The water needs to be deionized to remove free ions and minimize flow in the water in the tubes. This is accomplished with cation and anion resin beds, which require regular maintenance. One supplier, ABB does not vent the deionized water system and thus has oxygen scavengers in the resin beds as well. The other two suppliers vent to air as breakdown of the water into Hydrogen and Oxygen occurs at high voltage. In addition there are sacrificial anodes in the water circuit made of stainless steel (Alstom Grid) or sacrificial anodes of platinum (ABB and Siemens) to prevent corrosion. These must be checked periodically for corrosion or deposits.

The water circuit may be a single loop system (includes industrial grade glycol for cooler regions) or a double loop system. The double loop system will have DIW in the Thyristor Valve circuit and regular water or glycol in the outdoor cooling circuit, glycol for cooler regions. A single loop system brings the DIW to an outdoor water to air cooler whereas the double loop system has an intermediate heat exchanger. The Thyristor Valve cooling water pipes must be made of a material which has very high electrical insulating properties and to have a long life (25 years or more) must withstand continuous

Page 37

Guidelines for Life Extension of Existing HVDC Systems temperatures in excess of 60 oC. These cooling pipes were a problem in older valves as they had a short life but newer materials made of cross linked polyethylene (XLPE) and Teflon which have much longer useful lives have been used in modern valve cooling circuits.

Note that the remainder of the cooling circuit will be discussed separately.

Performance issues:

The cooling circuits have been and still continue to be a source of problems in Thyristor Valves. Deterioration of the older plastic tubing is a significant problem in some HVDC schemes such as Nelson River Bipole 2. Main issues verified in the past include:

 Corrosion  Fouling of the platinum electrodes, resulting in additional scheduled outages to clean them  O-rings and gasket material deterioration from DIW exposure Problems with the resin beds, bad resin and contamination of the resins are common and a major consideration should be a proven supplier and not price alone.

Thyristor aging:

There have been cases where thyristors have suffered from premature aging issues (not necessarily thyristor failure), either from inadequate design, manufacturing defects or operational stresses. The thyristor is the fundamental component of the HVDC system. Any aging issues that could affect the design parameters could ultimately lead to a cascading failure of the series connected thyristors [6]. As a HVDC scheme ages, it is likely that it can no longer get Thyristors from the OEM, or the cost of replacements is extremely high. There may be other suppliers which can fill the void but careful matching must be adhered to, especially with the Qrr or stored charge on turn off. To ensure an owner is prepared to handle any thyristor issues, a full data specification for the thyristor device as well as documented testing should be provided as part of the station documentation.

Some schemes are experiencing higher than normal failure rates of the Thyristors and the causes are not always known. Some are associated with mixing new Thyristors with old ones from different suppliers. In one particular case of Thyristor failures is that they always occurred after testing the Thyristor in the OEM supplied test set. The testing in the OEM supplied test set was discontinued and no more failures occurred.

Failure of valve reactor modules is fairly common with age due to lack of cooling flow (plugging) and severe vibration causing deterioration of the iron cores.

The fiber optic outer coating has deteriorated due to ultraviolet light in some schemes. Replacements can usually be found but can be a challenge because the connectors are obsolete. Also, the laser diode transmitter and receivers can be a challenge to find replacements for.

DC wall bushing failures increase with age, requiring their replacement. Flashover of the outside portion of the DC wall bushings is quite common due to the rain effect from the converter building, where a part of the bushing is dry as it was shielded by the building during a rain storm and the other part is wet. This causes voltage stress across the bushing, resulting in a flashover. Increasing the DC wall bushing length or creepage alone does not work as anticipated. An alternating long short shed profile or “booster sheds” or silicone rubber sheds may be required depending on the voltage level and pollution. There is an opportunity to move away from porcelain oil filled bushing to either gas filled or composite solid core bushings.

Page 38

Guidelines for Life Extension of Existing HVDC Systems

The Thyristor Control Unit (TCU) contains a number of printed circuits boards (PCB) and there can be poor solder connections as well as parts obsolescence. In one case over 50 000 PCB’s were re- soldered because of solder joint issues. While this was a big job it was a lot less costly than replacement. Parts on the PCB’s can also be an issue but again parts can usually be sourced. In some cases the individual electronic chips can be manufactured by a semiconductor supplier. In others the inputs and output signals on a PCB are replicated by a new PCB with modern, available parts. This has been done in the case of a few utilities but not all utilities have these capabilities and they must find them elsewhere or replace the cards of an entire TCU. In modern day HVDC schemes this may no longer be possible as everything is done in software and there may not be any access to the software code. In this case replacement could be the only option.

There have been two separate incidents of an entire Thyristor Valve failing. One was when a wall bushing flashed over just as a Thyristor Valve was starting to turn-on and the current had not spread out entirely in the Thyristor wafer thus having a large current in a small area. The second was when an AC breaker bay opened under load at an inverter station. It appears that the solution is to block the Valve Group when the trip signal goes to the breaker trip coil. While this is a very rare event it highlights the requirement to carry sufficient spare Thyristors with the same Qrr ratings to make up a complete valve. This requirement would be based on the cost of the outage versus the probability of the failure.

Problems with the air cooled valves include failure of the pulse transformers (for valves that use them) and deposits of dust and other contaminants.

Technical Life Assessment:

In general, the Thyristors themselves are showing no signs of end of life. For example, tests on the Cahora Bassa Thyristors recently showed no end of life and Nelson River Bipole 2 is not having abnormally high failure rates. Usually by about the 20 to 25 year mark in a 40 year life, the OEM will no longer supply any replacement Thyristors or the cost becomes very high. This is usually about the time spares run low and become expensive to stock. They can usually be obtained from another semiconductor supplier provided there is sufficient documentation required or sufficient units to test properly. This can be hit and miss especially because of the Qrr requirements and the mixing of units from different suppliers, but usually a successful replacement can be achieved. A semiconductor supplier may even custom make these which can be expensive but may still be a lot less costly than complete replacement. Utilities need to manage their spare stock of Thyristors as the project goes into service since the price is very favourable and there is no known shelf life. However, design and quality issues may be a concern.

The valve reactor modules are subject to intense vibrations as they turn on and off which can affect the steel cores and produce a reddish dust of iron oxide which is conductive. Overheating due to cooling blockage and similar problems is common in several projects. The OEM can usually manufacture replacement valve reactor modules but they are quite expensive (hundreds of thousands of dollars each) and come with the associated manufacturing risks, including quality. It is likely that additional valve reactor modules will have to be purchased or in some cases all of them replaced in order to maintain the life of the Thyristor Valve itself.

Lack of support from the OEM can cause problems with the OEM supplied “Special Test Equipment” for testing the Thyristors, valve reactor and TCU card checks. The OEM can and will design a new piece of test equipment but it will likely be expensive.

The cooling system is usually adequate for about 25 years after which it requires refurbishing and or replacement. Digital based cooling controls will likely require replacement prior to actual cooling

Page 39

Guidelines for Life Extension of Existing HVDC Systems hardware end of life as a result of availability of spares and the ability to communicate/program the controls.

The decision to replace or do a major refurbishment on the Thyristor Valves is very challenging. The redundancy of the equipment makes the forced outage numbers low and the durations short. The high cost of replacement plus the lost revenue due to the outage required makes a business case justification difficult. The replacement outage duration can be minimized by doing some up front work in the planning stages but will result in some additional costs so there is a trade-off. In spite of these additional costs, reducing the outage duration is usually beneficial.

One commonly used method to assess the refurbishment option is to create a matrix with the Components in a column and the factors weighted in each row as Satisfactory, Requires Improvement, and Unsatisfactory. An example is shown in Table 1 below. Each of these decisions must be supported with hard data discussed with a group of company representatives for the different areas of the company, as they can be somewhat subjective. This can then show if the problem is the valve cooling which, if replaced, can improve the system or if there is a general concern in most critical areas indicating that a complete replacement is necessary. The critical areas are usually the higher cost items such as the Thyristors, valve reactors etc. and must have a higher weighting than other items such as a snubber capacitors which is relatively inexpensive to replace. An example of this has shown that the cooling piping was causing a 0.3 % forced outage rate and replacement was easily justifiable. There was a perception that the entire Thyristor Valve required replacement before this analysis was complete. The cost of the piping was less than one year’s interest on the cost of the complete replacement.

Table 1: Example of an assessment matrix for refurbishment1

Forced Weighting Maintenance Staff Components Outage Replacements Factor Outage Rate skills Rate Thyristors 10 S S S RI

Reactors 10 U U S S

Snubber 4 S S U S Capacitor

Snubber 4 S S U S Resistor

Thyristor 10 S U RI S Control Unit Fiber Optics 4 S S U S

Cooling System 10 S S RI S

1 Satisfactory (S), Requires Improvement (RI), and Unsatisfactory (U)

Page 40

Guidelines for Life Extension of Existing HVDC Systems

Maintenance:

The Thyristor Valves themselves usually require very little maintenance. The exception is the deionized water and air cooling systems.

When the valve hall is in-service the TCU electronics, fiber optic cables, Break over Diodes (BOD), refiring of the Thyristors and Thyristors themselves are continuously monitored. There are redundant Thyristor levels ranging from 3% to 8% of the Thyristor Valve. Should the redundancy be used up, the Thyristor Valve will trip and cannot be de-blocked until it is repaired or replaced. Deionized water usage is monitored and if excessive could indicate a small leak. Topping up with deionized water from ground level is a must. A large water leak will be picked up by a leak detector and trip the valve group. In some cases the leaked water has mixed with dust in the fiber optic duct and caused a valve flashover destroying the fiber optic cables. Other in-service maintenance includes changing the resin beds, air filters and any defective fans or fan belts. Infrared scans should be performed twice per year and viewing ports may be required to access all portions of the valve hall. In one case viewing ports were added to allow infrared scanning while the equipment is in operation. The viewing ports are opened only for the duration for the infrared scans and are closed after use. If a device is available, ultraviolet light corona scope scans, are to check for defective insulators and problems with the corona shielding of the structure. These scans should be done once per year. Changes in audible noise levels can indicate an impending problem. Changes in the “smell” of the air also can indicate a component failure which is self- extinguishing and not necessarily picked by the fire/smoke detection. These should be investigated as required.

When the valve hall is out of service, it is important to analyze and replace any defective components in the TCU electronics, fiber optic cables and Thyristors themselves that showed up by the continuous monitoring. The OEM will have supplier test equipment and documentation to successfully diagnose and fix these problems and it is not intended to go into details in this document.

It is very beneficial to have check lists and have the person or persons initial and not just check the item when it is complete. This allows an investigation of a failure to know first-hand the condition prior to the failure by talking to the correct person, but is not intended as a punishment system for errors. If possible a person separate from the person doing the initial work should check to see that the equipment has been placed back securely. This is very important where modules are replaced, to make sure the electrical and water connections have been made properly but can also apply to other equipment. Errors can be very costly in terms of outage cost and equipment replacements.

Any outstanding infrared scan hot connections and corona scope defective insulators, among others, are addressed at this time also.

One of the most important aspects is a thorough visual inspection looking for anything out of the ordinary, dust accumulation, black spots, and spray water evaporation residue are some factors. A partial list is outlined below but is not intended to supersede the OEM’s check list:

• Check the electrodes for corrosion or deposits. • Check the valve arrestors visually, record and trend the number of operations. If the number of operations is more than a few per year this should be investigated and the arrestors tested as required. • A percentage of the Thyristors may be checked with the OEM supplied test set, but the Thyristor is very robust and this does not usually show up any failures. In one instance the testing of the

Page 41

Guidelines for Life Extension of Existing HVDC Systems

Thyristors in the test set was causing premature failures of the units and was discontinued. It is more likely to be used if a problem is suspected with a group of Thyristors for some reason. • Visually check the valve reactors for red dust indicating the cores are loose, vibrating, corrosion and any signs of overheating. If so they should be replaced, analyzed for the Root Cause of the problem and tested by the OEM supplied test set as required. • Clean the wall bushings, insulators and equipment based on dust accumulation. • Examine the fiber optic cables in the cable trough by opening and checking. Note: A check of the end to end light transmission will not show a deterioration of the outer protective coating. • Check the Valve Group Grading Capacitor capacitance and Dissipation Factor, if there are multiple groups in a pole. This capacitor may be installed in a porcelain housing and look like a support insulator and overlooked if not careful. In one instance the utility did not even realize that it was a capacitor until one failed. • Check the Thyristor Valve transient voltage grading capacitor for capacitance and Dissipation Factor. This is used to provide transient voltage grading within each valve tower and may be difficult to access, thus it may not be obvious that there is one to check. • Replacement components must be of high quality and are usually much higher in cost but well worth the money. As an example electrolytic capacitors used in the PCB’s in the TCU are available in 2 000 hour and 1 000 000 hour varieties with a significant cost difference. Obviously changing the capacitors every 3 months for the 2000 hour units is not viable. • Components may not be available from the suppliers or are very costly. Other sources are usually available but one must do sufficient reverse engineering to ensure that they will work properly or pay the price. • Check the safety interlocks and grounding switches as required. • Check the deionized water piping and manifolds for any cracks or leaks. • Check the cooling fans, filters and heat exchangers and replace or clean as necessary. • Replace any lighting and repaint as necessary. • Check the smoke, air sampling system and fire alarm detectors for proper operation, fouling and response time as required by RCM and or the fire code.

3.3. Converter Transformers

FIGURE 12 500 KV DC CONVERTER TRANSFORMER

Page 42

Guidelines for Life Extension of Existing HVDC Systems

Description:

HVDC converter transformers are, in many ways, similar to AC transformers. It is not intended to duplicate these aspects here, thus some knowledge of AC transformers is required. This section will address the differences between HVDC converter transformers and AC transformers, and the impact on design, testing, operations, maintenance, and life extension.

HVDC converter transformers have an AC voltage component across the windings and a DC voltage component or bias on the ungrounded valve-side to ground. The older designs and the 800 kV HVDC converter transformers will be a two winding transformer which have either a set of delta connected valve-side secondary windings or a set of wye connected valve-side secondary windings. More recent designs (and those rated below 800 kV) will be a three winding transformer and have valve-side windings with a set of both delta and wye connected windings. The windings carry a DC current with a roughly rectangular wave shape. This rectangular wave shape contains the harmonic currents, which together with the resulting harmonic fluxes, must be adequately catered for in the design of the HVDC converter transformer. The harmonic fluxes do not cancel in certain parts of the transformer and must be considered in the design. This was not always done in the past and has resulted in some failures.

The core must be designed to handle small amounts of DC current typically in the range of 3 to 15 amperes without saturation. The difference in impedances between phases must be 5% (guarantee) or less and have been specified at 3% or less on newer transformers. The valve-side winding is enclosed with cellulose-based insulation (pressboard and paper) to contain the DC voltage stresses. The DC bushing oil end above approximately 150 kV DC forms an integral part of the insulation structure of the unit and is enclosed with a DC insulation or molded pressboard barrier system making DC bushing interchangeability in most cases not possible.

Most of HVDC converter transformers have a top rating designed as Oil Directed Air Forced or ODAF. There may be some rated Oil Forced Air Forced (OFAF). This rating is for the cooling provided by the oil that is channeled by a system of pipes and manifolds directly from the oil pump(s) into the core and windings in the appropriate quantities and subsequently moved to the external cooler(s) or radiator(s). The core and windings then have oil ducts and spacers to allow them to be adequately cooled by the forced oil flow. The lead entry points usually have additional insulation making cooling in this area even more challenging.

Converter Transformers:

The design and manufacturing process is more critical in the case of converter transformers because of their complexity, DC cocoon, DC bushing part of the insulation structure, harmonics, and the need to consider both AC and DC current and voltages. Most of the issues are covered in “IEEE Standard C57.129 - IEEE Standard General Requirements and Test Code for Oil immersed HVDC Converter Transformers”. Note: this standard is currently under revision as a joint standard with IEC. The highlights will be covered here but the details are provided in the standard.

Page 43

Guidelines for Life Extension of Existing HVDC Systems

DC Bushings:

FIGURE 13 DC BUSHING

FIGURE 14 DC BUSHING OIL END BARRIER SYSTEM

The DC bushings at first glance appear similar to AC bushings but there are significant differences. Some are mounted vertically for outdoor operation and some are mounted horizontally for insertion into a valve hall.

The DC bushings have about 20% more insulation in the condenser core than an equivalent AC bushing. The porcelain is made from “special” low sodium, low resistivity material (resistivity of ~1010 m compared to ~1012-1014 m for ‘standard’ porcelain). “Composite” bushings with an epoxy resin core and silicone rubber sheds have become more common especially for new HVDC schemes as it removes oil from the Valve Halls. For DC bushings rated above about 150 kV, the oil end of a DC bushing (including the condenser core) is an integral part of the insulation structure of the converter transformer. The condenser foils are aligned with the DC barriers such that only a bushing with the exact make, type, and model can be used for a replacement. The corona shield either is attached to the oil end of the bushing or is a “tulip” imbedded into the DC barrier structure. A voltage tap may be used to provide a timing

Page 44

Guidelines for Life Extension of Existing HVDC Systems signal to the valve group. Additionally there is both an AC voltage component and a DC voltage component across the bushing. The dielectric withstand of the various materials to DC voltage is quite different and varies a lot with temperature so extreme care in the use of these materials and Finite Element modeling (FEM) is critical.

Tap Changers:

FIGURE 15 TAP CHANGER

The tap changer used in a converter transformer is, in many respects, similar to that in AC transformers however the number of operations is higher in converter transformers the OLTC is used to control the valve side voltage in order to keep the firing / extinction angles in a predefined band. The converter transformer tap changer will receive a raise and lower command from the DC controls and provide tap position status back to them.

A tap changer is made up of two major parts, the selector switch (no arcing) usually in the main tank and the diverter (does the switching) in a separate oil vessel.

Newer tap changers may be with a vacuum interrupter used for the diverter switch which then has no oil and hence much lower maintenance needs.

The wave shape of the recovery voltage, amongst other aspects, determines the arc quenching capability of the switching gap in the diverter. Therefore, the recovery voltages caused by the square wave shaped through-current of an HVDC transformer must be verified - it is required to know the steepness (di/dt) of the through-current at the current zero. It should be noted that the vacuum type LTCs are not affected by this effect as much as the conventional oil-type LTCs due to a very fast dielectric recovery of vacuum interrupters (~10kV/s).

Page 45

Guidelines for Life Extension of Existing HVDC Systems

Core and Windings:

FIGURE 16 DC SIDE CORE AND WINDINGS

The magnetic core is made up of many very thin sheets of electrical steel with a few percent addition of silicon which increases the electrical resistivity of the steel, in turn decreasing the eddy currents, resulting in a lower core loss. Typically these sheets are manufactured with 0.23mm or 0.27 mm thickness the thinner the sheet, the lower the eddy losses. They can be laser or mechanically scribed to further reduce the core loss. Each sheet is coated with a thin clear insulating material. It The assembled core is supported with a clamping structure, which is insulated from a core and a tank to prevent the flow of circulating currents, but still galvanically bonded and then brought out of the tank through a “spark- plug” type bushing to ground. The sheets may be glued together in some locations (typically the core legs and bottom yoke) for additional support. The sheets are cut to size using a computer operated “GEORG” core shearing machine and either automatically or manually stacked.

Each winding is individually wound and then inserted into a leg of the core with spacers. The top yoke of the core is manually stacked and then the winding connections are made. The DC winding is enclosed in a “DC cocoon”, i.e. several layers of pressboard insulation. The DC insulation may create a number of potential design concerns, which must be properly addressed, or the transformer may fail test, fail in the field, or have a reduced life. At the point where the DC winding is brought out additional insulation may be required to withstand the dielectric tests, however, this can reduce the cooling in that area and cause overheating. The DC cocoon also has more insulation than an AC transformer therefore it can lead to problems during the vapour phase drying process and higher concentration of moisture in inner pressboard layers. Numerous layers of solid insulation form a specific geometry at the winding ends, with a large distance to the yokes, much higher than that for a conventional AC transformer. This configuration forces the magnetic flux to change the direction from axial to horizontal at the winding ends as the core legs are closer to the winding than the core yokes. In result, the magnetic flux which throughout most of the winding had an axial direction and was penetrating through a narrower thickness of the strands, changes the direction to horizontal and it’s penetrating the wider dimension of the strands, significantly increasing eddy currents and eddy losses in these locations. This results in higher temperatures and cooling requirements.

Page 46

Guidelines for Life Extension of Existing HVDC Systems

For converter transformers with both a wye and delta valve side winding the harmonic fluxes add in the center of the winding plus other potential locations, which must be catered for, especially for enter/exit winding leads. This is why the hot spot must be calculated by the manufacturer for each transformer design and general hot-spot factors, such as 1.3 multiplier allowed by CSA standard, cannot be used. The overheating of these areas will not usually show up during the load run test as it is too short, but will later lead to gassing and possible failure in the field.

Some manufacturers of converter transformers still use the OFAF rating, but most manufacturers now use the ODAF rating. The temperature rise across an OFAF rated winding is approximately 30°C (86°F) whereas it is approximately 10°C (50°F) for the ODAF. In the ODAF case the oil is pumped into a manifold either directly or through some insulated pipes and then allowed to move through the winding or core by the size of the openings and the path restriction.

In the case of OFAF, the oil is dumped into the bottom of a transformer tank and then moved through the winding or core by convection from the heat generated and is channeled using the size of the openings and the path restriction.

Converter transformers may also have an ONAN or ONAF rating but this will likely never be used as this restricts all the other transformers in the bipole and DC is in series and thus it is better to ensure that the units are always at full rating capability.

Fiber optic temperature probes have been added to many converter transformers to try to measure the hot spot and the effect of harmonics, as they cannot be tested for in the factory. Initially they were placed in low voltage locations only, but more recently they were put in high voltage locations also. Great care must be used to route them out of the DC cocoon. They exit through a DC lead and then are eased out of the DC lead insulation at a scarf joint, (a scarf joint is where a connection is made and the insulation is “V” out. The “V” part is manually taped backed in once the connection is made). Fiber Optic probes can read several degrees Celsius difference, even in the same location due to bonding issues. Improvements in the quality of the fiber optic probes have reduced the failure rate during manufacture. They can be monitored periodically or as part of a transformer monitoring system.

Coolers:

FIGURE 17 BANK OF RADIATORS WITH FANS (COOLERS)

The coolers are essentially the same as that of AC transformers.

Page 47

Guidelines for Life Extension of Existing HVDC Systems

While radiator type coolers are still used for converter transformers, the alternative is a fin type cooler with a pump for each individual cooler. A converter transformer will normally have three to five coolers with one being a spare. The spare or redundant may be interlocked to prevent it from operating until one of the main units is taken out of service, in order to prevent excessive flow rates and the possibility of static charge build up or it may be allowed for overload ratings. During start up in extremely cold weather one or more pumps (usually two) may be equipped with a second set of overloads to allow them to start with cold oil without tripping.

Page 48

Guidelines for Life Extension of Existing HVDC Systems

Auxiliary Devices:

FIGURE 18 CONSERVATOR “DRYCOL” BREATHER (WHITE)

There are some auxiliary devices, which will be discussed here which can improve the reliability of a converter transformer (and similarly an AC transformer). One of these devices is an over pressure device called a Pressure Relief Device (PRD). There is usually one or more on the main transformer tank depending on oil volume and one on the diverter. However, these may not necessarily prevent a tank bulge or rupture as is commonly believed. For example, a test was done on a 10MVA AC transformer with the lid removed and a 10kA fault was placed at the bottom of the tank for a long time, resulting in a tank rupture. The IEEE has prepared a guide to make transformers more rupture resistant. IEEE PC57.156/D4.1, September 2015 - IEEE Draft Guide for Tank Rupture Mitigation of Liquid- Immersed Power Transformers and Reactors

For free breathing transformers, a maintenance-free breather (e.g. “Drycol”) can be added to the conservator to allow the unit to breath and yet maintain very dry oil. It acts like an air conditioner but uses a special resistive element. This element cools if the voltage is in one direction and heats if the polarity is reversed. Other devices use silica gel, which is heated automatically to revitalize the gel. Still others use a filter type element, which shows when it is spent and must be changed.

Performance Issues:

Converter Transformers

HVDC converter transformers have generally performed fairly well. The HVDC converter transformers that are used in a typical HVDC system are usually all designed the same although the designs may vary because of different AC voltages in different converter stations, and manufactured at the same time. Thus, any flaw in the design or the manufacturing process would potentially be common throughout and can result in a large number of failures or repairs across the system. This can result in a large impact on the performance statistics. Past history has shown that the HVDC schemes, which have had large numbers of failures for which the cause is known (or believed to be known by the Owner/Operator) has been correlated to a flaw in the design or manufacturing process. Most manufacturers are now aware of

Page 49

Guidelines for Life Extension of Existing HVDC Systems these causes. In some instances the failures may be due to more than one cause, thus making the analysis a lot more difficult.

DC Bushings:

One of these causes is failure of the DC bushing oil end. The bushing condenser foils, porcelain, and DC barriers, form a complex insulation structure, which can be subjected to excessive voltage stresses as the materials vary with temperature, AC and DC voltages. Finite element modeling (FEM), plus testing the DC bushing (complete with DC test barriers) at 115% of the transformer test levels have generally resolved this problem. In some cases older converter transformers do not have sufficient room for these 115% bushings and the lower voltage are used as replacements or the old bushings refurbished with a new condenser.

The CIGRE DC bushing guide recommends that the DC bushing be tested at a level 15% higher than the level it would be tested in the converter transformer, and also tested according to IEC/IEEE 65700- 19-03-2014 - IEC/IEEE International Standard -- Bushings for DC application.

The DC bushings have been a source of problems in the past. One manufacturer has eliminated the oil end porcelain such that the main transformer oil is used to insulate the DC bushing. This however requires a very high main transformer conservator tank for vertical DC replacement bushings as the conservator must be higher than the head of large high voltage DC bushings. This is less of a problem for the through-the-wall variety.

For outdoor bushings certain criteria may be followed to reduce the probability of flashover due to pollution. Manitoba Hydro has established a standard alternating large/small shed profile with a 0.70 ratio between them. There may have been other standards developed by other Suppliers that are not included here. Clearance to the next large shed below is 70 mm for an anti-fog type shed. Shed profile is an important factor at higher DC voltages. An “outdoor end creep to flash distance ratio” of not less than 3.44 is acceptable. A specific creep of not less than 35 mm/kV is acceptable depending on the pollution level but 43 mm/kV has been specified as well. However, many converter transformer manufacturers choose to go with composite bushings and silicone rubber sheds as they have superior seismic and pollution performance. For indoor DC bushings in the valve hall the silicone rubber sheds may be eliminated and only the epoxy core used.

Other attempts to reduce the probability of flashover have included coating with grease and the addition of “Booster Sheds”. A Booster Shed is made of a glastic material or silicon rubber material and put around several sheds down a porcelain bushing creating vastly different shed profile as well as increased creepage.

Tap Changer Leads:

A second “contributing” cause has been the failure in the design to cancel the currents in the tap changer leads. This results in the load current increasing multiplying for each increase in tap position in either direction from the neutral point. Thus with a rectifier transformer operating at its normal position of say, some 10 000 amperes can be inducing circulating currents in the core. This was a problem for at least two HVDC Systems, which resulted in gassing and reduced life. However, other factors also contributed to the actual failure. These other factors were reduced cooling in the winding ends due to the requirement for additional insulation for voltage transients, and increased losses from flux lines changing direction due to the large amount of “DC cocoon” insulation. Pairing the tap changer leads for current cancellation can resolve this problem; however, each manufacturer has taken a different approach. Some use different size wires at the ends of the winding to cut down on eddy losses, and also bringing

Page 50

Guidelines for Life Extension of Existing HVDC Systems out the lead in the center of the winding and using a capacitance technique to distribute the transients rather than increasing the insulation.

A converter transformer tap changer may have a larger number of tap positions depending on voltage fluctuations and the length of the transmission line. In addition, for weak AC systems the tap changer may automatically go to the tap position, which inserts the maximum impedance to the converter transformer to limit inrush currents.

Harmonics:

Another HVDC System experienced several failures and all of the transformers had to be rewound. The cause was the six pulse harmonics not canceling each other out, thus causing additional heating in the center of the windings and other places.

Harmonics cannot be tested for in the factory; traditionally a factor has been added to the full load current during the load run test. Experience has shown that it is not possible to adequately test for the hot spot in the factory. Thus factors such as 1.3 to 1.5 times the gradient over top oil cannot be used to estimate the hot spot properly. It must be calculated accurately by the manufacturer.

Oil:

One of the more controversial areas with respect to transformer failures is the presence of mercaptains in the oil causing copper sulfide deposits on the windings. In one instance the supplier and the utility agreed that the copper sulfide, along with voltage transients and overheating due to reduced oil flow caused the transformer failures. Thus, some HVDC Systems that have found mercaptains in the oil added an inhibitor and have had no failures to date. In another instance involving mercaptains with voltage transients, the supplier and utility have not been able to agree on the cause of the failures. The remaining failures are “one-of-a-kind” failures where no trend exists

Many conductors are coated with insulating “enamel” which prevents the oil from attacking the copper but the oil should be tested for the presence of “mercaptains” before being used and also for topping up a transformer in the field. Note that the “ASTM D 34 87 - 00 Standard Specification for Mineral Insulating Oil Used in Electrical Apparatus” does not does not include a test for “mercaptains” so it must be done separately. This test method is available from CIGRE or IEC.

The remaining failures are “one-of-a-kind” failures where no trend exists. Moisture in the oil can “age” the converter transformer oil quicker and should be kept as low as possible. High levels of 60 ppm have been known to cause failures. Because of the very large amounts of cellulose insulation in the converter transformer, it is possible to measure the moisture at low temperatures and when the unit heats up the cellulose insulation driving the moisture into the oil and causing a failure.

Core and Coils:

The DC insulation “cocoon” can cause cooling problems due to restricted oil flow and increased losses due to electrical fluxes crossing the ends of the windings radially instead of axially. Harmonic fluxes have a higher impact in the middle of the windings. In a twelve pulse converter transformer (wye and delta windings on the valve-side) the six pulse harmonic fluxes do not cancel but are additive.

The tap windings are another source of concern. The tap leads must be paired such that the flux generated by them cancels each other out. This has been a problem in at least two DC transmission systems. Because of the large number of tap positions there are a large number of connections that can potentially cause gassing problems. Solutions that have been employed to reduce this risk are “double crimped connectors” with a hole visually ensuring proper conductor insertion. Inductive braising and

Page 51

Guidelines for Life Extension of Existing HVDC Systems qualifying the welder on each type of braises if the person has not done one in the last six months are also important considerations.

Therefore, a comprehensive design review is required for the hot spot as well as other critical areas. The manufacturer must demonstrate how the hot spot is determined in operation considering the effect of these harmonics, harmonic fluxes and cooling flow in the DC cocoon.

CIGRE has published several reports on conducting design reviews for converter transformers that can be referenced. IEEE and IEC may have similar guides but are not referenced here.

• TB 406 HVDC Converter Transformers - design review, test procedures, ageing evaluation and reliability in service. • TB 407 HVDC Converter Transformers - for conducting design reviews for HVDC converter transformers The pump bearings are a major cause for concern as when they fail they spew metal particles into the winding and core. Sleeve type bearings can be used instead, with a bearing wear monitor.

Control problems are usually with the fan or pump contactors, winding or top oil temperature set point, or the sensor themselves.

Technical Life Assessment:

Converter Transformers

The conditional assessment of a converter transformer is similar to that of an AC power transformer. The following sections below indicate where there may be a difference or where additional care is required due to the differences.

DC Bushings

Periodic testing of the capacitance and power factor should be done. Other tests would be similar to AC bushings. It is possible to take oil samples, but is not recommended by the bushing manufacturers in general. To take the sample from the bushing head risks contamination of the oil and may require the replacement of a nitrogen gas cap. This should only be done if absolutely required, and with instructions supplied by the manufacturer. In some instances a valve is provided at the bushing flange, which makes this a lower risk operation.

Tap Changers

The tap changers condition is verified by the maintenance records kept by the utility. Fortunately the advanced technology available today makes it easier to monitor all of the operating characteristics of a tap changer in real time. Tap changer monitoring packages are available from some tap changer suppliers, or it may be part of an overall transformer monitoring system.

Number of operations and the measurement of the amount of wear on the arcing contacts are indicators of end of life or refurbishment.

The diverter conservator filling or removing of oil while in service is an indication of an oil seal or O-ring failure between the main tank and the diverter vessel.

Core and Windings

On line hydrogen monitoring can give a quick indication of any gassing concerns which would then be followed up with DGA sampling and analysis. Testing converter transformers for assessing the condition

Page 52

Guidelines for Life Extension of Existing HVDC Systems of the unit should follow the manufacturer’s instructions. Tests could include frequency response analysis, core vibration tests, Partial Discharge testing as well as the normal periodic dielectric tests. Note though, that the sequence of any dielectric tests in the field is important as it can take up to 24 hours for the DC charges to dissipate after a DC test. Thus any subsequent test could have an unintentional DC bias.

Should a converter transformer fail or be repaired in the factory this could provide an opportunity to assess the condition of the unit and extrapolate this information into the other transformers, as they are usually similar in design.

Coolers:

Visual inspection plus maintenance records can determine the condition of the coolers. An infrared inspection or other temperature monitoring equipment can be used to determine that the coolers are still functioning efficiently. Pressure testing can determine the integrity of the coolers for leaks.

Analysis:

The analytical tools used to do the analysis are the same as used for AC transformers but the analysis of the results is specific to converter transformers (e.g. AC field stresses the oil and DC field stresses paper cellulose based insulation). The analysis itself is complex and is compounded in that there are fewer units in-service to gain experience from and because all the converter transformers are designed and fabricated at the same time, design issues become significant. Thus it is very important to determine if a failure or high gassing rate is a quality issue likely affecting only one unit or if it is more systemic and potentially affects all units of a similar design.

The following criteria should be used to assist in the analysis process and are based mainly on experience and standard techniques:

1. The analysis will be applicable to most converter transformers but given the complexity there may be exceptions. 2. A hot spot may not produce any gassing or furans and can still fail without warning. This is based on a converter transformer that had a multi-gas analyzer installed and furan testing done and then failed without warning a few days later. 3. If there are larger amounts of paper involved in a hot spot, it will show up in a gassing and furan analysis. This is based on a unit where the winding shifted after a through fault and blocked some of the cooling ducts. 4. A transformer can produce a large amount of gas and if there is a hot lead connection even if it is covered in cellulose tape may not indicate any furans or out of ordinary CO/CO2 ratio. This is based on one unit which had a loose core plus a bad lead connection. It indicated all the gassing was in metal parts as the amount of cellulose was small. 5. A bad CO/CO2 ratio is normally associated with cellulose paper degradation but can also be indicative of oil degradation. This is based on two spare converter transformers which had never seen service but still had bad CO/CO2 ratio readings. 6. In a bladder or air cell unit which is sealed from the air, the reduction of oxygen is associated with cellulose paper degradation. An increase in oxygen may mean that there is an air leak. 7. Wherever possible more than one indicator should be used to confirm that there is a problem. 8. Trends are more important than a single reading which could be faulty and should always be verified by another sampling. 9. Furan analysis, although developing, can still be a useful tool in the analysis process. If there are no furans the levels are low or the levels are very high.

Page 53

Guidelines for Life Extension of Existing HVDC Systems

10. While there is a high probability that transformer with high gassing rates will fail, it is not possible to predict when they will fail. 11. Accurate and an abundance of maintenance records will make the analysis process easier and produce better results. At a minimum annual DGA, furan analysis and oil quality should be performed. Dissolved Gas Analysis (DGA):

There are three methods that are commonly used for the DGA. They are Duval’s Triangle, Roger’s ratios and Doernenberg ratios. Many test instruments use Duval’s Triangle because of its simplicity. IEC 60599 - Mineral oil-impregnated electrical equipment in service - Guide to the interpretation of dissolved and free gases analysis uses the Duval Triangle Method. ANSI/IEEE C57.104-1991 - IEEE Guide for the Interpretation of Gases Generated in Oil-Immersed Transformers uses both the Roger and Doernenberg ratios. Some companies have used both methods when a transformer is gassing to see if one or the other gives a better idea of the problem.

Note: Converter transformers are very complex custom built devices and as such it is not possible to provide an accurate assessment of their condition by using the information provided by DGA. The information provided is primarily here to give a high level indication of a problem, know when to get a better assessment and provide a base knowledge. However because of the high cost of converter transformers to replace or refurbish/repair it is usually required by the Owner or senior management to get an expert opinion on replacement or repair. It is quite common for the original supplier to perform an assessment and then for this to be reviewed by a knowledgeable expert.

It is recognized that each supplier, each utility and each OEM will use a slightly different methodology to analyze the converter transformer. While in most instances the results will be similar in certain cases the results may be completely different in which case an Independent Expert second option is beneficial because of the high cost of converter transformers.

This section is not intended to be a document that determines the cause of a converter transformer failure or gassing as there are a number of guides to help with that process. It is more related to an understanding of what the analysis means especially since a converter transformer and its failure modes may be different from AC transformers.

One important difference is that AC voltage stresses the oil and DC voltage stresses the paper and pressboard. That is the bulk of the DC voltage is contained by the paper or cellulose insulation. Once it breaks down, the DC voltage can span at 500 kV, 16 feet or more of oil to ground.

Oil Sampling intervals vary but most common is an annual sample. It should include the dissolved gasses as listed below, oil quality and furans. While there is no qualitative guide on the various levels of furans, the IEEE, to name one, is discussing the need for a relative guide to confirm the CO2/CO ratio, e.g. if there are no 2 Furfuraldehyde (2-FAL) the ratio is due to deterioration of the oil and not the cellulose, so an important confirmation. Also trending is more important than single readings and cannot be stressed enough, it could be a bad reading and should be resampled. The 2-FAL concentration up to 100 ppb is considered normal for most transformers and the range of 400 ppb excessive loss of insulation life. Different types of cellulose generate different levels of furans so the furan concentrations are best used as a confirmation of the CO2/CO ratios.

The sampling is done according to ASTM D 3613. The goal is to collect a representative sample, avoid contamination and preserve it until it is analyzed. It is normally taken from the bottom of the tank and the

Page 54

Guidelines for Life Extension of Existing HVDC Systems sampling pipe flushed so a “good” sample is achieved. The unit must be positive pressurized if it has a rubber bladder or air cell to avoid sucking in outside air.

The sample analysis is very specialized and it must be done by a qualified laboratory so the results can be relied upon. The laboratory techniques are described in ASTM D 3612. Two laboratories will give slightly different results, even using the same methods, and bad readings are not uncommon even with a qualified laboratory.

Carbon monoxide and carbon dioxide (CO and CO2): These show the deterioration of the cellulose paper even at operating temperatures but if they or the ratio is outside the normal range accelerated ageing is likely occurring. However this should be confirmed with furan analysis as deterioration of the oil with time can also yield similar results. Note: The authors are aware that there were two spare transformers which never saw service in 30 years but still showed similar results which are rather difficult to explain.

Hydrogen H2 - This is indicative of low intensity electrical discharges such as Partial Discharges (PD). There will also be some Methane and fewer quantities of other gasses (H2/CH4 >10).

Methane CH4 - Methane is the first gas to show up as the temperature of the overheating increases (>150 oC). It is not unusual to have a transformer with some methane which can have a long life.

Ethane C2H6 – Ethane is the next gas after Methane to show up as the temperature of the overheating increases (>250 oC).

Ethylene C2H4 – Ethylene is the next gas after Ethane to show up as the temperature of the overheating increases (>350 oC).

Acetylene C2H2 – This indicates a high intensity electrical discharge or power arcing. It indicates a very high temperature of over 700 oC. The presence of the gas is usually a major concern. If the fault is developing in the cellulose the transformer will not last long, even if it is in metal the rapid deterioration of the oil is also a major concern. While the standard uses 35 ppm concentration of acetylene as normal, some utilities use an even lower number which may depend on the design of their converter transformers. It is also usually accompanied by higher levels of Hydrogen.

From the above it can be seen that the types of gases show the overheating that is occurring and if cellulose is involved, it shows the seriousness of the problem. However the analysis is not quite that simple and the ratio of the gases is important to determine exactly what is going on.

In sealed transformers with an air cell or rubber bladder the Oxygen and Nitrogen is also measured. The Oxygen is consumed over time but a rapid increase is likely to be a leak in the seal of the unit.

Degree of Polymerization:

One of the most dependable methods of assessing insulation life is Degree of Polymerization (DP). The cellulose molecule is made up of a long chain of glucose rings. DP is the average number of rings in a molecule. As the paper breaks down the number of rings is reduced and the chain is shorter. When it is new it has a DP of about 1200 and when it gets to 200 there is no more usable life. Very little loss of life has a DP equal to 800, moderate life loss is 500 and end of life is 200.

The problem with DP is the leads are not representative of the windings or the hot spot. So the only way is when a transformer fails, one can get an idea of the remaining converter transformer as they will all age at the same rate. One has to be careful though as the AC voltage rating may be different at each converter station and the units may be star and delta in separate transformer.

Page 55

Risk Matrix

Hydroge Ethylen Ethan Carbon Carbon Methane Acetylene n e e Monoxide Dioxide

C2H2 C2H4 C2H6 TDCG H2 ppm CH4 ppm CO ppm CO2 ppm CO2/CO O2/N 2 ppm ppm ppm ppm

Normal 100 120 35/5 50 65 350 2500 720 5 to 9 30%

Low 700 400 50 100 100 570 4000 1920

Medium 1800 1000 80 200 150 1400 10000 4630

High > 1800 > 1000 > 80 > 200 > 150 > 1400 > 10000 > 4630

Water D1533 Interfacial Neut # Color Visual Dielectric Specific ppm Tension D971 D974 mg D1500 Inspection D877 in kV Gravity mN/m KOH/g D816* Relative

Recommende d < 10 32 (min18) 0.1 0.5 New > 30 0.91 New

Page 56

Replacement/Refurbishment:

Converter Transformers

The replacement or refurbishment of a converter transformer is similar to that of an AC power transformer. The sections below indicate where there may be a difference or where additional care is required due to the differences.

DC Bushings

Replacement can only take place with the exact make, type and model as DC bushings are not interchangeable at the higher voltages (which require the DC barriers). If there are no DC barriers and the oil end does not form part of the insulation structure they can be interchanged. DC bushings are very expensive compared to AC bushings, and so it may be more economical to replace the core and or seals if required. This would have to be done by the OEM as they would be the only ones that would normally have a DC “test” barrier to adequately test the bushing after refurbishment. The current revision to the IEEE converter transformer standard requires the manufacturer to provide sufficient information so that replacement bushings can be obtained if the OEM is not available for any reason. For lower voltage bushings that do not have a DC barrier there may be other suppliers. In some cases the utility or contractor may be capable of replacing seals, processing, and testing.

Tap Changers

Parts for older tap changers may be difficult to find and/or costly to procure. There are companies that will reverse engineer a part or parts and some utilities have done the same. In other instances the entire tap changer has been replaced with the assistance of the transformer manufacturer. Retrofit to vacuum type diverter switches and tap changer is offered by many suppliers and may also be considered if entire replacement is warranted. This can reduce maintenance cost and extend maintenance intervals.

Core and Windings

If only the windings or the core have to be replaced, it is usually economical to replace either of them. However, if both have to be replaced, it is most likely not economical. A business case analysis should be done in any instance, taking into consideration any improvements in the design and their benefits. The remaining life of all the other components such as the bushings, tap changer, and others, must also be included in the assessment.

Depending on internal access, a “bad” tap changer lead connection or other connection may be able to be repaired in the field. These should be done under the manufacturer’s supervision or by them. Loose core laminations can be somewhat fixed and loose blocking can be repined, again depending on access. For loose core laminations, on-line degassing can allow the converter transformer to stay in-service until a scheduled shutdown. This was successfully done by one utility for several years. An on-line multi-gas analyzer (5 or more gases) can be added during the degassing, or to transformer of concern, to provide instant notification of any changes to the gassing pattern.

Coolers

Typically the refurbishment or replacement is done during the course of normal maintenance. Technicians replace parts as needed over the years. When spare parts are no longer available or the rating of the cooler is insufficient to meet the requirements of the circuit, the utility may consider the replacement of the entire cooler.

Auxiliary Devices

These devices are for the most part replaced if defective.

Page 57

Guidelines for Life Extension of Existing HVDC Systems

Refurbishment:

A refurbishment may be required after a number of years, say 25 to 30 years, where the units are still believed to be in good condition due to testing and DGA results. This may include some or all of the following:

• Replacement of the AC bushings, DC bushings, AC arrestors. The arrestors may be gapped silicon carbide and newer zinc oxide will provide much better protection. • Replacement of the OTI, WTI, PRD Bucholtz and rewiring of the control cabinet. Removing of the 600 V or other higher voltage from the control cabinet for safety reasons. • Cleaning of the coolers or replacement of the coolers, fans and pumps • Refurbishing the tap-changer, selector switch, diverter, gasket and O’rings • Internal inspection, replacing blocking and tap lead supports • Replacing the wiring on the converter transformer and to the control building • Adding additional monitoring such as on line DGA or a transformer monitoring system.

Trouble Modes Converter Transformers

The trouble modes are:

• Excessive Gassing • Test failure • Heat run failed • Tests passed but fails in field DC Bushings

The trouble modes would be similar to an AC bushing. The trouble modes are:

• Leaks • Damaged sheds • DGA - Gassing • Bad test data C1,C2 or PF • Poor Connections • Bad PF tap grounding • Defective oil level gage Tap Changers

Tap changers are very complex devices made up of many components and subcomponents, which all must work together properly. The manufacturers of the tap changers provide inspection and maintenance manuals with recommended practices to keep the device working reliably.

The trouble modes are:

• Leaks • Timing • Excessive motor currents • Overheated connections • Selector switch springs losing pressure • Worn parts • Linkage failures

Page 58

Guidelines for Life Extension of Existing HVDC Systems

• Contact damage • Diverter vessel equalizing valve Core and Windings

The manufacturers of the converter transformers provide inspection and maintenance manuals with recommended practices to keep the unit working reliably.

The trouble modes are:

• Loose core laminations • Oil contamination • Shorted laminations • Blocking moved or missing • Tap-changer lead support structure in bad shape • Moisture contamination • Gassing Coolers

Coolers once installed are relatively trouble free for many years with regular maintenance. Items such as regular cleaning of the coolers, inspection for leaks and bearing monitoring are required.

• Contamination • Bearing failure • Corrosion • Leaks • Control problems • Fan failures • Pump wired backwards • Defective Valves Auxiliary Devices

These devices are relatively simple and can easily be fixed or the entire part must be replaced.

The trouble modes are:

• Defective breather • Plugged breather • Defective PRD • Closed conservator valve causing the PDR to operate • Moisture in the oil Converter Transformers – Maintenance:

The maintenance of a converter transformer is similar to that of an AC power transformer. The following sections below indicate where there may be a difference or where additional care is required due to the differences.

Maintenance includes oil sampling, checking for leaks, and infrared inspections.

Page 59

Guidelines for Life Extension of Existing HVDC Systems

DC Bushings

DC bushings require visual inspections noting the condition of the shed surface area for signs of excess heating, tracking, dirt build-up, damaged sheds, or contamination. The most common problem is external contamination due to climate and environmental conditions. Periodic cleaning may be required. Oil leakage or weeping is another issue. The oil gauge may indicate too high or too low indicating a seal failure internal to the bushing. Measurement of C1, C2 and Tan delta with trending and comparing to factory results is required on a regular basis.

Tap Changers

Due to the number of high operations found in the converter transformer tap changers, regular inspection and overhaul is required to keep them operating reliably. Because of the complexity, highly trained staff and/or the use of a manufacturer’s representative are required.

The drive linkages must be inspected for wear, and in some cases the safety guards or shields must be removed to do this inspection. A recommended practice is to move the tap changer periodically through the entire tap range in both directions checking for binding or excessive motor current. Remove the diverter, inspect it, and clean it according to the manufacturer’s recommendations. Replace with new oil. Ensure that the diverter vessel and diverter are out of oil for only a minimum amount of time, as contacts will oxidize and cause overheating problems.

On-line filtering of the oil has been implemented on some oil filled diverters and generally if more than 15 000 operations per year are required.

Some newer diverters have vacuum bottles and the authors are aware of at least one older HVDC that did have vacuum bottle diverters as well. In this case it is necessary to measure the contact wear every 4 or 5 years and trend it also to determine remaining life.

Core and Windings

There is very little that can be done in the field for problems with the core and windings.

Coolers

Coolers are relatively simple devices. The coolers manufacturer’s recommendations and guidelines for maintenance and maintenance schedule should be followed. Maintenance is normally a visual inspection for leaks and contamination. Ultrasonic monitoring can detect bearing failure, but a highly trained technician is required and it must be done monthly as the bearing can fail rapidly unless the bearing manufacturer can advise otherwise. On-line monitoring is also available. Fan motor and contactors should also be checked periodically.

In some cases a valve may show that it is in the open state, whereas the valve is partially closed or a pump may be wired/running backwards. A periodic infrared inspection twice per year will pick this up as a cooler will be running cooler than the rest.

Auxiliary Devices

These devices are for the most part replaced if defective, although some repairs can be done to the Drycol control cabinet. The PRD can be checked for proper operation at the set point. The Drycol will alarm if defective.

Page 60

Guidelines for Life Extension of Existing HVDC Systems

3.4. DC Control and Protection

FIGURE 19

Description:

HVDC Control and Protection systems, like any power electronics based converter control and protection, can be differentiated from AC system protection. The DC system protection is closely integrated together with the control system and is partially implemented in control of the converters. Traditional AC protection systems exclusively trip circuit breakers as a means of protection. For DC systems, protective action can include circuit breaker trips as part of a protective sequence. In addition, control actions such as changing the control firing angle from rectifier to inverter operating region, which would reduce the system DC current to zero or blocking the firing pulses, and thereby stopping the conversion process, can be components of a protective control sequence. For early HVDC projects, the control and protection system was analogue based and highly customized from project to project, even within the same vendor. More recently the HVDC controls have become digital based and include a large number of integrated systems or packages within the same basic control architecture. HVDC control systems remain highly customized packages still based on the requirements of the particular project. The useful lifetime, the ability of an owner or HVDC supplier to continue supporting analogue or digital control systems is a pertinent question being asked by HVDC System Owners. Any control system replacement, whether analogue or digital based, is generally necessary for one of two reasons. Either there is a lack of spare parts to replace the failed, failing components, or there is a lack of staff expertise (either in-house or external) that can maintain, repair or reprogram the existing control systems, obsolescence, lack of compatibility for hardware or software may cause a replacement.

CIGRE statistics from 1993 to 2008 cite 8% and 2009 to 2010 cite 8.9% of the HVDC outages are due to Control and Protection events.

HVDC Analog Electronics:

Page 61

Guidelines for Life Extension of Existing HVDC Systems

FIGURE 20 ANALOG VG CONTROLS (EDDY COUNTY)

It is technically possible to repair and refurbish existing analogue control systems. The long-term challenge of this approach is obtaining and retaining the knowledge base, components, staffing, and funding to undertake such a program. Replacement of analogue control systems with digital technology has been considered by many utilities. Others continue to maintain analogue control systems as long as components and staff expertise can be obtained. An average service life of twenty-five years has been assigned to analogue electronic systems assuming routine maintenance and availability of spare parts. Analog control systems can be redundant. Analog controls, by their nature, consist of many discrete components, each of which can age and deteriorate, thus changing the overall response of the control system and its failure performance.

HVDC Digital Electronics:

FIGURE 21 DIGITAL CONTROLS (HIGH GATE)

Rapid improvements in technology, design, hardware, and software as well as changes in system requirements tend to make the service life of digital electronics relatively short. An average service life of fifteen years has been assigned to digital electronic systems assuming routine maintenance and availability of spare parts. Digital control and protection systems are far more compact than their

Page 62

Guidelines for Life Extension of Existing HVDC Systems analogue system counterparts are. They are often installed with redundant systems. With process self- monitoring capability and redundancy, a single failure mode does not normally result in a transmission system outage.

Description of HVDC Control and Protection:

HVDC Control and Protection systems cover a broad organization of control systems or subsystems. HVDC systems are constructed using various configurations of several elements, namely valves, valve groups, poles, and bipoles. The following section itemizes the composition of typical HVDC systems and highlights the control and protection sub-systems commonly found in most HVDC systems. HVDC controls tend to be hierarchical in nature. In order for a HVDC system to operate, controls specific for each level are required. Higher level control system can serve one or more of the lower level control systems, but a lower level system is dependent upon the higher level systems remaining in operation. For example a bipole control level can provide functions for two pole controls or there can be a Bipole controller for each pole. A pole control system can operate or interface with two (2) or more valve group or bridge controls. Each valve group control system can function with one or two 6-pulse bridges. The valve group control interfaces to a valve base electronics (VBE) control system, generally with one firing pulse command for each Thyristor Valve. Within the VBE the individual requirements for each individual Thyristor are developed. Depending on the size and configuration of the system, individual controllers at each level may exist or all of the control levels may be implemented within a single control system. Figure 16 illustrates the HVDC control hierarchy.

Master VG 1 VBE 1 Pole Or Bipole Control

Or System VG 2 VBE 2 Control

To Other Pole Control Input from System Control Centre

HVDC Control and Protection Hierachy

FIGURE 22 HVDC CONTROL HEIRARCHY STRUCTURE

Graetz 6-Pulse Bridge

The most common configuration for 3-phase HVDC power system, utilizes a 3-phase leg or 6-pulse Graetz bridge configuration as shown in figure 17

Page 63

Guidelines for Life Extension of Existing HVDC Systems

+

Valve

-

The 6 Pulse Graetz bridge is the Basic HVDC Module Bridge

FIGURE 23 GRAETZ BRIDGE IS THE BASIC HVDC MODULE BRIDGE

Valves

The power electronic valves consist of many devices. Thyristors for example placed in series to increase voltage withstand capability, or placed in parallel to increase current capability. A valve is illustrated in Figure 23. Each individual Thyristor within a valve require firing or gating signals as inputs. They have some form of individualized protection. A Break over Overvoltage gating Diode (BOD) for example, or a more sophisticated form of protection firing, a power supply for the electronics, and reporting or diagnostics output signals to indicate the state of health for each specific thyristor level that collectively make up the valve. Each thyristor will have a parallel resistor-capacitor (RC) circuit. This circuit, commonly called the RC snubber circuit, limits dv/dt transients that can damage the Thyristor. The valve assembly includes saturable reactors for limiting di/dt, which otherwise could damage the thyristors. It also may include some form of intertier voltage grading capacitors. A surge arrestor is also placed in parallel with the overall Thyristor Valve assembly for lightning and switching transients.

Thyristor Trigger or Electronics Gate is electrical or optical

VBE interface is optical

‘N’ thyristor levels

Valve

FIGURE 24 SERIES CONECCTED THYRISTOR VALVE

Valve Level Electronics

The power electronic Thyristor Valves receive an isolated (generally optical) signal from the valve base electronics (VBE). They require both a source of supply energy and device level electronics for shaping and producing the gating or firing pulse that is required for each individual valve level. Some HVDC projects use direct optically fired Thyristor devices. Each valve level provides some feedback diagnostic information. Examples of these indications include a sufficient voltage has developed across the device,

Page 64

Guidelines for Life Extension of Existing HVDC Systems indicating that it is not short circuit, and an indication the device is gated by a level protection device (could be BOD operated or a more sophisticated form of protective firing). This reporting information is transmitted through an isolation method (generally optical) to the VBE for further processing.

Valve Base Electronics (VBE)

The VBE control system accepts firing (gating) commands from a higher level Valve Group or Bridge control, which generates and distributes the required number of valve level gating signals for each valve. The VBE receives the report information from each valve level and provides a level of valve protection. Protective control action (e.g. refire all of the Thyristors within the entire valve or x levels have individually protectively fired) or protective trip action, (loss of device level redundancy) are usually implemented in the VBE. The VBE can be implemented in analogue electronics or digital programmable electronics but most modern ones are digital. The VBE system may also include the valve fault monitoring system. The VBE will usually also record the operations of valve surge arrestor operations, except for the Siemens design. This system provides indication of which valve levels are faulty or healthy, and generally contains an interface to the system's human machine interface (HMI), Sequence of Events Recorder (SER) and/or the station's alarm system.

FIGURE 25 VALVE BASE ELECTRONICS (BLACKWATER)

Valve Group Converter Control

Valve groups can be configured as either 6 or 12-pulse configurations. Early HVDC systems used 6- pulse configurations. Today, it is typical to have 12-pulse configurations. Figure 26 shows the valve group configurations. Valve group converter control functions determine the timing of the firing signal for each of the six (6) or twelve (12) valves within the Valve Group (VG). Typically, the timing of the firing signal is determined based on the measurements of the valve voltage waveform and/or valve current waveform, together with the phase locked oscillators. Other methods for generating fire commands such as a bias cosine delay angle determinator have been implemented. In some 12-pulse systems the 6- pulse star and 6-pulse delta arrangements will be referred to as Bridges within the VG. VG control requires the fastest control response time. The valve group converter control will provide firing signals to the VBE. VG Start/Stop sequence controls for placing the VG in and out of service as well as VG

Page 65

Guidelines for Life Extension of Existing HVDC Systems protection signals are included in this level. Interfaces to VG cooling system and switchgear associated with the VG will be included in this control system.

Valve Group (VG) r 6 Pulse VG is identical to a Bridge (either star or delta)

12 Pulse VG is a 6 Pulse Star VG In series with a 6 r Pulse Delta VG

Valve Group

FIGURE 26 VALVE GROUP CONFIGURATION

Pole Control

Pole control and protection system is a higher level control system than the VG level. Separation may be difficult to identify, depending on the overall configuration of the system, the VG, pole, and bipole control systems. One example of a system with difficult control separation would be a system with only one VG in the pole. For systems with more than one VG per pole, connected either in series or in parallel, control separation is essential. This ensures that the system can be operated at partial ratings during the variety of different types of pole or VG equipment outages. Possible pole configurations are shown in figure 21.

A pole can consist of Series, Parallel or a single VG

Pole

FIGURE 27 POLE CONFIGURATIONS

Pole control accepts the DC current order signal from the higher level master or bipole control system. Pole control contains the necessary control loops to determine the firing angle for a given set of system conditions. It includes the protective limits, which may be applied based on issues, which affect the entire pole. An example of a pole sequence may be DC line fault or line differential protection. DC voltage control or limits are applied in pole control. High level tap changer control may be implemented within

Page 66

Guidelines for Life Extension of Existing HVDC Systems pole control; however the actual tap changer control will usually be included within the VG level. The firing angle is the main signal that is passed from pole control to the VG control. HVDC measurement of pole quantities such as DC voltage and DC current is implemented in pole control. The interface to the DC switchyard for the pole switchgear and equipment like DC filter, high voltage DC line disconnects, and DC electrode line disconnects are located here.

Pole protection functions differ from VG protections in the scope and zones of protection covered. For example, DC line protection or pole differential protection is a pole protection function, while VG control will contain VG related protection like commutation failure protection. DC protection may provide a control action, or initiate the operation of circuit breakers, or a combination of both, depending on the circumstance.

Master Station or Bipole Controls

Bipole control functions are generally the highest level of controls. Master station or bipole controls serve as the interface for parameters necessary for the operation of the overall system. These may include one or more DC bipoles, particularly when parallel or multi-terminal operation is utilized. A bipole configuration is shown in Figure 28. Input parameters are generally provided by the System Control Center (SCC) Operators or Independent System Operator (ISO) over Supervisory Control and Data Acquisition (SCADA) systems. They include DC power flow set points, power flow direction, and AC voltage set points as in the case of a SVC function or synchronous condenser function implemented in the HVDC system. In addition, any special AC system control functions, sometimes referred to as RAS (Remedial Action Schemes) would be included at this level. These control signals are designed to assist and preserve the AC system during system disturbances. Features such as power swing damping, frequency control, reactive power control, AC system undervoltage protection, reduction of DC transmission levels or selective tripping may be included in a RAS. Common bipole control functions include calculation of DC current for a particular power setting, the current control loop, di/dt control and inputs with any special AC system controllers, which may be required. The limited or final current order signal is transferred to pole control. A transient overload or continuous overload may also be implemented especially for loss of a pole.

+ +

Pole 1

Pole 2

_ _ Bipole

FIGURE 28 BIPOLE CONFIGURATION

Page 67

Guidelines for Life Extension of Existing HVDC Systems

Special AC System Control Functions

A HVDC system can include additional control functions that integrate together with the pole (or bipole) controls. Examples of this category of controls include such items as AC system damping controls or AC system undervoltage controls. This controller may be physically separated from the bipole or pole controls or the control function may be integrated within them.

AC Switchyard Protection System

FIGURE 29 AC PROTECTION PANELS (HIGH GATE)

In addition to the protection functions provided in the HVDC control and protection system, a more classical AC protection system is also implemented. This distinct protection system is very similar to AC station protection based on voltage and currents, and will result in operation of an AC circuit breaker for clearing of any faults. This protection is generally physically separated from but coordinated with the HVDC control and protection functions. It serves as a primary means of avoiding equipment damage for the AC apparatus associated with a HVDC installation. Typical examples of protection would include transformer protection, for overcurrent and differential currents, harmonic filter protection with overcurrent, differential current, and capacitor unbalance protection.

Additional Control Systems associated with HVDC converter stations:

Digital Fault Recorders (DFR) and Sequence of Events Recorders (SER)

A HVDC plant tends to have additional fault diagnostic and recording systems. In older plants this equipment was stand-alone and used separate digital/analogue (D/A) inputs from the control systems. As integrated digital systems have developed these recording systems are often integrated into the overall control system, with the inputs derived using network connections within the control system.

HMI (Human Machine Interface) or Operator Controls

Page 68

Guidelines for Life Extension of Existing HVDC Systems

FIGURE 30 HMI COMPUTER SCREENS (SQUARE BUTTE)

The operator control interface can consist of control panel interface with discrete interface contacts, switches, indicating lamps for earlier systems. Typical digital based systems will operate the plant on personal computer (PC) based software controller connected to the control via the programmable serial digital interface.

Supervisory Control and Data Acquisition (SCADA) System

FIGURE 31 SCADA AND RTU PANELS (BLACKWATER)

These systems usually use a Remote Terminal Unit (RTU) to interface the HVDC station to one or more remotely located system control and dispatch centers. Typically the alarms displayed at the dispatch center are a subset or grouping of alarms that is available at the actual station fault recorder. Earlier versions of RTU hardware had a discrete (point by point) interface with the station and communicated to the dispatch center using a variety of different serial protocols, depending on the equipment at the dispatch center. More modern digital control systems can have the RTU function embedded in the programmable controller software of the overall plant control system.

Communication Systems

Telephone networks, microwave transmitter/receiver, power line carrier (PLC) and fiber optics communications systems are normally constructed of many replaceable parts, which are usually duplicated and can be serviced under normal maintenance programs. However, complete replacement of networks or systems may be required at intervals because of changes in operating requirements or obsolescence. It is becoming more common to have internet connectivity direct access to the control system. The security (NERC compliance in the USA) of this remote communication access to the station control access is an issue that needs to be addressed, but it should be noted this is an issue for

Page 69

Guidelines for Life Extension of Existing HVDC Systems traditional AC substations, generating plants as well as HVDC stations. Point-to-point HVDC systems characteristically use high speed communication systems to transfer control and protection signals between the rectifier and invertors stations. While the DC link operation can continue without communications between the stations, usually a degraded operation mode is required. Communications for back-to-back HVDC converter stations does not require the same level of high speed communications to continue operation, because both the rectifier and inverter are physically situated at the same site. Many newer HVDC schemes have optical fibers embedded in the overhead transmission line skywire and is called OPGW.

HVDC Control Description Summary

HVDC control and protection systems are at the heart of the flexibility offered by a HVDC solution. As a result, the associated control and protection systems can be significantly more complex and involved than some of its counter parts in the AC station. One additional significant difference is the requirement for redundancy, sometimes triple or duplicated high speed controller transfer schemes including the closed loop control functions. HVDC control and protection systems can have numerous levels or layers that are interdependent on each other. This interdependence needs to be taken into account when evaluation of the lifetime extension or controls replacement project is undertaken. Another consideration is the measuring devices such as the DCCT and VDR, whether they may need to be replaced as well. Failure or mis-operation of the HVDC control system can have serious negative consequences on expensive apparatus. For example, if the valve firing control mis-operates then damage to the Thyristor Valves may result. Standardization of design practices is not widely practiced, although discussion at the Institute of Electrical and Electronic Engineers (IEEE) and CIGRE committees is beginning to address this issue. See reference [8] for more information.

Performance Issues:

Trouble Modes

The control and protection system has an impact on virtually every piece of apparatus within the HVDC station. Identification of common aging and failure modes for the controls are similar to methods used for other equipment. Historical outage statistics and the duration of the outages resulting from the control and protection system are prime indicators of lifetime issues associated with the control system. Root Cause Analysis (RCA) will assist in identifying the root cause of the outage. It is important to note that control system failure may result in failure or premature failure of other equipment. For example, control failure limiting high valve fire angles, may result in premature failure of Thyristors, specifically if Thyristors are failing. A detailed analysis to find the correct control failure mode or other root cause will be required. A single Thyristor failure normally will not result in a forced outage as there are always redundant Thyristors. Therefore, collection of the system information during the event can occur a long time after the initial event, which may make analysis more difficult.

In addition, there may be circumstances when it is desirable to modify the existing system control in order to accommodate changes in the larger electrical system. The ability to implement these changes in either analogue hardware or a software/firmware program or not is an important factor to consider in the life extension analysis.

Page 70

Guidelines for Life Extension of Existing HVDC Systems

Technical Life Assessment:

FIGURE 32 DIGITAL CONTROL

Condition assessment is defined as that testing outside the scope of testing normally performed as part of scheduled maintenance.

Some forms of condition assessment are:

• Review of Digital Fault Recorder (DFR) / Sequence of Events Recorder (SER) Records based on system events that occur o For these system events or system disturbances, did the plant perform and respond as expected, or as it has responded to a previous event, provided there is an organized data management of previous faults? • If system simulation models exist, (PSCAD™, PSS/E, RTDS, PLSF), it can be very valuable to compare DFR recordings to similar events performed in the simulation of the system. Confidence of the system simulation model of the controls and protection is validated and can be utilized to determine response to system events or conditions, which have not yet occurred. Conversely differences in simulation to system response can be determined to identify weakness or limitations of the simulation models. • Monitoring of specific control functions with portable recorders/analysers. This allows in-depth analysis of hard-to-find or intermittent failures in control equipment. This method is more effective with analogue controls as the desired monitoring point(s) can be better accessed in analogue equipment. Refurbishment:

The decision to replace or refurbish HVDC Controls and protection requires a detailed station evaluation of many factors.

Factors, which impact control and protection refurbishment or replacement decisions:

• Availability of qualified personnel • Assessment of risk of plant outage or unavailability • Economic evaluation for plant outage or unavailability • Economic evaluation for the expected duration of the refurbishment or replacement program • Historical outage / failure performances • Availability of spare parts / components

Page 71

Guidelines for Life Extension of Existing HVDC Systems

• Technical resources to maintain and troubleshoot faults • Available documentation to maintain and trouble shoot • Requirements to add or modify control functions in order to support network growth and changes • New software may not be compatible with existing hardware and the piece hardware needs to be upgraded. If the piece of hardware is upgraded it may not interface well with other hardware resulting in replacement of more hardware than first anticipated. • Availability of programming tools and source code (among others) to maintain or modify controls • Evaluation of the systems and associated equipment integrated together with the control system. o A good example would be the Thyristor Valves, switchgear, cooling plant, breakers, and transformers etc. approaching end of life and require replacement, or refurbishment. Each individual control system should also be considered at the same period of time of those replacements. Replacement of Control and Protection systems:

These systems are complex and have complex integration with hierarchical systems either that are one level higher or one lower than the particular subsystem being evaluated. In a review of the system or subsystem the following questions can be pertinent:

• Definition of the scope of the upgrade or refurbishment project - Which systems require replacement? Can the replacement program be performed in phases or all at once? • Is there enough documentation and particularly interface documentation - This includes overall description of the control functions, system drawings, and definitions of the interface signals (timing, voltage current levels, Pulse width, message protocol, and so on.). • Is the documentation up-to-date and valid? Have all modifications made to the system included in the documentation? Updating documentation may be the first step. • What is the risk that once the replacement program has been defined and started that the defined boundaries will change? For example, one instance includes replacement of control cable that was not originally intended as part of the upgrade program, but once the work started it was determined that the existing cable was leaking oily like substance and had to be replaced as well. • Are there new features that are required and should be included in the controls replacements? Has the system configuration or operating requirements changed since the plant was originally installed? Has operation of the HVDC plant been limited due to “missing” functions? • Particularly for digital control systems, are programming tools, application source code, and a knowledge base required to perform program modification available for the existing systems? Once the software changes are made in the existing control systems, what method(s) will be utilized for offline testing and validation? Will the new components interface with the existing/remaining components and also can the old components be upgraded? • If the controls are not scheduled for replacement, but other system(s) are replaced, can the existing control system accommodate the new plant? For example, if the cooling plant is replaced, the new cooling plant must be specified to interface with the existing system. • What is the level of commitment from supplier of new equipment to provide training and maintenance support for plant staff? Will the supplier work with staff during commissioning to pass on knowledge base? • If the HVDC link is a two terminal owned by separate parties will the other partner be changing their controls? If not will the new controls and protections communicate with the old controls? What would the cost and potential reliability impact?

Page 72

Guidelines for Life Extension of Existing HVDC Systems

FIGURE 33 DIGITAL CONTROL SYSTEM (LAMAR)

Digital based control and protection systems are a common place reality. Experience since the initial installations of these digital control systems indicate clearly that at least one control system replacement can be expected in any system prior to the retirement of the entire system. Control systems have been assigned a fifteen to twenty-five year life expectancy. Thyristor valves, converter transformers, high voltage circuit breakers, have a life expectancy of thirty five to forty years, while the transmission line life expectancy may be in the range of seventy-five or more years. A relevant question that can be asked is; “What can be specified and designed into the protection and control system, at the project conception stages that would facilitate a control system replacement project in the future?”

The following possibilities are presented:

 Definition of control interface boundaries  HVDC control systems are very hierarchical or layered in nature as described in the overview of this section. If the interfaces between these control systems were standardized or at least documented clearly, it may be easier to facilitate a replacement in the future. IEC 68150 standard for digital protection relays provides a mitigation path for more traditional protection systems. Further work is required for the acceptance of a similar standard for HVDC control system.  Maintenance and staff training opportunity with HVDC system simulators  Teaching, training, and maintaining of new staff, as there will be staff turnovers  Control systems, particularly programmable digital controls with full redundancy do not generate a lot of outages. Opportunities for maintenance staff to become familiar with the controls systems, documentation for the controls, and the tools to program, upload, download, verify software are minimal in the operating system. The procurement of an operational set of programmable controls is desirable. The simulator environment with which the controls were factory acceptance tested would offer an ongoing opportunity for maintenance staff to remain familiar with control system without an outage of the operational plant. This would also allow staff to test or refresh any maintenance procedures prior to the maintenance outages, which are increasing shorter in duration. It would be possible to replay system disturbances, using the system simulator and control system. Finally a complete set of operational controls would be available to facilitate an emergency control replacement in event of a major control system failure.  Spare parts and programming tools  Procurement of short life assets such as PC’s and other programmable devices

Page 73

Guidelines for Life Extension of Existing HVDC Systems

 The programming tools used to generate, compile, and transfer the “control” program(s) has a similar life expectancy of the control system itself. The ability to procure complex digital spare parts and the tools needs to be accessed at the start of the start of the project. This issue is strongly related to the staff training item mentioned previously.  Password Security  Cyber Security: Equipment access only at the physical site and by dedicated computers that are normally locked up. Remote control only by a dedicated control and communications system.  Digital systems, by their nature, offer the opportunity for various levels of authorizations and security. Operators can view and make changes to certain category of items, while control changes require a different level of authorizations. It is important to ensure that the OEM transfers “all” password privileges to the owners at some point in the project. The onus is on the project owners to ensure the overall security protocol for system(s) remains effective and provides for life extension opportunities in the future.  Standardized Communications Protocols With new digital control systems, different protocols are used for serial communication between systems and subsystems by different suppliers (examples of protocols include PROFIBUS, MODBUS, and DNP). Standardizing these protocols would not only make maintenance easier, but also allow for a smoother updating of equipment.

Maintenance:

The maintenance of analogue and digital systems will be completely different. It is likely each converter station and HVDC link will do maintenance slightly differently.

Documentation

There is generally significantly more documentation for an HVDC installation than for an AC hydraulic generating station or AC terminal substation with similar power ratings. This is due to the complexity and customized nature of HVDC installations. The associated documentation systems will have a unique look and feel to them, depending on the OEM, which adds requirements for staff training and acceptance. Documentation will be found in following categories:

 Operational and Maintenance Manuals  System Drawings o Schematics o Cabling, wiring, I/O connection sheets o Setting Information: any setting or parameter that can be changed should be recorded.  System Block Diagrams; Program Flow Charts o Simulation Models: the overall control system structure or system block diagram is becoming documented more often in a simulation model (such as PSCAD).  Software / Firmware o Listings o Past revisions  Commissioning results o Original settings o Modifications o Problems during commissioning  Maintenance results

Page 74

Guidelines for Life Extension of Existing HVDC Systems

o Historical results o History of problems o History of setting changes  Specification and Technical Descriptions for the HVDC protection and control systems Weaknesses in Documentation

Major areas of concern during the assessment of a control system for life extension are the quality, availability, and accuracy of the documentation for the plant. Questions need to be addressed such as whether the drawings are up to date, including all the modifications performed to the plant. Increasingly with digital based control systems the documentation for the controls is the controller software. Often the documentation will consist of a high level system block diagram, followed by the detailed software blocks, making it difficult to follow or understand the control system implementation.

Control and protection system maintenance is defined as periodic scheduled inspection and testing of equipment to ensure trouble-free performance.

FIGURE 34 ANALOG CONTROL CENTER (MILES CITY)

Difference between Control and Protection functionality

Control systems are required in order to maintain operation and the control loops that are in service and are known to be working. For example, a HVDC plant, that is online, and producing the correct response (power flow, voltages, and more) indicates that a large percentage of the control system is functioning. Protection, on the other hand, does not respond until it is required. However digital metering functions in protection relays can prove that the input signals (A/D) circuits are functional. Digital control and protection systems generally have an advantage over their analogue counterparts. They include self- monitoring circuits such as watch-dog timers or input signal monitoring, which are able to detect failures within portions of the system.

Maintenance Records and Tracking Program

There are different maintenance issues for analogue and digital control systems not unlike the issues involved with analogue and digital protection systems.

Page 75

Guidelines for Life Extension of Existing HVDC Systems

Analog System

FIGURE 35 MIMIC BOARD (EDDY COUNTY)

Analog control performance can change with time as components age. Offline maintenance checks are required periodically to ensure that the parameters and performance remain within acceptable pass/fail limits. Timers and level detectors are of particular interest. Detailed procedures for performing this periodic maintenance are required. This documentation is specialized and customized for the particular installation. These procedures would include the following:

 Clearance and isolation procedures to take control equipment out of service safely without causing outages  Description of the test set-up and procedure: these procedures would reference the schematics required, terminal pins, waveforms, and monitoring points etc.  Instrumentation involved in the test, and test signal generator: specialized test instrumentation may have been supplied as a part of the initial installation.  Acceptable criteria for pass/fail for each part of the test/maintenance program  Restoration procedures, which describe what should be checked prior to return to service  Similar to the testing of protection relays, a database can be used for the detailed charting and tracking of the maintenance functions. When analogue circuitry is modified in the field it is relatively easy to re-commission the new circuit. It is common practice to check from the point at signal path input of the circuit, which was not modified, to a physical point beyond the signal path output. Large control modifications can be tested by using digital transient network analysers (TNA) methods.

Modifications to the printed circuit boards must the also be completed on the spare PCB’s.

Page 76

Guidelines for Life Extension of Existing HVDC Systems

Digital Systems

FIGURE 36 DIGITAL CONTROL CENTER (HIGH GATE)

Maintenance of digital based control systems still requires clearance, isolation and restoration procedures as well as a description of the actual test set-up procedure along with necessary test equipment. However, it can be modified to only include verification of input and output signal paths. The internal control algorithm (operating in software or firmware) which was tested during commissioning should not have changed or degraded in any way with time. Self-monitoring circuits will ensure that the control processor(s) are operational. It is common to have sanity checks of the input signals performed to identify failure of an input. These checks may include confirmation that the signals are within an acceptable range, comparison tests with redundant inputs, and the utilization of signals that require a valve for a valid zero level input. It is a relatively common feature to have the ability to monitor internal parameters within the control algorithm. Maintenance timeframes for digital controls can be considerably shorter than the analogue counterparts. This may be seen as a benefit, but familiarity with the equipment is sacrificed making it more difficult to troubleshoot when on-line problems and failure occur.

A maintenance tracking system and version control for digital controls is essential. This tracking system contains similar elements to the system required for analogue systems, although the number of functions that require maintenance checking will be significantly less.

 Listing configuration settings for each digital board o Jumpers, optional add-in hardware, firmware revision (to name a few) o It is possible that one type of controller boards is used in many different locations within the overall control structure. When a spare card is installed replacing a defective controller, it is important to ensure all components of the “spare” replacement match the board removed from service. This includes hardware settings, firmware, software, and all the programmable settings that are “user” definable within the software.  Test procedures for calibration of input/output signals o Procedures include input or output interface terminals, signal levels and calibration, schematic references, and instructions to observe the change or activate the output from within the software environment. This procedure may have been documented as part of the pre commissioning test plan.

Page 77

Guidelines for Life Extension of Existing HVDC Systems

 Documentation regarding software settings o This can include controller parameters, time constants, thresholds and/or alarm limits. Any setting that can be changed by the control owner should be documented and tracked. If a setting change is required, a rigid process of checks justification and approval should be followed as well as documentation and drawings updated.  Documentation of software version installed  Password identification for every security level  Procedures to install maintenance monitoring tools One of the issues regarding digital control is the uploading, replacement, or upgrade of the software and/or firmware of the controller. In a maintenance program it is not likely the software program will be modified, but is possible that software will have to be uploaded when installing a spare “replacement” controller. When new software is uploaded and installed, the entire control structure (algorithm) is replaced. From a maintenance point of view, a question regarding the depth and completeness of commissioning or re-commissioning tests required is a valid one. Checks and verification are required to ensure that the software version intended for installation is the software version that is actually installed and that it uploaded properly.

On-line Maintenance:

The following is a list of on-line maintenance possibilities:

 Controls redundancy o A redundant control system allows the possibility of testing the one control system while the plant remains in operation with the other control system controls. It is possible, in theory at least, to perform maintenance on the deactivated control system (with its control outputs blocked). The decision to perform maintenance with the system operating on the redundant path will depend on how serious the threat is to the HVDC transmission system is if something goes wrong. Human error will be the greatest concern. If the system load is low, online maintenance of the deactivated control and protection system is usually allowed. o Scheduled transfer of a redundant controls system from one system to the other while it is online will exercise the transfer scheme. The advantage here is the configuration of the system, system loading, and timing of the procedure is known. Maintenance staff is on site and present to assist in troubleshooting, should the transfer mal- function.  Built in test functions o For example, the exercise of plant control systems that are occasionally used but not continuously in service. Examples include activation of either thyristor switched capacitor (TSC) test (1 second) scheduled or during a special online testing period. Another example may be controlled (staffed) exercise of parallel operation for an HVDC plant. Testing of this nature risks system outages or disturbances should either the procedure or the equipment fail during the test. The advantage of course is that the timing for the test (system configuration) can be controlled so the impact is less.  Review of DFR and SER records: this ensures that the normal sequences are functioning as expected. For example, a review (or comparison with a previous record) of a normal or typical block or deblock sequence will provide information on the status for a significant portion of the control system. Usually a sampling of alarms will be triggered into the SER from different I/O boards as a test.

Page 78

Guidelines for Life Extension of Existing HVDC Systems

Training of Personnel:

Various levels of training required:

 Plant operators  Plant maintenance staff  Plant engineering staff  System engineering staff

3.5. References: [1] Guidelines for Life Extension of Substations EPRI 1007517, Chapter 6 Relay and Controls Systems

[2] IEEE panel session paper 0-7803-7322-7, “Celilo Converter Station Mercury Arc Valve Replacement,” W. Litzenberger, 2002

[3] IEEE panel session paper 0-7803-7322-7, “Rehabilitation of Chateauguay and Madawaska HVDC Interconnections,” J Primeau et al., 2002

[4] IEEE panel session paper 0-7803-7322-7, “HVDC Life Extension for Nelson River System,” R. Valiquette, 2002

[5] IEEE panel session paper 0-7803-7322-7, “Life Extension of the New Zealand HVDC Link,” M. Clark, 2002

[6] IEEE panel session paper 0-7803-7322-7, “Renewal of the Sylmar Converter Station”, C.T. Wu, 2002

[7] IEEE Power Electronics Building Block (PEBB) Concepts; Document Number: 04TP170

[8] IEEE PES i8 WG on "Power Electronic Building Blocks has a Task Force working on Characterization of Control Architecture for Power Electronic Building Block Systems". "Guide for Control Architecture for High Power Electronics (1 MW and Greater) used in Electric Power Transmission and Distribution Systems".

Page 79

Guidelines for Life Extension of Existing HVDC Systems

3.6. Valve Cooling Description:

This section discusses the cooling that is external from the Thyristor Valve itself and outside the Valve Hall. This section is not intended for discussing the building heating and cooling systems commonly called HVAC here as they are usually separate from the equipment cooling systems. There are so many different variations and ways to implement the cooling systems it is not possible to list them all but only to describe the more commonly known ones here. The valve hall air is pressurized to prevent dust incursion, humidity controlled to prevent flashovers, and cooled to around room temperature on older installations. In modern day installations the Valve Hall air temperature is allowed to run at 50 oC or 60 oC. In colder climates the outside air may be used to cool the air as humidity rapidly decreases in the air below freezing as it is cost effective. The valve hall air is cooled by Refrigerant Condensers (RC) for water cooled Thyristors. A typical 500 MW valve hall would require two 50 ton RC’s, one for the main and one for back up. The air requires filtering with a coarse filter in series where a fine filter or an electrostatic filter is used. These systems are usually duplicated to allow for maintenance and outages.

FIGURE 37 VALVE HALL REFRIGERANT CONDENSERS

Air cooled and air insulated Thyristor Valves require large amounts of air and this poses a challenge for filtering and environment control. Very large fans are used to drive large volumes of air and it is restricted generally to lower MW HVDC schemes of 1 100 MW or less.

Page 80

Guidelines for Life Extension of Existing HVDC Systems

FIGURE 38 AIR COOLING SYSTEM FANS

For the water cooled valves in cold climates, there is usually a two stage system with an intermediary heat exchanger but there could be a single loop using a high grade of glycol. This could be large chillers, shell type heat exchangers, or plate and frame heat exchangers. The systems could be combined to include the heat from the valve hall air systems. The plate and frame heat exchangers have become more common recently because of their compact size, low cost, high heat capacity and ease of maintenance. These systems are usually duplicated for 100% redundancy or with three units for 50% redundancy.

FIGURE 39 PLATE AND FRAME HEAT EXCHANGER

The secondary circuit is usually glycol filled in cold climates but may be water only in moderate climates. The end device heat exchanger transfers the heat to the air, many with the assistance of water. The water is used when the air temperature is higher with just air used at lower temperatures. The water can be sprayed on a fill type material to make use of the water evaporation high heat dissipation capacity, be sprayed in an air curtain with the cooling air drawn through the water spray or be sprayed directly on the coolers themselves. Another system is geothermal with water being drawn from deep wells and the discharge water into shallower wells some distance away. The secondary circuit will usually have a spare cooler and fan which can be valved-off for maintenance and repairs

Page 81

Guidelines for Life Extension of Existing HVDC Systems

With the modern day schemes the use of water in the air has been eliminated wherever possible.

FIGURE 40 DRY TYPE COOLER

When the Thyristor Valves are in-service they generate very large amounts of heat. However during lengthy outages in cold climates, heat is required to prevent the Deionized Water (DIW) from freezing. This heat is usually provided by electric resistance elements or natural gas.

The control for the valve cooling in most of the older schemes is not duplicated but the protection trips are normally duplicated, the sensors may not be but should be. The older valve cooling controls have analogue controls with the newer ones run by a PLC. Sensors are used to monitor temperature and flow rates and are usually mechanical in nature. Pumps are normally 100 % redundant or 50 % redundant. There is usually a three way valve for water cooled systems and heat recovery for air type systems for recirculation.

Performance Issues:

The cooling systems being mechanical in nature have several performance issues.

The largest problem for refrigerant condensers is the loss of the cooling media (Freon). This is a particular problem in colder climates.

For the primary heat exchangers using chillers or shell type of plate and frame units, there is the reduction in efficiency over time. The tubes must be cleaned by reaming and the thickness must be measured by X-rays.

Whenever water is used in the secondary heat exchangers for evaporation or as an air curtain, it is likely that it will have to be treated to prevent the build-up of scale. It will then be treated via a zeolite softener but getting rid of the salt brine becomes another issue to deal with. If the water is recycled there is a potential for bacteria, algae and other contaminants to concentrate in the water. Treating the water with chemical ozone or other means is required. Water is more complex than it first appears and any change to the water source can have dramatic effects on the equipment performance.

Page 82

Guidelines for Life Extension of Existing HVDC Systems

In the case of water and or water-glycol systems, leaks are the major concern. Glycol is toxic and is an environmental concern for spills and releases. It is very difficult to contain and clean up because it readily mixes with water. Also it may leave a sticky residue to be cleaned up.

The cooling system controls generally cause a number of the problems. Many of the parts have short life times and being mechanical requires significant testing and adjustment.

Technical Life Assessment:

Most of the valve cooling equipment has a life of 25 years +/- 5 years assuming proper maintenance has been performed and the equipment has not been abused. Thus, about 5 years before the end of life a life assessment should be performed.

In the case of Refrigerant Condensers, the Freon is ozone depleting substance and a global warmer and thus subject to strict environmental regulations governing releases and usage. Phase out of the use of a Freon such as R12 in the past and R22 in the future may limit the useful life of the equipment unless compatible refrigerants can be found.

For the primary heat exchangers, using chillers or shell type of plate and frame units once all the spare tubes are used up or if many of the tubes become too thin the equipment must be replaced.

In the case of a fiber fill evaporation system the fill will deteriorate over time and must be inspected and replaced as required.

In many cases the life assessment begins on the hottest day of the year when at high loading; the system is unable to keep up and an alarm and or trip occurs. An assessment of each piece of equipment is necessary and if the cause is a general deterioration and the system is older (20 years or more) it is likely that replacement is the best option.

The cooling system controls generally have a shorter life time in the 15 or more years range. However it is not uncommon for many of the individual parts to be replaced over time to allow the controls to continue until the entire cooling system can be refurbished or replaced.

With the air driven systems the air intake is critical to avoid vehicle exhaust, BBQ fumes and smoke from fires getting into the valve hall.

Refurbishment:

The refrigerant condensers being standalone units and redundant are usually replaced when they fail. Some engineering may be required as the new unit is unlikely to be exactly the same as the exiting one and this is best done prior to a failure so the redundancy can be restored quickly.

If the cooling system is 20 years or older, it is likely that the complete system needs to be refurbished or replaced. It is also likely that a replacement may consider a completely different type of cooling system that was originally designed depending on the problems with the existing system, technological developments and options available.

Maintenance:

One utility has taken to removing the Freon prior to winter and the pressure testing and refilling in the Spring. If there is a Freon leak it must be reported if over the required amount. The unit must have the remaining Freon removed and stored, the unit pressurized and checked for leaks, repaired and the Freon reused.

Page 83

Guidelines for Life Extension of Existing HVDC Systems

With an air cooling system, changing filters on a regular basis, fan belts, checking contactors and replacing defective fans is required.

3.7. Station Auxiliary Supplies Description:

The station auxiliary supplies include the station service transformers, medium voltage switchgear, motor control centers (MCC), battery banks, first grade supplies, and uninterruptable power supplies (UPS). There can be one MCC for each pole, or more.

The auxiliary supplies are normally duplicated or in some cases even triplicated.

Performance Issues:

There are usually a large number of performance issues associated with the auxiliary supplies mostly because they are not considered important, well thought out and there are conflicting requirements. The operating staff would like the duplicated auxiliary supplies to be run in parallel to avoid downstream transfer issues. However this increases fault levels and a fault can take the entire system down. A reliable design would have the duplicate system separate to minimize the amount of power that is lost and rely on automatic downstream transfers. However the medium voltage switchgear and Motor Control Center (MCC) are usually under maintained and the transfer does not occur or is slow and may cause protections to operate. Even so, the breakers may need to be exercised (operated) periodically to ensure they will work when required. In some cases the AC auxiliary supplies will be manually paralleled only for a short time to prevent downstream transfers as a comprised solution. Older transfer controls and relays can prove to be troublesome and unreliable and should be replaced if miss-operation becomes a problem.

Battery banks may appear healthy and yet not pick up the load when required. They should be load- tested to ensure they are okay. Some systems will have three battery chargers and two battery banks in separate rooms for each battery system.

The worst thing about first grade supplies is that loads are added which should not be on this type of supply, (e.g. computers printers and the system reliability gets comprised). UPS system are usually just allowed to fail and be replaced as they are duplicated, failure rates can be quite high so adequate spares are required.

Technical Life Assessment:

The battery banks usually last about 12 years, at which time they are replaced. The battery chargers are usually allowed to fail and are repaired.

The first grade supply feed off a separate set of battery banks and inverters. They are treated just like the traditional battery banks and chargers.

Refurbishment:

The medium voltage switchgear and Motor Control Center (MCC) breakers are usually refurbished after some 25 to 30 years and in some cases replaced. A complete overhaul is usually done and they are swapped out with spares and the unit removed from service is sent out to a suppliers shop with new contacts and overloads.

Page 84

Guidelines for Life Extension of Existing HVDC Systems

Maintenance:

The battery banks will be inspected periodically for post leaks, tested for bad cells and the electrical connections cleaned from corrosion. The battery room air exhaust system, hydrogen detection will also be inspected and tested regularly.

The medium voltage switchgear and Motor Control Center (MCC) breakers will also be inspected periodically, timing checked, number of operations and contact wear.

3.8. Ground Electrodes and Electrode Lines (does not include sea electrodes) Description:

Ground electrodes are of two types; vertical also known as deep well or bore hole electrodes and horizontal electrodes which come in many shapes and sizes, circular, linear, star shapes and segmented or continuous. A typical one would be circular, buried below the frost line and consist of many sub- electrodes made of 1.5 inch mild steel in a 2 foot by 2 foot coke bed. The coke bed serves to make the conductor appear larger and reduce the current density at the bed edge. This is an example of the Manitoba Hydro electrodes. Please also see EPRI - HVDC Ground Electrode Overview # 1020116 Technical Update, November 2010, and General Guidelines for the Design of Ground electrodes for HVDC Links, CIGRE Technical Brochure, Working Group 14.21 TF2, 1998. Ground electrodes are normally designed in a bipolar system to run at full load for 30 to 40 days and last approximately 50 years. In the case of underground or undersea cable system they need to last as long as it takes to repair a cable.

There is a myth that ground electrodes are not allowed in the United States of America. This is based on the National Electrical Safety Code (NESC) IEEE C2-2014. This code actually allows Bipolar HVDC Systems to have ground electrodes and there are 4 systems currently in use in the USA. The code does not allow for the continuous use of an electrode ground current but allows bipolar systems to run mono- polar during emergencies and maintenance with no time limit specified. In some operating systems permits allow for emergency operation via ground electrode for a limited time and at some reduced current only.

There do not appear to be any other regulations known in other countries, which explicitly exclude ground electrodes. No survey of these countries has been conducted for the purpose of this Technical Brochure, however, opposition by environmental and political groups have delayed and stopped HVDC projects or caused them to go with a Dedicated Neutral Return (DNR) metallic conductor.

The electrode line is typically similar to a 25 kV or 35 kV or higher voltage distribution line. There are usually two separate conductors, each capable of the entire DC current. The electrode line can be installed on wood pole, steel pole or attached to the DC transmission towers or AC transmission towers or it may be DC cables.

The DNR has become common in recent times and more of an issue with protection and detection of faults and their location. Schemes to do this are being developed.

Page 85

Guidelines for Life Extension of Existing HVDC Systems

FIGURE 41 VERTICAL ELECTRODE TYPE

FIGURE 42 HORIZONTAL ELECTRODE TYPE

Performance Issues:

The performance of a ground electrode depends on having a low resistance upper and lower geographic layer. This is not always possible and thus the HVDC system can run into performance problems. Additionally, a good supply of moisture must be available and the current density, usually in the micro ampere per square centimetre range, be designed to be low at the coke bed or conductor edge.

If the current does not return via the ground electrode path, it can get into the grounded neutrals of transformers causing the core to saturate. They then absorb excessive amount of Vars and generate harmonics into the AC system. The absorption of Vars can cause the voltage to collapse and the harmonics can cause overheating in electrical equipment. There can also be a number of problems in the transformer with gassing, increased deterioration of the insulation and steel overheating. Blocking capacitors in the transformer neutrals, plus a bypass for system faults, has been successfully employed

Page 86

Guidelines for Life Extension of Existing HVDC Systems in a number of cases. There is a commercially available product for distribution transformers but larger power transformers require a custom product. The currents can also cause corrosion of pipelines, and so on. This is usually less of a problem as the pipeline owners also have to cater for Geomagnetically Induced Currents (GIC’s) due to sun spot activity any way. GIC’s are quasi-DC currents which flow from the poles in a general North - South orientation. GIC’s can and have caused the same problem in converter transformers

Other problems can occur if the electrode has to be used continuously for more than the 30 to 40 design days. This causes the moisture to be driven away from one ground electrode; however replacing the moisture can extend the electrode life. In one instance this was successfully employed by Manitoba Hydro in 1979 when the Nelson River Bipole 2 system was required to run mono-polar for many months due to equipment problems. In other cases such as New Zealand, the electrode conductor has been corroded away and requires a new electrode segment to be installed periodically.

The performance issues of the electrode line are similar to any distribution line; however, the difference is normally there is no voltage or very little current in bipolar operation. Thus when the HVDC system goes mono-polar and the electrode line is required, this is not a good time to find out that there is a broken insulator or burned wood pole or DC cables. There are protection schemes available that can detect these types of problems, such as pulse echo systems or forcing current down the electrode temporarily to do an impedance measurement.

Page 87

Guidelines for Life Extension of Existing HVDC Systems

Technical Life Assessment:

In general, if there are no problems at the initial use of the electrode, they are relatively maintenance free for many years. However, if there are extended periods of mono-polar operation or once they have aged after 25 to 30 years, they should be checked and or inspected. For the horizontal electrodes this usually means digging up the electrode at several locations and measuring the conductor for corrosion and the soil for moisture. The iron rod will likely be caked with the Coke material and needs to be wire brushed off to take any measurements. A sump pump may be required if there is a lot of water. A safety procedure is a must for the dig. It is unlikely that the entire Bipole can be shut down and there can be a transient DC voltage if the HVDC system goes mono-polar. If the ground electrode is segmented it may be possible to disconnect the segment being inspected. Resistance measurements can also be taken. Injection of at least 100 Amperes DC and measuring the voltage to a remote ground reference point some 1.6 km away should give a good resistance reading. For example, one HVDC scheme has on line monitoring of both voltage and currents and can do regular resistance measurements and comparison.

For a deep well inspection it would be necessary to dig it up using a “soft dig” system. This system uses water to make a slurry of the ground cover and coke material and then suck it to the surface. In each well there may be sections that can be removed independently and this will likely require a crane to lift the electrode. For deep wells at greater depths, this may not be feasible. A resistance check should be done prior to the inspection so the worst deep well can be the one dug up.

The life assessment of the electrode line or neutral return conductor would be similar to any AC distribution line.

Refurbishment:

The electrode conductor can be replaced or a new segment added as necessary. Refurbishment may also be necessary if the DC current of the HVDC link itself is increased for any reason.

The electrode line or neutral return conductor refurbishment would be similar to any AC distribution line.

Maintenance:

A thermo-couple can be buried with the electrode conductor at various locations and read during extended periods of mono-polar operation. However once the temperature starts to rise the life of the electrode is limited.

Another option with segmented electrodes is to measure the current difference between the various segments and alarms if a certain differential is exceeded. By adding a Voltage Divider with a remote reference point, the resistance can be measured and if plotted over time can show any deterioration of the electrode. One HVDC scheme has on-line monitoring of both voltage and currents and can do regular resistance measurements and comparisons.

The maintenance of the electrode line or neutral return conductor would be similar to any AC distribution line. However if live line maintenance is performed a detailed assessment of the transient voltages should be conducted.

It is possible to add a Pulse Echo System (PES) or a small signal of single frequency (other than the system frequency) to the electrode line if less than 200 km long. The PES system injects a pulse into the electrode line which is reflected at the electrode ground and produces a “signature”. The system with single frequency measures the impedance to the electrode and reacts if the impedance deviates from the pre-fault measurements. It is also possible to increase the current in one pole and decrease the current in the other pole to provide electrode current and take an impedance measurement. This can be

Page 88

Guidelines for Life Extension of Existing HVDC Systems done periodically each day. There is usually a voltage divider at the converter station to detect an open circuit line. There are also lightning arrestors on the electrode line. The electrode line and or neutral return conductor may have current differential systems. These systems require maintenance as well and face all the same issues described above for the electronics, instrument transducers and other DC equipment.

3.9. Reliability Centered Maintenance Maintenance systems have evolved over time to improve reliability and availability and reduce maintenance costs.

The first systems used were calendar based or annual maintenance as it was called where equipment was taken out of service every year or every second year. The results were OK but the costs were high and sometimes right after maintenance there would be failures.

Many suppliers recommended the calendar based or annual maintenance and thus would be required during the warranty period to maintain the warranty but even they have progressed as higher availability requirements have been inserted in the specifications.

The next maintenance system to come along was Predictive Maintenance or Condition Based Monitoring. This system relies upon inspections and testing to categorize the equipment’s deterioration and to eventually recommend or predict when the maintenance of that equipment is required.

Reliability Centered Maintenance (RCM) uses some of all the tools listed above but in addition looks at the importance of the equipment and redundancy. RCM is also a “living system” and needs to be updated as information is available. For example a bathroom fan is allowed to fail but an exhaust fan on the battery room would not be unless it is duplicated. The UPS systems are duplicated so allowed to fail rather than maintained. However, if the failure results in more damage that it would be to maintain and repair, it would still be maintained. Information gathering and Root Cause Analysis (RCA) of failures are important to ensure that the maintenance is effective. There are many ways to do this and Manitoba Hydro uses RCA with the assistance of a computer tool to guide one through the analysis and store the results for future reference. Other systems use Six Sigma or just a trend analysis. The information must be drilled down to the component level otherwise a system or a larger piece of equipment may be replaced other than just a less expensive component.

Manitoba Hydro has used RCM and RCA and has seen dramatic improvements in the forced outage unavailability and a reduction in maintenance costs even though the Bipoles were built in the 1970’s. Each Valve Group is on a 2 year minor and a 4 years major whereas years ago it was every year.

Basslink Case Study Basslink implemented RCM, the study and the results are outlined in Appendix A.

3.10. References [1] EPRI “Guidelines for the Life Extension of Substations,” EPRI document TR-105070-R1CD, Electric Power Research Institute, Palo Alto, CA, 2000

[2] EPRI “Guidelines for the Life Extension of Substations,” EPRI document TR-1001779, Electric Power Research Institute, Palo Alto, CA, 2002

[3] IEEE Standard C57.129 - IEEE Standard General Requirements and Test Code for Oil immersed HVDC Converter Transformers

Page 89

Guidelines for Life Extension of Existing HVDC Systems

[4] IEC/IEEE 65700-19-03-2014 - IEC/IEEE International Standard -- Bushings for DC application

[5] ASTM D 34 87 - 00 Standard Specification for Mineral Insulating Oil Used in Electrical Apparatus

[6] Reference IEEE (1998), “Thyristor Aging”, Milan Cepak, Chandra Krishnayya.

Page 90

Guidelines for Life Extension of Existing HVDC Systems

4. Chapter 4 – Guideline for assessing Techno-Economic Life of Major Equipment

4.1. Operational Issues – Maintenance Cost / Management and Availability of Spares

4.1.1 Types of components used within HVDC systems HVDC installations use a wide variety of very different components supplied by a wide range of suppliers. These can be categorized into component types from the spares replacement point of view.

Commercial off the Shelf (COTS) components.

These are available on the open market and delivered as a complete product. They are ordered against a manufacturer’s catalogue number. These components include power supplies, processor modules, ethernet switches, relays, terminal blocks, fuses, CTs, VTs, MCBs, capacitors, reactors, AC switchyard components, and more.

Product design life, operational life, and reliability figures are defined by the component manufacturers as part of their standard literature.

Configured Products

Some manufacturers offer a configuration service for their own or a third party manufacturer’s COTS products. The configuration service may specify special adjustments or settings that require physical changes or modifications that make the product different from the standard product they are based on. These include power supplies, terminal blocks fitted with standard third party components. These components will have an order number that is specific to their individual use.

Bespoke (customized) Products

These products are designed to meet the functional specification provided to the supplier by the Owner and are normally associated with the DC itself. Normally the specialist supplier is responsible for all the detailed design, and must contractually guarantee that the product meets all the criteria specified in the technical specification. These guarantees will include the performance criteria (sometimes risk mitigated by the suppler via acceptance testing), the design life, and Reliability Availability and Maintainability guarantees.

Typical equipment of this type includes:

 Converter Transformers  Thyristor Valve Cooling Plants  DC Capacitors  DC Surge arrestors  Smoothing Reactors  DC Reactors  DCCT’s/DCOCT’s  VDR  DC switches, disconnects and breakers  HVDC Controls

Page 91

Guidelines for Life Extension of Existing HVDC Systems

The purchase order for these items should also include contractually binding spares replenishment time and specify a minimum time over which spares parts will be available.

4.1.2 Management of Obsolescence. Obsolescence of a component can be caused by a number of factors that are more or less likely for particular types of components.

COTS, Configured COTS components and Bespoke components.

For COTS components the marketability of a product is severely decreased if a better (faster/smaller/more functional) replacement product becomes available. Alternatively changes to legislations that govern the processes or materials used to manufacture the components (lead free solder, greener PCB cleaning technologies, etc.) can make a product very unattractive to manufacture.

Bespoke designs normally use third party components that are subject to the same lifecycles as the COTS products.

Most mass market manufacturers therefore tend to renew their product offering over a period of about 3 to 10 years. The normal sequence is as follows:

1. Market a new product 2. State that the old product is not to be used for new designs 3. Increase the price of the old product 4. Send an End of Design Life (EOL) statement to all customers, often accompanied with a “last time buy” opportunity 5. Part only available on the “grey” market 6. Part available from a “secondary” market (e.g. a supplier who used to work for the OEM) 7. Parts available directly from the sub-supplier of the OEM when the OEM no longer has a contract with the sub supplier 8. Part not available

4.1.3 Components designed to meet a specific Specification. In general; components of this nature do not become obsolete unless the manufacturer goes out of business.

Physically simple components such as capacitors and inductors and resistors normally use a standardized manufacturing process that is parameterized (via specifications and drawings) by the manufacturer’s in-house design team to provide a particular value or type of component. The lowest risk solution is normally to ask the original manufacturer to re-use the original design information to create an exact spare. While changes to legislation and materials used can still affect their ability to produce a spare, the manufacturer can normally be expected to provide a compatible spare, this is the next lowest risk solution. As components of this type are made to a functional specification it is also possible to request a different manufacturer to create a spare, though in this case the specification may have to be reviewed to ensure that the new component is fully compatible.

The above process and spares replacement service should be subject to contractually binding spares replenishment times that permit the assumed site delivery spares replenishment times to be met. The price and service level for this spares service should be agreed as part of an overall framework agreement with the supplier but will likely be challenging. These critical items such as HVDC control and protection should be indented early and a plan in place to supply or an increased number of spares carried by the Owner

Page 92

Guidelines for Life Extension of Existing HVDC Systems

For more complex components such as Thyristor Valves, the manufacturer should be contractually obliged to provide exact spares or compatible replacements within the spares replacement period for at least the longer of the contract warranty period or availability guarantee periods, and then for a specified time frame of 10 to 15 years.

Page 93

Guidelines for Life Extension of Existing HVDC Systems

5. Chapter 5 – Recommendation for Specification of Refurbishing HVDC System

5.1. Introduction It is not intended that this section be of sufficient detail to specify either a Brownfield or Greenfield replacement but only to highlight the major areas that are required.

Converter stations hold several types of AC equipment and DC equipment. Each of these elements has been specifically design for the global product. The heterogeneity of these elements will lead to a difference of life duration for each one. Thus, the refurbishment may be done at different moments, depending of the concerned element, of the lifetime of the converter station which is generally based on 40 years. Due to the specific use and the uniqueness of each converter station, it appears that it is important to specify as detailed as possible the specification of the piece of equipment that needs to be refurbished. It is important also to keep in mind what the Vendors normally supply to keep the costs to a minimum.

The second important aspect that needs to be taken into account is interfaces. Some elements will have a limited interaction with the global converter station, like transformers. Other elements will have interaction with all the pieces of equipment, like control systems.

In any cases, specifying every single interaction is vital in order to have a functional final product.

The number of components and system replaced for Brownfield is less than included in a Greenfield. Greenfield will imply a majority of the components or the most costly components (Example of Greenfield: DC switchyard, valves, controls, converter transformers are replaced but the AC filter and AC switchyard is not).

All the elements of the converter station may be considered by a refurbishment:

• Thyristor Valves, • Associated cooling • Converter transformers • Smoothing reactor • DCCT/DCOCT and VDR • HVDC protection and control system of the converter station • And possibly air conditioning/heating of the valves halls (if present). • DC switchyard area • AC switchyard area (outside the scope of this document) • AC and DC filters (outside the scope of this document) The choice of refurbished elements should be based on a technical and economic analysis as described in chapter 4.

The major difficulty for a Contractor is the implementation of new equipment and a new logic in an existing installation and an existing logic.

5.2. Main Components of a Converter Station: Guideline for the Specification

5.2.1 Thyristor Valves The following list defines the several aspects to identify in the specification and which are the most important aspects:

Page 94

Guidelines for Life Extension of Existing HVDC Systems

Technical part:

 The compliance with existing components like transformers in terms impedance between phases, taping range current and voltages  The existing valves hall and it constrictive capacity (dimensions, capability of the roof, of the floor, door access and interfaces)  Fire and smoke clearing aspects  Fault aspects: to define the faults that need to be detected by the system if required  Insulation co-ordination: to define the requirements in terms of standards ,studies, higher capability arrestor and results  Required lifetime  Design part  Environmental Issues – oil containment  Erection/Installation part  Mitigation of transport damage  Required data, drawings and documentation during the design phase and for each Thyristor Valve. This information depends of the level of detail required by the Owner. Hereby an example of what could be required:

 Total number of thyristors, in all valves together  Installed number of thyristors per valve  Minimum number of non-faulty thyristors per valve for de-blocking  Minimum number of non-faulty thyristors per valve for safe operation  Number of non-faulty thyristors per valve at tripping  Losses for one thyristor level, at rated load, min. load and blocked  Losses for one valve, at rated load, min. load and blocked  Circuit diagram of converter bridge  Circuit diagram of a valve  Thyristor Qrr matching if required  Max. DC voltage stress  Maximum Thyristor junction temperature  Normal operating Thyristor junction temperature Maintenance part:

(The aim is not to specify the time or the tools, or to say how many spare parts the Owner’s needs, it is about asking the OEM to give the information and to highlight some required points):

 The needed time and tools to replace a faulty thyristor  Spare parts including failure rates and optic fibres  Maintenance planning  Estimated time between maintenance periods Redundancy part:

 To be precise and prove that the number of redundant Thyristor’s is compliant with the required availability Testing part:

Page 95

Guidelines for Life Extension of Existing HVDC Systems

 All the tests that the Owner needs to be performed based on international standards or the Owner’s standards  Associated reports (level of design details required, type of test report)  The witness of type tests by the Owner, if required

Page 96

Guidelines for Life Extension of Existing HVDC Systems

5.2.2 Cooling of the valves The following list defines the several aspects to identify in the specification and which are the most important aspects:

Technical part

 Design requirements: o The cooling shall be designed in the same project of valves refurbishment o Type of cooling (water or air cooled) o Single or double loop o Material used in the piping o De-ionized water loop o Operation temperature interval o Types of additives in the water cooled if required o Earthing of all these components o Pumps and fans o Physical interface o Flexibility  Erection/Installation part  Required alarms and trips  Mitigation of transport damage  Expected drawings and documentation Maintenance part:

(The aim is not to specify the time or the tools, or to say how many spare parts the Owner’s needs, it is about asking the Supplier to give the information and to highlight some required points)

 Needed spare parts  Needed tools  Maintenance planning  Estimated time between maintenance periods. Redundancy part:

 Which parts are important for the Owner to be redundant if any? Alarms may not be redundant but protection should be, including the sensors. Testing part:

 All the tests that the Owner needs to be performed based on standards or the Owner standards  Associated reports (level of design details required, type of test report)  The witness of type tests by the Owner, if required

5.2.3 Converter transformers In case of a refurbishment of single phase transformer, it is advised to change all three single phase transformers at the same time in order to have a balanced technical solution. The risk of replacing only one single phase is the possibility of unbalanced voltage and current on the valves unless specified properly for less the 3 to 5 % difference in impedance between phases.

Page 97

Guidelines for Life Extension of Existing HVDC Systems

The following list defines the several aspects to identify in the specification and which are the most important aspects:

Technical part:

 Design requirements: o To stay as much as possible close to the technical characteristics of the existing transformers o The easiest way to specify is to use the specification and tests reports of the original transformers. o The main technical characteristics to specify are : . Available space . Current and voltages . Impedance and reduced tolerance variance 3 to 5% . Harmonic current requirements . Tapping range . DC barriers, DC bushings and DC insulation system design details . Current switching . Cooling systems  Erection/Installation part  Required alarms and trips (OTI and WTI)  Expected drawings and documentation  Pressure Relief Device (PRD) and Bucholtz relay  Mitigation of transport damage Transportation (IEEE C57.150 - IEEE Guide for the Transportation of Transformers and Reactors Rated 10 000 kVA or Higher)  HVDC controls interface  Auxiliary power interface  Civil and foundations Maintenance part:

(The aim is not to specify the time or the tools, or to say how many spare parts the Owner needs, it is about asking the Supplier to give the information and to highlight some required points)

 Needed tools  Maintenance planning  Estimated time between maintenance periods. Redundancy part

 In general, each converter station has one spare part transformer. It is up to the Owner to decide if it is required. This part is not necessarily fulfilled in the specification. Testing part

 All the tests that the owner needs to be performed based on standards or the Owner standards  Associated reports (level of design details required, type of test report…)  The witness of type tests by the Owner, if required

Page 98

Guidelines for Life Extension of Existing HVDC Systems

5.2.4 Smoothing reactor Oil-filled smoothing reactors are similar to transformers. The main idea is to have as close as possible the same technical characteristics as the existing smoothing reactors but the new smoothing reactors will likely be air core reactors. However with new Thyristor Valves, it may be possible to reduce the inductance value of the smoothing reactors and this should be investigated, prior to the Bid specification being finalized.

In many cases the existing oil filled reactors are being replaced with air core reactors to reduce the risk of fire, oil containment, reduce maintenance and reduce capital cost. This will require new foundations and a new DCCT/DCOCT as the existing DCCT is normally installed in a DC bushing turret in the oil filled reactor. A new zinc oxide DC arrestor may also be required.

The following list defines several aspects to identify in the specification and which are the most important aspects:

Technical part

 Design requirements  To stay as much as possible close to the technical characteristics of the existing smoothing reactor, unless a small one is feasible  The easier way to specify is to use the specification and tests reports of the original smoothing reactor.  The main technical characteristics to specify are : o Available space o DC current o Harmonics o Inductance o Interfaces o Existing foundations  Erection/Installation part  Required alarms and trips if oil filled  Expected drawings and documentation  Mitigation of transport damage Maintenance part

(The aim is not to specify the time or the tools, or to say how many spare parts the Owner needs, it is about asking the Supplier to give the information and to highlight some required points)

 Needed tools  Maintenance planning  Estimated time between maintenance periods. Redundancy part

 It is up to the Owner to decide if redundancy is required. This part is not necessarily fulfilled in the specification. Testing part

 All the tests that the Owner needs to be performed based on standards or the Owner standards

Page 99

Guidelines for Life Extension of Existing HVDC Systems

 Associated reports (level of design details required, type of test report…)  The witness of type tests by the Owner, if required

5.2.5 Control system Control system is the most difficult part to specify because of the needed functionalities that may be different from the original ones and the interaction with all the components of the converter station such as the VBE and the communication with the Owner several entities. It is normal to interact and discuss with potential suppliers before coming up with a final specification.

The following list defines several aspects to identify in the specification and which are the most important aspects:

Technical part:

 Design requirements: o Specification of the hardware The Owner needs to define all the standards that he requires and special components that they are used to if applicable (like specific LV cables)

 Specification of the software o Specification of the control principles o Specification of the required functionalities to be maintained from the original station and the new ones o Specification of the automatic mode and the manual mode o Specify all the communication system o Provide the I/O of all the converter station components that interact with the control system o Specify the minimum required protection o Specify the alarms and trips: in this part the requirements are most turned on the important alarms that need to be seen by the operators. The second aspect concerns the categories and the types of these alarms (definition of urgent, non-urgent, audible alarms, non-audible alarms, colour of the alarms on the HMI…) o Specification of the HMI, web based or not. Some web based may have a longer life. o Specification of the AC and DC measurement: definition of the existing components. To ask the OEM if it is compliant with its design.  Auxiliary: description of the existing.  Earthing  Erection/Installation part  Expected drawings and documentation  Mitigation of transport damage  Available space Maintenance part

(The aim is not to specify the time or the tools, or to say how many spare parts the Owner needs, it is about asking the OEM to give the information and to highlight the required points). This is one area however where additional spare parts may extend the life of the controls

 Needed spare parts  Needed tools

Page 100

Guidelines for Life Extension of Existing HVDC Systems

 Fault analysis documentation  Maintenance planning  Estimated time between maintenance periods. Redundancy part:

 To precise and prove that the designed redundancy is compliant with the required RAM Testing part:

 All the tests that the Owner needs to be performed based on international standards or the Owner’s standards  Associated reports (level of design details required, type of test report…)  The witness of type tests by the Owner  Real Time Digital Simulator (RTDS) testing

5.3. Interfaces

5.3.1 Generality The definition of the interfaces in the case of a Brownfield project is critical and more complicated than in a Greenfield project. The definition of the exact list of remaining and refurbished equipment is the first step. Then, for each component that will be refurbished, the interfaces with the existing and remaining equipment shall be described in a high level of detail. This description implies I/O, auxiliary, electrical interfaces, civil and foundations. Other interfaces concern the future replaced component which are its environment, space and mechanical interfaces.

5.3.2 Electrical interfaces The Owner shall provide the electrical characteristics of the remaining components and to specify those of the refurbished components if applicable or needed.

5.3.3 Mechanical interfaces The Owner shall provide the mechanical characteristics of the floor, the roof or any support to the refurbished components. In case of a lack of information, it could be asked to the Supplier to define if reinforcement works are needed or not. This is likely for existing foundations for converter transformers and smoothing reactors and oil containment. This shall be specified in any case.

5.3.4 Environmental interfaces The Owner shall provide the environmental characteristics like fire protection, noise, temperature, oil containment, humidity, wind, snow or any other characteristics that may have an impact on the operation or the aging of the component that will be refurbished.

5.3.5 Space interface The Owner shall inform, via the specification, the Supplier of the available space for the refurbished components. The Supplier shall take into account space constraints, existing foundations and provide solutions during the design phase.

5.3.6 Auxiliaries interface The Owner shall describe the auxiliaries related to the components that will need to be replaced with more or less capacity, different voltages and ask the Supplier to confirm if it is feasible to make the changes that are required.

Page 101

Guidelines for Life Extension of Existing HVDC Systems

5.3.7 I/O interfaces In the case of control system refurbishment the I/O of the remaining components to be integrated shall be provided to the Supplier.

In the case of component refurbishment, the Supplier will have to take into account the existing I/O of the refurbished component and to provide the new one with the same I/O to fit the remaining control system.

5.3.8 Example: valve and control system refurbishment Figure 43 is an example of a valve and control system refurbishment. It is a view of the scope of supplies with the equipment, which have to be refurbished, and the remaining equipment that give all the constraints for the new equipment.

PROJECT SCOPE ‐ EQUIPMENT REMAINING EQUIPMENTS Valve Equipments

‐ Thyristor Valves

‐ Thyristor Valves Cooling ‐ Network constraints ‐ Thyristor Valves cooling Control

‐ Filters DC Control an d Protection system

‐ Transformers ‐ Smoothing reactors ‐ Pole Control ‐ AC Protections ‐ Bipole Control ‐ Low Voltage AC ‐ Harmonic Fi lter control DEFINED ‐ DC cables ‐ TFR (Transient an d Fault recorder) ‐ … Other

INTERFACES TO BE ‐ Heating

‐ Fire Detection

FIGURE 43 EXAMPLE OF VALVE AND CONTROL SYSTEM REFURBISHMENT

The interfaces are mainly made between:

 The physical control system and all the remaining low voltage equipment  The valves and the high voltage components (wall bushings, disconnects, and surge arresters of the smoothing reactors)  The valves, other interfaces which include converter transformers, controls, AC Filters and the buildings  The cooling system and the buildings of the converter station

5.4. Maintainability including spares requirement Maintainability of the converter stations implies different aspects; correct design of the new equipment, good quality of equipment, a routine preventive maintenance plan and the adequate spare parts which shall meet the required availability of the station.

Any single component failure shall not affect the power transfer capability of the HVDC system.

Page 102

Guidelines for Life Extension of Existing HVDC Systems

Preventive maintenance operations must be carried out with the hardware and software in operation, taking into account the redundancy required. If an operation cannot be carried out with the station in operation then the Contractor shall list the maintenance activities that are outage reliant along with details of resources required to complete the work (e.g. staff required, outage time etc.). The Contractor shall state the maximum hours per pole required for maintenance per year.

The Contractor shall propose a spares structure that enables the Owner reliability and availability objectives to be met and shall detail corresponding deadlines for set-up and restocking and the costs involved.

Spare parts shall be tested and furnished as are required for the safe and reliable operation of the plant to the target availability level. The required numbers of each spare for a life of 10 to 15 years shall be detailed by the Contractor.

All spare parts shall be readily interchangeable with those that they are to replace. They shall have the same physical, mechanical, material and electrical properties as the installed equipment (e.g. specified conditions in relation to tests etc.) of the installed equipment shall also apply to the spares.

All spares shall be properly packed in such a manner as to allow prolonged storage at site, taking into account the ambient conditions there.

It is very important that the spares in stores be updated as modifications occur to the in-service equipment so that they are viable spares.

During the warranty period the Contractor shall provide a spares management service and shall ensure the quality and integrity of stored spares on an annual basis.

5.5. Cost minimization All interconnections have commercial considerations, such as cost minimization. One of the key factors of a cost minimization would be to reduce, as much as possible, the outage and site testing period.

The following list is a first draft of propositions that could reduce cost:

 The spares shall be divided between the stations so as to optimize RAM.  To determine and finalize the project details at the concept stage as much as possible and then freeze the design.  Any design changes after that require management approval, supplier approval and detail the reason for the change, financial and technical justification and cost.  To have a realistic but tight planning of the project  To have the same Supplier for both or more sites  To have as much as possible detailed specifications while still giving the Supplier the flexibility to design and reduce costs  To optimize the actions: to do as much as possible actions in parallel as feasible.  To produce contingency plans for possible delays, i.e. transport damage  To perform deeply all the required studies (Civil works, complete software FAT, complete equipment FAT)

5.6. Replacement time minimization Replacement time minimization can be reached by multiple ways:

 Training program (training on mock-up for instance)

Page 103

Guidelines for Life Extension of Existing HVDC Systems

 Clear and simple O&M Manual. The O&M manual shall contain replacement sheet. Each replacement sheet shall describe the actions to be taken, step by step, the required tools for each action and the level of security  Holdings and availability of spare parts  Considerations for replacement during design to ensure ease of access, removal and replacement. In addition to spares, a full set of all special tools and test equipment to enable full operation, fault diagnostic and maintenance of plant shall be included in the purchase and supply sufficient spares to provide adequate redundancy of the tools and equipment that will see the greatest levels of wear and tear.

Furthermore tools and test equipment shall be provided on the same basis as spares (e.g. management of obsolescence).

5.7. Operation outage minimization

5.7.1 Outage due to refurbishment works: Brownfield and Greenfield Greenfield needs a longer outage than Brownfield as the perimeter of replaced components is bigger.

To minimize the outage of Brownfield and Greenfield, several steps and verification need to be followed:

 A complete factory testing in order to minimize as much as possible onsite testing  A well-defined and realistic schedule with sequencing and resource estimates  To promote the works in parallel if possible  Financial incentives if justified to speed up the process The main points related to Brownfield are the following:

 An exact description of the implied interfaces, done by the Owner  A detailed study of each one of the interfaces, done by the Supplier  Supplier and Owner need to work jointly before and during the construction phase

5.7.2 Outage due to a forced maintenance Forced or corrective maintenance is unscheduled maintenance, which is carried out in the event of malfunction or damage to the stations’ equipment.

The malfunction or failure of the equipment may require a pole or bipole to be disconnected immediately, or at a more convenient time if the operation of the converter is not challenged (e.g. where redundant equipment exists).

Two scenarios have thus been identified according to their impact on availability and an organization of corrective maintenance should be provided by the Supplier.

In case of requested site intervention, the following action would save time:

 All the needed spare parts on one of the two sites or needed spares at each site  Technical fluency by the intervention teams in the language of the country  A first quick action is to have an intervention over the telephone for a first analysis of the event.  Authorization (depending on the country where the visit takes place) and skill of the intervention teams, effectiveness of the visit: resolution of the problem identified on site as quickly as possible  Comply with “safety rules” and “environmental rules” see Chapter 7 of the Owner

Page 104

Guidelines for Life Extension of Existing HVDC Systems

5.7.3 Outage for programmed maintenance Programmed maintenance needs to be prepared with a planning, a work preparation, inventory of the spare parts and tools and to be sure to have all the needed resources.

Guarantees, Performance and Warranties:

The guaranteed losses for the entire HVDC scheme and or individual equipment need to be specified. The reliability performance needs to be specified such as the failure rate per bipole forced outage (i.e. one every 10 years), pole forced outage rate (i.e. one per year) and valve group forced outage rate.

Also the availability of forced outage equivalent (i.e. one percent) and schedule outage rate (i.e. one percent) as well as maintainability requirements are normally part of the guarantees. Guarantee of spare parts availability may also be included such as for 15 years,

Warranties may include labour but normally only materials, and can vary from 2 to 5 years or more.

Page 105

Guidelines for Life Extension of Existing HVDC Systems

6. Chapter 6 – Testing of Refurbish/Replacement Equipment

6.1. Introduction It is very import to check the equipment for transportation damage when it arrives at site or significant delays and costs can be incurred.

Onsite pre-commissioning tests allow the verification of the correct erection and connections of the replaced components. Commissioning tests prove that global correct operation of the converter station and confirm the successful integration on the new component. The aim of this chapter is to provide a high level list of the tests that need to be done for each refurbished component for both high voltage and low voltage. These tests concerns pre-commissioning and commissioning.

6.2. High Voltage Equipment

Valves The pre-commissioning of the valves is composed of different tests:

 Coolant hydraulic pressure test  Joint resistance measurement  Fibre optic continuity check  Visual checks and cleaning  Leak detector operation  Level by level testing The description and the objective of the tests are detailed in the following table.

Page 106

Test Description of Test When Performed Objective, goal and purpose

Coolant hydraulic pressure tests: the test is The objective of this test is to performed to check the absence of leak in the detect an eventual leak of the Coolant hydraulic Before fibre optics are coolant circuit in the valves hall. cooling system in the valves pressure test installed Leak free is defined as totally dry surface. and the valves hall or to prove Leak is defined as any wetness the absence of leak. All bus bar joints assembled at site to be The test shall prove that joints contact resistance measured. have been correctly performed Inter-Tier and Tier to Measurement must be taken directly at the joint and that the resistance does not Stress Shield Bus bar Soon after joint assembly result should not include any additional bus bar exceed the expected level. joint resistance in the measurement loop. measurements. The values shall not exceed a level given by the Supplier. The main purpose of this test is to check that all the fibres are Fibre Optic continuity Visual continuity checks to be carried out on Before or during correctly incorporated in the check each fibre. installation. modules and in the valves and to detect faulty fibres. Cleaning of : Dust and debris on the valves The goal is to handle the final Module-module bus bars and joints cleaning of the valves and the Flexible silicone sheds on the inter-tier column valves hall. The second After all valve assembly Visual inspection and insulators objective is to verify that complete. clean nothing has been forgotten on

The test shall allow the verification of absence the valves modules. Visual of : inspection for leaks. Equipment/tools left at each level of the valves Damages wiring and components

Page 107

Guidelines for Life Extension of Existing HVDC Systems This test is performed to prove the mechanical Only after VBE and The goal of this test is to prove Check of Leak operation of the Valve Coolant Leak Detectors Engineers Interface the correct operation of the Detector operation. and their associated alarm and system tripping sequences. Commissioned leak detector.

Valve level pre-commissioning confirm that each gate unit can be charged and send data back via optical fibre. The test is required on Before any bus bar The primary objective of these valves, level by level. The VBE will be power connections to the tests is to verify the correct Level by Level required and must be set up to perform single valves. performance of each Thyristor functional checks. level tests. These tests should be performed on Only after VBE and Level gate electronics in the each Thyristor level before energization of Engineers Interface valve and correct location of valves. The status of the levels would be commissioned. fibres in the VBE. checked via the Engineers Interface.

Page 108

Converter transformers There are no special tests due to refurbishment for a replacement of 3 single phase transformers or a 3- phase transformer.

The factory tests shall follow the IEC 61378-2 (HVDC) or IEEE C57.129. For the onsite tests, the first step is the control of the transformers erection. A transformer test and check plan shall be prepared and followed for all the components. Then, transformers shall be energized. The final tests are the on load test by increasing the power.

In case of replacement of only one single phase transformer, some specific measurements shall be performed on the 2 remaining single phase:

 An onsite measurement of the mutual inductance  An onsite measurement of the capacitance between windings on the valves side  An onsite measurement of the capacitance to ground  An onsite measurement of the reactance between units less than +/- 5% These tests shall also be performed on the new single phase transformer in factory. The results of the onsite and factory shall the same.

Once the single phase transformer is installed and erected, power tests shall be performed and special attention shall be provided to operation of the Thyristor Valves (commutations).

Smoothing reactor • There are no special tests due to refurbishment for a smoothing reactor replacement. • The factory tests shall follow the IEC 60076-6 or IEEE C57.16 • For the onsite tests, the first step is the control of the reactor erection for possible physical damage • Ensure that the cooing vents are free of clogging and debris. • Ensure that there is no cracking or other damage due to transportation or installation • Measure resistance and inductance • The final tests are the on load test by increasing the power and the check of the require signal • There is no pre-commissioning test needed.

Surge arresters The purpose of this pre-commissioning test of the surge arresters is to ensure that no mechanical damage is done to the arrester during the transport and installation.

This test shall be carried out by inspecting (visual inspection) the surge arrester for any wear and tear after it is installed.

Once in service the leakage current can be measured to ensure it is within an acceptable value

Voltage dividers The voltage dividers may not be refurbished equipment. Their pre-commissioning is necessary because of the information of voltage and current given by these apparatus.

This test is performed to prove the measurement of the voltage dividers have not been affected by the replacement work of the control system.

The test implies a disconnection of the high voltage bus bar for the divider and a connection to the new control system.

A first measurement campaign shall be undertaken before dismantling the old installation. The results would be the reference for the second measurement campaign after the installation of the new

Page 109

Guidelines for Life Extension of Existing HVDC Systems equipment. The second campaign shall take place after the final installation, connections and pre- commissioning of the control system cubicles and before the commissioning tests of the converter station.

6.3. Low Voltage Equipment

Control system

Site Installation Tests

During the pre-outage period the system is unpacked and a visual inspection is carried out to check for any obvious delivery damage. The equipment is moved into the final position, all mechanical mountings are completed followed by connection and checking of power supplies and powering up. Interconnections to other new equipment and communication systems are made where possible without disruption to an existing plant. The running and preparation of cables is completed, but not the connection to the existing equipment. The final stage of installation can only be completed after the start of the outage period; this mainly entails the final connection of the existing plant signals to the pre- prepared cables.

Pre-Commissioning Tests

Once the system has been connected a signal by signal check of all the plant connections will be carried out, this will be done against a simple signal list and an agreed pre-commissioning test specification. The actual testing entails the manual operation/injection of all inputs and the monitoring of all outputs as close to the plant equipment as possible. This is an end-t- end check, from the plant equipment to the control system software and can be proven by using of the control systems event logging facilities. Test records are created by “green lining” or ticking a copy of the plant signal list and test specification to create a test report. Supporting evidence is also created by storing the relevant sections of the control system‘s log files.

As the control system continuously monitors and time stamps all changes, as well as point to point signal checks, correct and incorrect multiple signal routes can be easily identified.

Typical tests would include:

 Individual equipment standalone tests (power supply, basic communications)  Inter-equipment communications tests (paying particular attention to interfaces between old and new equipment)  Digital I/O checks (including links to outstations)  Analogue I/O checks (including secondary injection of VTs/CTs)

Transient Fault Recorded (TFR) The purpose of these tests is to ensure that the TFR has been configured correctly and that all the required signals are routed to the correct places.

This will usually consist of secondary injections of transducers and the forcing of points in the control system to generate the correct events.

6.4. Auxiliaries

Cooling

Page 110

Guidelines for Life Extension of Existing HVDC Systems

The tests are intended to check that the cooling system has the characteristics required for correct operation of the system in which it is implemented and are composed of the following steps:

• Installation controls • Verification of the power supply • Verification of the water quality and the conductivity • Filling, venting, tightness testing and flushing of the system • Leak verification • Cooling towers • Test of the manual functionalities • Operation tests • Interface testing • Audible noise testing • Parameter monitoring These tests and verifications shall be done before the commissioning of the converter station.

Battery chargers The tests are intended to check that the battery charger has the characteristics required for correct operation of the system in which it is implemented.

It consists of a 2 phase testing: a. Preliminary checks:

 battery charger cabinet grounding  nameplate for conformity with reference drawings and supply  connections for conformity with the reference drawings and their tightness  For loose connections  That all printed circuit boards are correctly plugged-in  That the fuses are fitted, that their ratings and types conform to the manufacturer’s manual, and that their micro switches are set  That the ventilation grids are not obstructed b. Testing and commissioning:

 Electrical checks: correct earth of the cabinet and the insulation resistance of input and output power cables  Check at the rectifier output terminals that alarms, indications and measurement returns are correct.  Setting and testing of alarms, float settings and overcharging This initial energizing of the battery shall be performed only if all the checks have been successfully completed.

6.5. Fire detection The purpose of this pre-commissioning test of the fire detection is to ensure the correct operation after the transport and installation on site.

It consists of the checks of the installation, wiring, connections, and power supplies in accordance with the locations and installations drawings and to test the operation of the entire connected fire detection system

Page 111

Guidelines for Life Extension of Existing HVDC Systems

6.6. Tests of the connection with the existing equipment Verification of the status of the AC and DC switchgear areas: status of the disconnects, of the circuit breaker and other AC equipment.

Commissioning procedures Once the pre-commissioning tests have been completed, for all elements of the new system, Energized or System tests can be performed to an agreed Energized Test Specification.

Commissioning tests It is possible to split the commissioning tests into different groups:

Energization tests:

The purpose of these tests is to verify the correct installation and connections of the valves, to test correct installation and connections of the control system cubicles and to detect faulty optic fibers. These tests correspond to the open circuit tests and the open line tests.

Low power tests

Mono-polar tests

Bipolar tests

High power test

Mono-polar tests

Bipolar tests

The tests shall be done with telecoms and without telecoms. The emergency stop should be tested in the first tests (energization tests). A system changeover shall be performed during testing, during ramp up and between each test in order to prove the correct operation of each redundant cubicle.

6.7. Example of commissioning procedures: IFA 2000 refurbishment

Page 112

Guidelines for Life Extension of Existing HVDC Systems

Open circuit test:

Configuration sheet

Test No: 4 Essai No:

Tests required to be complete before this test can commence: None Essai No.s doivent être effectués et réussis avant le début de cet essai :

Test Lead: Essai dirigé par:

Test Stage: 1 Essai Etape Power Transfer Direction: N/A Direction de transit de la puissance: Power Loading (Max) MW 0 Puissance maximale durant l'essai MW

Operating Mode: Mode de Fonctionnement:

Cable Configuration: Configuration du câble

Station Earth Location: Sectionneur de Terre Fermé

Master Station Selection: Sélection station maître

Control Cubicle: Station Bipole CCP Pole 3CCP Pole 4 Contrôle Cubicle: Active Control Lane: Sellindge Lane 1Lane 1Lane 1 Voies en contrôle: Les Mandarins Voie 1Voie 1

Automatic Bipole Control: Automatique

Active Control Point: Sell OWS Point de Contrôle Actif:

Test Stage: 12 Essai Etape: Emergency Stop: Arrêt d'urgence

Page 113

Guidelines for Life Extension of Existing HVDC Systems

Test Sheet

Executed Time / Activity No. Location Activity By Date Name TEST 4B - FIRST ENERGISATION TEST POLE 3

T 4.0 See "Configuration" sheet for setting up for the test.

INITIAL CONDITIONS FOR POLE 3 - FIRST ENERGISATION TEST - TEST 4B

T 4.1 Operator Work Station Selected for Use: PO séléctionné pour l'essai :

….. ….. initial conditions for BP testing to be prepared ….. ….. …..

T 4.7 POLE 3 FIRST ENERGISATION TEST - TEST 4

T 4.8 Establish that all communications systems are in place and working

T 4.9 On Bipole 2 Select Control to Operator Work Station

T 4.10 AC and DC switchgear configuration to be prepared to reach the Open Circuit status

T 4.22 Ensure Transient Fault Recorder Trigger arrangements are in place

T 4.23 Close the Circuit breacker Test is Expected to be 30 seconds in duration

T 4.24 From the viewing gallery observation points, check for signs of distress on equipment in the Valve Hall

T 4.25 Make Observations and Logs as the On-Load Test Specication

T 4.26 Carry Out Live Trip Test by Operation of the Emergency Stop Button

T 4.27 Confirm Trip of CB

T 4.28 Confirm CB X1310 Auto Opened

T 4.29 Examine the results from the TFR and the additional recording equipment. Record the results of the transformer and valve hall observations

On Bipole 2 T 4.30 Confirm all Protection relays reset Confirm all Alarms reset.

T 4.32 First Energisation is Now Complete On Bipole leave switchgear in a suitable state to proceed with the next planned test or shutdown for valve hall T 4.33 inspection

Page 114

Guidelines for Life Extension of Existing HVDC Systems

7. Chapter 7 – Environmental Issues Environmental issues could be the driver behind a requirement to upgrade or replace a piece of equipment when it is necessary to comply with revised environmental legislation or to meet company policy or standards which are beyond the original capabilities of the installed equipment or systems.

Addressing environmental issues could also be a further benefit associated with a planned replacement of equipment or systems under consideration for life extension issues. Life extension projects could provide an opportunity to enhance environmental performance or accommodate revised or future anticipated environmental requirements and these opportunities should be considered as part of any upgrade plans.

Environmental issues that place constraints on how work is completed should be considered as part of a life extension project to ensure that unexpected demands to comply with environmental requirements do not add additional time and expense to a project. Also, environmental issues need to be considered prior to a life extension project to avoid any inadvertent environmental damage. Items such as handling and disposal of contaminated soils or asbestos for example, need to be considered when developing life extension plans.

All life extension work needs to consider present and future environmental issues to ensure a station designed and constructed under different environmental rules and conditions will continue to comply with new regulations.

Environmental requirements are very much driven by government regulations, specific interpretation of regulations and Owner policy/standards. As such the region and business environment in which the Owner operates dictates the specific actions and investment put forward to address environmental issues and manage environmental risks. Additionally, specific site conditions and parameters, such as location to environmentally sensitive areas play a role into what and how certain environmental risks are managed.

Unrelated development within the vicinity of an HVDC converter station and evolving environmental legislation can subject the facility to unplanned environmental risk. It may be possible that a site is affected by development in the immediate area, resulting in more stringent environmental constraints than considered in the original design. For example, residential development in the vicinity of a converter station that was considered an isolated site, could lead to more stringent regulations with respect to issues such as spill containment and audible noise.

Chapter 7 comprises the following sections which identify environmental issues to consider and potential opportunities to address those issues during life extension projects.

7.1. Insulating Oil Many pieces of equipment within an HVDC switchyard are filled with oil, which is necessary for cooling and providing electrical insulation for that device. The requirement for spill containment is generally legislated and the degree of containment could be dictated by interpretation of that legislation and company standards. Site location and site conditions play an important role in evaluating appropriate containment to minimize risk and consequence associated with any spill.

Oil filled equipment varies in size and associated volumes of oil; the volume of oil present typically defines the spill containment requirements and the type of containment.

Oil filled equipment within an HVDC station includes;

Page 115

Guidelines for Life Extension of Existing HVDC Systems

• Pad mounted transformers • Converter transformers • Smoothing reactors • DCCT’s • Voltage dividers • Wall bushings • Transformer bushings As aging equipment is replaced or upgraded, evaluation of oil containment requirements and opportunities to reduce oil on site should be included in the development of life extension plans.

It may be possible to take advantage of a life extension project to enhance oil containment or replace oil filled equipment with new technology which eliminates oil completely for that piece of equipment. For example oil filled smoothing reactors could be replaced with air core smoothing reactors. The benefits of eliminating the oil includes elimination of oil release risk, elimination of spill containment requirements, and elimination of oil fire hazard.

Oil Spill Containment:

Planning criteria for spill containment, based on existing regulations, anticipated future regulations, site specific hazards and company policy should be developed and consider the following;

• Size of equipment and oil volume requiring spill containment • Type of containment to be used for specific volumes of oil • Size of containment • Handling of oil beyond containment infrastructure • Evaluate central collection system requirements, compare localized holding vs. collection and centralized separation • Site Discharge • Removal of oil from sensitive areas in case of oil fire Note that there are numerous industry standards, guides and reports along with local legislation to consult when evaluating oil spill containment requirements and developing a strategy.

7.2. Polychlorinated Biphenyl Polychlorinated Biphenyl’s (PCB’s) were used as coolants and insulating fluids based on their chemical stability (low flammability) and electrical insulating properties. Although the manufacture of PCBs has been banned throughout the world since 1979, they may still be present in equipment manufactured prior to the ban.

Environmental legislation is aimed at controlling and eliminating the release of PCB’s into the environment. As a first step to comply with any regulation, an inventory of equipment containing PCB’s, the concentration in parts-per-million (PPM) and the volume needs to be established.

Various world agreements, federal, provincial/state and municipal regulations will dictate (usually the most onerous will apply) the levels of concentration that are allowable and the time frame for removal. As part of any life extension planning the opportunity to eliminate PCB contaminated equipment should be considered when considering replacement of equipment or evaluating life extension options.

Additionally, there may be equipment that is sealed, where the ability to obtain oil samples for verification of PCB content is not possible. It may be possible to verify the presence of PCB’s by consulting equipment manuals or reviewing previous maintenance records. If the presence of PCB’s cannot be

Page 116

Guidelines for Life Extension of Existing HVDC Systems reliably confirmed it may be necessary to remove pieces of equipment or a sampling of pieces (where there are many) from service for destructive testing to verify PCB content. In this case procurement of replacement equipment is necessary.

Any equipment identified as PCB contaminated will require appropriate handling according to regulations to ensure the PCB’s are disposed of in a compliant manner. This may require sampling of soil around the equipment and cleaning/disposal of any soil removed from contaminated areas. Disposal will require documenting and tracking shipment of PCB’s to a licensed facility for destruction and verifying that destruction has been completed.

From a life extension project perspective, the additional complication of handling PCB’s and the disposal costs can be a major effort in terms of time and expense and could drive costs up significantly.

7.3. Sulphur Hexafluoride Gas Sulphur hexafluoride gas (SF6) is an inert gas used extensively for dielectric insulation and current interruption in circuit breakers, switchgear, voltage dividers, bushings, bus work and other electrical equipment. Despite the many advantages associated with SF6 gas, including reduced size of equipment, its use is not without challenges.

SF6 is one of the strongest greenhouse gases mentioned by the Kyoto protocol. It has global warming potential about 23 900 times greater than carbon dioxide (CO2). It is also persistent in the atmosphere, having a lifetime of 3 200 years.

SF6 gas is heavier than air. In enclosed areas, it can displace breathable air, thus posing a safety hazard in confined locations. The toxic by-products released when SF6 gas interrupts an arc in a circuit breaker are also a concern. Decomposition products are toxic and adequate personal protective equipment and training are essential for personnel safety.

SF6 gas can be handled many times during product life: gas filling at the beginning of product life, during maintenance top up, sampling and gas recovery at the end of product life. Although regulations involving SF6 gas are not presently stringent there is generally a requirement to monitor and audit gas stocks and to manage the filling and recovery of gas.

From a life extension perspective the requirements for handling SF6 gas in a compliant manner when removing existing equipment or installing new equipment, should be considered during project planning and design. Recovered SF6 gas can be cleaned and recycled for reuse or destroyed in an approved manner.

Alternatively, when considering equipment options the presence of SF6 gas and plans for handling, monitoring and accounting of gas need to be considered when making equipment decisions as well as preparing for equipment installation, commissioning and operations.

7.4. Halon Gas Halons are a class of halogenated chemicals containing bromine that have been and continue to be used as gaseous extinguishing agents in a wide range of fire and explosion protection applications. Halons are very potent stratospheric ozone depleting chemicals when released to the atmosphere. Halons were phased out of production under the Montreal Protocol in 1994 except in Article 5(1) countries where continued production of Halons is permitted through 2009. The phase-out of Halon production has had a dramatic impact on the fire and explosion protection industry.

Page 117

Guidelines for Life Extension of Existing HVDC Systems

Halons are clean, non-conductive, and highly effective extinguishing agent. Halon is an extraordinarily effective fire extinguishing agent, even at low concentrations; it extinguishes a fire by interrupting the chemical chain reaction. It stops the fuel, the ignition and the oxygen from dancing together by chemically reacting with them. Halon 1301 in particular is safe for people when used at concentrations typically employed for “total flooding” fire extinguishing systems and explosion prevention (inert gas) applications. Halon 1211 was widely employed in portable fire extinguishing units for use in what are called “streaming agent” applications.

The primary benefit of Halon gas was its ability to quickly extinguish a fire without damaging items within the room. It is non-conductive, non-volatile, and leaves no residue once the fire has been suppressed. This makes Halon a popular choice for computer labs, museums, libraries and other locations where use of water-based suppressants could irreparably damage electronics or vital archival collections. Halon was also considered an effective choice for protecting electrical installations.

Halon gas has been identified as an environmental risk, along with many other types of refrigerants and chemicals (CFCs) linked to ozone depletion. The production of Halon has been banned in developed countries since 1994. In the USA existing Halon systems are still permitted to remain in service until end of life. However in Canada, Provinces have started to mandate the removal of Halon systems (Manitoba December 31, 2009) except for military and aviation. Existing supplies are carefully monitored to provide reuse and recycling for maintenance and refills of critical systems as needed. All Halon must be recycled or disposed in accordance with regulated guidelines to minimize adverse effects on the environment. Current supplies of Halon are expected to last at least through 2030.

Life extension projects should consider impending changes to regulations, the limited life cycle of Halon fire protection systems and evaluate the opportune time for replacement. While properly maintained systems could remain in use both Halon supplies and system parts are becoming harder to acquire and there are fewer qualified people capable of servicing the older units.

Since Halon manufacturing was banned, fire extinguishing agent alternatives to Halons, in the form of non-ozone depleting gases, clean agent systems, gas-powder blends, powders and other not-in-kind technologies (e.g. non-gaseous agents) as well as water based misting and small droplet systems are now available for virtually every fire and explosion protection application once served by Halons. (Gaseous extinguishing agents that are electrically non-conductive and which leave no residue are referred to as “clean” agents.)

Selection of the best fire protection method in the absence of Halons is often a complex process. Either alternative gaseous fire extinguishing agents, so called in-kind alternatives, or not-in-kind alternatives may successfully replace Halon but the decision is driven by the details of the hazard being protected against the characteristics of the gaseous agent or alternative method and the risk management philosophy of the user.

7.5. Refrigerants Refrigerants are generally classified into one of three substances, CFC, HCFC or HFC.

CFC refrigerants are banned from use or production in most countries. CFC refrigerants have the highest ozone depleting rating and are also a greenhouse gas.

HCFC are banned from production or new use since 2010 in most countries and a phase out is underway. HCFC refrigerants such as R22 have an ozone damaging potential and are also a greenhouse gas.

Page 118

Guidelines for Life Extension of Existing HVDC Systems

HFC refrigerants are used extensively. There is no current ban but responsible use and equipment inspections are mandatory. The HFC refrigerants have no ozone depletion potential but do act as a greenhouse gas.

It is clear from a facilities life extension perspective that considerable management and control of refrigerants is necessary to comply with regulations and reporting requirements. Opportunities to move from banned substances to environmental friendly alternatives should be considered part of facility life extension plans.

7.6. Asbestos Asbestos became increasingly popular among manufacturers and builders in the late 19th century because of its sound absorption, average tensile strength, its resistance to fire, heat, electrical, and chemical damage, and affordability. It was used in such applications as electrical insulation building drains, floor tiles, ceiling tiles, fire stop material and in building insulation. When asbestos is used for its resistance to fire or heat, the fibers are often mixed with cement (resulting in fiber cement) or woven into fabric or mats.

Asbestos fibers are known to cause health hazards to humans. When asbestos containing materials are damaged or disturbed by repair, remodelling or demolition activities, microscopic fibers become airborne and can be inhaled into the lungs, where they can cause significant health problems.

Regulations dictate appropriate handling procedures when asbestos is encountered on the work site. Life extension efforts should consider the cost associated with handling asbestos and the extent that removal from the worksite is required. Where regulations are not dictating, industry best practices should be implemented to handle and manage asbestos. Sampling for asbestos early in a project is recommended as discovery can result in health risks and budget over expenditures.

7.7. Audible Noise Most equipment within an HVDC converter station will generate some audible noise, with transformers, filters banks, and forced air coolers being the largest contributors to noise. Normally an audible noise study will be required from the HVDC Supplier. Factors that also need to be considered are that the actual equipment noise levels may exceed the levels used in the study and field mitigation may be required. In the case of field mitigation, exceeding the maximum operating temperature for acceptable life of the equipment is a significant concern. This problem is compounded by the fact that most equipment cannot be tested for audible noise in the factory with harmonics; only without the harmonics and estimating the audible noise impact of the harmonics. It will be necessary to ensure audible noise levels within the station fence remain at levels to ensure site safety for plant staff. Additionally, audible noise at the station fence must be controlled to within any local standards or regulations. Existing stations that have had development around them may look at re-evaluating the level of audible noise that is acceptable.

To complicate this issues many pieces of equipment cannot be tested adequately in the factory for noise resulting from the harmonics associated with HVDC systems.

If the converter station is located in a populated area, noise mitigation measures may need to be installed and included in any design of replacement equipment or added to the site design. Beyond replacing equipment, other mitigation measures could include physical barriers such as noise damping walls, barriers or earth works.

Page 119

Guidelines for Life Extension of Existing HVDC Systems

7.8. Electromagnetic Effects HVDC converters produce a variety of harmonic voltages and currents that are capable of being radiated from the converter station. Electromagnetic interference (EMI) has the potential for being coupled into various other communication systems.

Equipment replacement specifications should consider acceptable maximum values for the following classes of EMI;

• Television Interference (TVI) • Telephone Carrier Interference (TCI) • Railroad Signal Interference (RSI) • Power Line Carrier Interference (PLC) • Radio Interference (RI) Additionally, as equipment is removed from service for replacement the configuration of the remaining equipment needs to be evaluated to ensure that the EMI interference levels remain within compliance in the interim.

7.9. Mitigation of Environmental Issues Originally developed as part of Cigre Working Group B4-44, the following simplified table identifies potential environmental issues and suggests methods of mitigating those issues based on specific equipment.

Equipment Environmental impact Mitigation Converter transformers Visual Enclosure / housing Audible noise Sound barriers, low noise design Oil Oil containment, plus oil separators Thyristor Valves Audible noise Building Radio interference Shielded building and RI filters Valve halls large/tall: visual Distance to neighbours / hide halls impact in terrain Distance from emission source, EMF shielding IGBT valves Radio interference Shielded building/container and RI Audible noise filters EMF Distance from emission source, shielding Smoothing reactor (air Audible noise Sound barrier core) Visual Enclosure / housing Smoothing reactor (Oil Audible noise Sound barrier filled) Visual Enclosure / housing Oil Oil containment, plus oil separators

DC and AC wall Oil if used Oil containment bushings if required SF6 gas leakage if used Use composite bushings Potential gas leakage monitoring DC isolating switches Radio interference Distance from sensitive area

DC ground switches Audible noise Distance from sensitive area Radio interference

Page 120

Guidelines for Life Extension of Existing HVDC Systems

Thyristor Valve cooling Audible noise Sound barriers Glycol leakage Glycol containment or avoid glycol Chemical cleaners use dry air coolers Auxiliary power and Audible noise Housings station service Oil Oil containment equipment SF6 gas leakage Leakage Monitoring Air handling systems, Audible noise Low noise equipment Enclosures heating and air conditioning DC bus and Radio Interference Conservative design connectors Corona rings

Page 121

Guidelines for Life Extension of Existing HVDC Systems

Chapter 8 – Regulatory Issues With large expansions and development of power grids all over the world, the role of electricity regulators is becoming more and more important, particularly to ensure a tariff structure which is fair and acceptable to both the transmission utility and the consumer. The tariff should be structured in such a manner so that it does not make the consumer pay through his nose, nor does it place any burden of unprofitable operation and subsequent reliability of supply on the transmission utility. At this stage a Transmission System Operator (TSO) may, in the interest of reliability and security of operation, be compelled to replace certain items of equipment or carry out some Renovation and Modernization (R&M). This would therefore cause additional capital expenditure which in all probability would not be covered in the existing tariff structure. Therefore a mechanism has to be developed to ensure that the TSO or asset owner is suitably compensated to take up this additional capital expenditure in the interest of grid security.

These issues can even be more aggravated when the Regulatory Framework of a particular country determines the “end of concession term” for an Owner, say, after 30 or 35 years of exploitation of the service granted. The last years of concession, for such cases, can be very critical in terms of investments for keeping high reliability and availability indices against a pragmatic approach of optimizing assets costs if the concession period (term) is foreseen to be soon terminated

The main reasons for this additional expenditure are:

1. A large number of HVDC stations and AC sub-stations in the world are going to complete their useful life in near future and also some HVDC stations are relatively old. An unreliable element in the transmission system is a potential danger to the EHV Grid System and the same may lead to outages and consequent huge losses. It is therefore essential to keep all elements of Transmission assets in reliable operating condition to meet any contingency. 2. Transmission System elements such as Transmission Lines, Transformers, Switchgear, and HVDC systems are expected to serve for predetermined life span which is the “Useful life” of the equipment and transmission usually from the Commercial Operation Date of the system. The “useful life” in a country like India with hot tropical climate is generally taken as 25 Years for AC Sub-Station/HVDC converter stations and 35 Years for transmission lines. 3. Based on the feedback from users on problems and failures of equipment and also due to technological advancement, manufacturers are often incorporating major changes in equipment to rectify design and performance deficiencies. Such new and improved and modified features need to be incorporated in the existing and operating assets also for increasing the performance and life and to prevent failures. Replacement of certain equipment may also be required due to damage by natural calamity, accident, breakdown, malfunctioning and mal operation. 4. Existing transmission systems are getting continuously augmented with various new system strengthening schemes and addition of generating stations, etc. This leads to the changes in the Fault Level of the transmission network and the existing installed equipment are subjected the changed fault levels and changed load conditions. Hence it is necessary to review the existing equipment with reference to these parameters and the changed conditions may warrant for the replacement of the equipment itself. Another aspect of concern would be the incremental increase on system (background) harmonics due to large loads that can “pollute” the system and by this, stress HVDC filter ratings previously designed. This situation has happened in the FURNAS – Itapúa 600 kV HVDC system, which ended up leading to extra 3rd/5th AC harmonic filter branches installation to cope with the current background harmonics level coming from the network level which increased after the initiation of the scheme’s design. There are a number of

Page 122

Guidelines for Life Extension of Existing HVDC Systems

HVDC schemes were 5th and 7th harmonic filters are required even though it is a 12 pulse system. In order to carry out improvements and changes necessitated due to the above mentioned reasons, substantial expenditures are incurred by the utilities. Normally the cost of regular maintenance, condition monitoring and overhauls are covered in the O&M tariff of the utility. However the problem really arises when the equipment and systems need to be replaced for reasons attributable to changed system parameters or failures and deterioration of performance caused by external factors like natural disasters. Therefore in order to ensure that utilities do not defer costs on this unforeseen expenditure, the regulators deciding tariffs should clearly specify the guidelines and tariff mechanism under which the Transmission utility can carry out the required equipment replacements and modifications during operation stage. The additional tariffs can broadly be divided into two heads:

• Additional Capitalization • Renovation and Modernization An important aspect to be considered too is the Regulator’s policy for the periodic tariff revision application, which directly affects the remuneration of the Owner, and therefore, may affect their policy for O&M investments. Tariff revision, by itself is a very complex subject, but in principle should calibrate the transmission tariff (a fair tariff for consumers), with Utility/TSO’s assets capital return, cash-flow and O&M expenditures coverage enough to assure a stimulated service of good quality.

The useful life of transmission system as given above may be considered for the purposes of Renovation and Modernization (R&M). However given practical realities, some of the equipment may require replacement or major repair for reliable and efficient operation of the transmission system before completion of its useful life. The replacement of some of the equipment may also be necessary in transmission lines like insulators due to pollution. For replacing these equipment and systems, regulators should make provision for this capital expenditure under additional capitalization and adjust the tariff accordingly – tariff revision. The objective here should be to enable Owner’s to overcome operating problems, equipment failures and equipment obsolescence so as to sustain higher availability and reliability of the transmission system.

8.1. Renovation & Modernization The expenditure incurred for extending the life beyond useful life of the transmission system can be taken under the head of Renovation and Modernization (R&M).

The transmission utility responsible for meeting the expenditure on Renovation and Modernization (R&M) for the purpose of extension of life beyond useful life should prepare a detailed cost-benefit analysis along with estimated life extension.

The expenditure incurred or projected to be incurred after a prudent check based on the estimates of Renovation and Modernization expenditure and life extension, can form the basis for determination of tariff after deducting the accumulated depreciation already recovered from the original project cost.

8.2. Conclusion The above proposals will result in improvement of the performance of the HVDC system and increase efficiency and reliability. The revenue for the expenditure incurred for the same has to be realized through the tariff. This expense should not be linked to the regular operation and maintenance tariff as activities on this account which should be carried out as per norms within existing O&M tariff mechanism.

Page 123

Guidelines for Life Extension of Existing HVDC Systems

9. Chapter 9 – Techno Economics

9.1. Financial analysis of refurbishment options

9.1.1 Objective of financial analysis When embarking on a refurbishment project, there will generally be a host of possible approaches and scopes under consideration. The objective of the financial analysis is to compare the financial costs and benefits of the options under consideration, and then to select the option that aligns best with the financial objectives and constraints of the organization. In certain circumstances, especially when the utility has public responsibilities, this is broadened to economic analysis of the costs and benefits to all economic stakeholders.

There is no general rule as to which option is best. For example, if an organization has high levels of retained earnings, an option which has a higher initial cost or higher operating cost, but substantially longer life may be preferred based on the NPV over say 35 years. By contrast, a cash-strapped organization may prefer an option with a low operating cost and shorter lifetime because of the improved cash flow. Therefore, at the outset, it is important that the organization vigorously debates the broad financial environment in which it is operating, and achieves a degree of consensus on the financial objectives of the refurbishment. Where the scope of the analysis includes all economic participants, this is even more important.

9.1.2 Preliminary designs For each of the options under consideration, a financial model is required. Technical issues have a major impact on the financial model though. Therefore, one of the first activities to be undertaken is preparation of preliminary or sketch designs for each option. The initial design concepts must be engineered to the level that reasonably accurate capital cost and ongoing maintenance expenditure can be made. The methodology for implementing the option must also be considered carefully, so that any temporary facilities are provided for. If possible, first order optimizations should be done at this stage, to reduce the number of options as far as possible. At the end of the preliminary or sketch design phase, ideally there should only be a handful of options. A report documenting the options clearly should be prepared and accepted by the critical stakeholders.

9.1.3 Reliability and availability models Having completed the preliminary designs, a reliability and availability model for each of the designs is prepared. Using the Cigre reporting protocol, the impact of each option on the key reliability and availability parameters is quantitatively estimated. The most difficult option to assess is generally the “do nothing yet” option. It will probably be accepted that the ongoing repair and maintenance cost is increasing and the availability (and hence revenues) are decreasing, but a quantitative prediction will almost certainly be contentious. Although it dramatically increases the analysis effort required, it may be required that a “best case”, “likely case”, and “worst case” scenario is prepared for this option.

Since the Cigre reports generate availability parameters that are averaged and annualized, it is normally most practical to model the system on a year by year basis. Trying to model the system in shorter intervals is normally counter-productive. In fact, given the high reliability of HVDC links, there may be some justification for modelling the system over longer intervals.

• Considering the range of options, the sub-systems whose reliability and availability will be affected are listed.

Page 124

Guidelines for Life Extension of Existing HVDC Systems

• Based on the available data, the Mean Time Between Failures (MTBF) and Mean Time To Repair (MTTR) of each of these sub-systems in estimated, as well as the MTBF and MTTR or the systems that will not be refurbished. • Calculated reliability and availability of the whole system based on the MTBF and MTTR of the individual sub-systems should be checked against the input data, to verify the base model. • For each of the options under consideration, the impact of changes to the MTBF and MTTR of the refurbished sub-systems on the reliability and availability of the entire system is modelled. Extrapolating reliability data for sub-systems that are close to the end or their useful life is difficult, because failure rates increase almost exponentially for such sub-systems. Failure rates may increase so rapidly that an entire population of components can fail within a very short space of time, resulting in a sub-system which moves from a high availability to extremely low availability in the same space of time. This can be as short as six months to a year. Nevertheless, parametric investigations of the onset of rapidly increasing failure are possible in these circumstances, and will indicate which sub-systems have the biggest impact.

Reliability models will generally be prepared by a team that includes the operations and maintenance staff as well as specialist reliability engineers.

9.1.4 Financial models Having completed the reliability and availability models, a financial model is prepared for each option. These models will:

• Identify the overall lifecycle cost of the refurbishment, including capital cost and the effect on operation and maintenance costs and losses • Show the financial impact of the improved reliability and availability, in terms of increased revenues and reduced operating and maintenance costs • Show the financial impact of the implementation methodology in terms of the cost of temporary facilities, outages, and constraint charges • Show the financial impact of the extended life The financial model should be on the same basis as the reliability and availability models. It will be prepared by a team that includes both engineers and power system economists.

9.1.5 Impact of discrete events on financial models. In many situations, the timing of discrete events may have a massive impact on the financial model. For example, in a deregulated market the energy charges may be strongly dependent on the status of other links, so a failure in another portion of the network may increase the potential revenues hugely. If the HVDC link is out of service for rehabilitation at this time, the potential lost revenue could be very large. Accurate treatment of these is almost impossible in the financial model. Statistical methods may be applicable to some extent, but even these are generally of limited benefit. Therefore such events during the implementation of refurbishment are normally ignored, unless they have a very high likelihood of occurring. If they are likely, then the implementation methodology should make provision for temporary facilities or for ways of temporarily returning equipment that is not yet refurbished equipment to service quickly if something happens.

9.1.6 Cost-benefit analysis Using the financial models, the financial internal rate of return (FIRR) for each option is determined. The discount rate used in this analysis will generally be uncertain to some degree, so the analysis will be done for a range of values. Typically a best-case, worst-case and expected case will be investigated.

Page 125

Guidelines for Life Extension of Existing HVDC Systems

Where full economic analysis is required, this will be expanded into an economic internal rate of return (EIRR) analysis. This analysis should be performed by qualified power economists together with engineers.

Depending on the criteria to be used in comparing the options, the net present value (NPV) over a defined life, for a range of discount rates, may also be calculated. This has traditionally been one of the most widely used indicators, but it is open to abuse because it is very sensitive to both the life and the discount rate, and the NPV can easily be manipulated to give a result that was selected a priority. FIRR and EIRR are not as easily manipulated.

Example:

The following is a simple example of a cost benefit analysis used by one company called the Capital Project Justification ( CPJ) Process.

Background:

It was identified through root cause analysis that the nylon tubing and associated manifolds, connectors had reached the end of their life and was causing forced outages of 0.33 % per year on a bipole basis. It was expected that without intervention that the number of forced outages would increase. Almost all of the forced outages occurred during the day when loading was much higher than at night. The company involved has a program to perform the NPV tasks so to illustrate here the example will be greatly simplified. The real discount rate used was 6%. The real discount rate is the difference between the cost to borrow the money at 8% and the real of inflation at 2%. The period of 15 years was used as a life extension, as the Thyristor valves would only have that much additional life. The cost to replace the tubing including material, labour and scheduled outage costs was $ 6.0 Million.

Alternatives:

1. Do nothing: The FOA of 0.0033 times 8760 hour per year times 4 valve groups = 115.6 hour or dived by 8 = 14.5 days. The average cost of a daily outage was $ 60 000 USD. Thus the outages were costing 14.5 times $ 60 000 = $ 870 000 per year. For simplicity 15 years times $ 870 000 = $ 13.05 Million USD. This is reasonable as the discounted rate would show a lower number but as the outages will increase the dollar value of the outage costs will be higher in the future. Thus the NPV = $ 6.0 Million - $ 13.5 million = -$ 7.5 Million 2. Replace the Thyristor Valves: The cost to replace the Thyristor Valves was $ 200 million USD. Using the real discount rate of 6% times $200 Million = $ 12 million per year. Plus one has to recover the costs of the $ 200 million over 15 years which is another $ 13.3 Million. Thus the cost of this option is $ 25.3 Million per year. Again for simplicity 15 years times $ 25.3 million = $ 379.5 Million USD. One could argue that the $ 200 million should be amortized over 40 years instead of 15 years but that will still not change the result. Thus the NPV = $ 6.0 Million - $ 379.5 million = -$ 373.5 Million. 3. Replace the tubing: In this case the saving is $ 13.5 Million over 15 years and the cost is the $ 6.0 Million. Normally the savings would not start totally in the first year as this project would take 2 to 3 years to implement, but for simplicity we will assume so. Thus the NPV = $ 13.5 Million - $ 6.0 million = + $ 7.5 Million

It can be seen from this simple example that the best option is life extension for the Thyristor Valves of replacing the tubing instead of the Thyristor Valves themselves

Page 126

Guidelines for Life Extension of Existing HVDC Systems

Appendix A – Basslink Case Study Introduction:

Basslink commenced commercial operations on the 28 April 2006. This case study examines how the maintenance management system was implemented on a practical level and what successes were achieved through examples of implementation.

The maintenance system implemented was based on the principle of Reliability Centred maintenance (RCM), planning the appropriate maintenance for the equipment incorporating a combination of:

 Calendar Based Maintenance  Predictive Maintenance or Condition Based Monitoring Maintenance Management System:

Early in 2006, a purpose built Computerised Maintenance Management System (CMMS), was populated with the maintenance requirements collated for each item of equipment from the following sources:

 Local standards and codes of practice for inclusion of statutory maintenance  Manufacturer’s recommendations from the project documentation The CMMS produces maintenance schedules from the following categories:

 Maintenance during normal operations o Time based preventative maintenance –with the scheduling dates selected to suit local conditions and service provider availability while maintaining a balanced work load for the maintenance team o Condition based preventative maintenance – where data collated from the time based inspections is used to predict the next service date including: • Converter transformer tapchanger operations • AC filter circuit breaker operations and SF6 analysis • Measuring system LASER current trending • Motor bearing vibration analysis • Thermographic surveys • Maintenance requiring an outage o The scheduled maintenance tasks requiring an outage were grouped by bay or area Asset Management:

The CMMS was also populated with the spare parts inventory; a “Spares and Replacement Lead Time Review” was performed resulting in the purchase of additional items.

The Life Cycle of the installed equipment was determined and a Life Cycle Management Plan implemented to ensure continued operation of the Basslink interconnector.

Scheduled Outages:

All information associated with outage planning is recorded on the “Scheduled Outage” network share within folders using a reverse date naming convention so event records appear in chronological order, this information includes correspondence, Microsoft Project plans, contractor scope of work, job tasks, switching plans, data, reports and any photographic records.

Page 127

Guidelines for Life Extension of Existing HVDC Systems

All tasks identified requiring an outage are added to the Microsoft Project plan in the “Next Outage” folder for opportunity maintenance.

The outage project plan includes a task by task analysis organised by bay including provision for converter sequencing and plant isolation and restoration, an outage work package is created for each task including all the necessary information required to safely complete the task.

Corrective Maintenance:

As well as being recorded on the CMMS, all information associated with a fault or event relating to the Basslink Interconnector is recorded on the “Faults & Events” network share, within folders using a reverse date naming convention so event records appear in chronological order, this information includes correspondence, system records, reports and any photographic records.

A root cause analysis is performed on each event and any change identified to mitigate the event is implemented via a change management process.

Performance Monitoring:

The information recorded is summarised annually to report the reliability performance of the Basslink Interconnector to CIGRÉ B4.04

FIGURE 44 CIGRE MONOPOLE AVAILABILITY

The figure above displays the historical performance of Basslink (blue trace) compared with the performance of other monopole HVDC thyristor systems reporting to CIGRÉ.

Page 128