Medium-Term System Adequacy Outlook

MEDIUM-TERM SYSTEM ADEQUACY OUTLOOK

Callie Fabricius, Ntokozo Sigwebela, Sonwabo Damba, Dr Sipho Mdhluli, Prudence Rambau, Dipuo Magapa, Sharon Makopo

Generation System Adequacy for the Republic of South Africa

30 October 2019

SYSTEM OPERATOR | 102 Refinery Road, Germiston, 1401

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

IMPORTANT NOTICE

While the System Operator has taken all reasonable care in the collection and analysis of data available, the System Operator is not responsible for any loss that may be attributed to the use of this information. The changing environment in the South African energy industry means continually changing data that might not have been included in the modelling of this study. The study is not intended to be used as a plan, but rather to explore how possible different futures might test the adequacy of a generation system. Prior to taking business decisions, interested parties are advised to seek separate and independent opinion in relation to the matters covered by this report and should not rely solely on data and information contained here. Information in this document does not amount to a recommendation in respect of any possible investment.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

Table of Contents

1. INTRODUCTION ...... 5 2. Methodology ...... 5 2.1 Adequacy metrics ...... 5 2.1.1 Capacity Adequacy ...... 6 2.1.2 Energy Adequacy ...... 6 2.2 Modelling Framework ...... 7 3. Key assumptions ...... 7 3.1 Demand Forecast ...... 8 3.2 Eskom’s existing and committed supply resources ...... 8 3.2.1 Eskom’s committed build schedule ...... 8 3.2.2 Eskom Fleet Shut Down ...... 9 3.2.3 Eskom’s existing and committed sent-out capacity ...... 10 3.2.4 Eskom Plant Performance ...... 10 3.3 Renewables from Independent Power Producers ...... 13 3.4 Existing non-Eskom installed capacity ...... 14 3.5 Small-to-medium scale embedded generation (solar PV) ...... 14 4. Study cases ...... 16 5. Results ...... 17 6. POSSIBLE Risks to system adequacy ...... 18 7. Conclusion ...... 19 8. Appendix: System Operator Statistics ...... 20 8.1 OCGT Utilisation ...... 20 8.2 Frequency Control ...... 20 8.3 Unserved Energy ...... 21 9. References ...... 23

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

List of Figures

Figure 1: MTSAO methodology ...... 7 Figure 2: Energy Forecast and Actuals ...... 8 Figure 3: Shutdown of Eskom coal-fired station units ...... 9 Figure 4: Eskom installed sent-out capacity (MW) ...... 10 Figure 5: Eskom Fleet Energy Availability Factor ...... 13 Figure 6: Capacities (MW) of non-Eskom plant ...... 14 Figure 7: Estimated Rooftop PV Capacity ...... 15 Figure 8: MTSAO Study Scenarios ...... 16 Figure 9: Year-to-date OCGT utilisation ...... 20 Figure 10: Number of frequency incidents ...... 21 Figure 11: Estimated weekly load shedding ...... 22

List of Tables

Table 1: System Adequacy Metrics ...... 6 Table 2: Eskom New Build Dates ...... 9 Table 3: REIPP Cumulative Capacity (MW) ...... 13 Table 4: Reserves Requirements (MW) ...... 16

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

List of Abbreviations

Term Definition AREP Association for Renewable Energy Practitioners CSP Concentrated solar power DEA Department of Environmental Affairs DMRE Department of Mineral Resources and Energy EAF Energy availability factor FOR Forced outage rate IPP Independent Power Producers IRP Integrated Resource Plan MES Minimum Emissions Standard NERSA National Energy Regulator of South Africa OCGT Open Cycle Gas Turbine OCLF Other capability load factor PCLF Planned capability load factor PQRS Power Quality Renewables Services PV Solar photovoltaic REIPP Renewable Energy Independent Power Producer SSEG Small scale embedded generation SO System Operator UCLF Unplanned capability load factor

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

1. INTRODUCTION

The South African Grid Code (NERSA, 2014) requires the System Operator (SO) to publish a review (called the Medium-Term System Adequacy Outlook or MTSAO) on or before 30 October of each year of the adequacy of the integrated power system to meet the medium-term (five-year future) requirements of electricity consumers.

In preparing the MTSAO, the SO considers the latest information provided by Eskom generators, independent power producers, other non-Eskom generators, the national transmission company, transmission network service providers, and distributors such as –

• possible scenarios for growth in the demand of electricity consumers; • possible scenarios for growth in generation available to meet that demand; • committed projects for additional generation; • demand management programmes; and • reasonable assumptions about imports and exports, and any other information that the SO may reasonably deem appropriate.

The MTSAO measures the capability of the generation system to meet the expected country demand, within component ratings and voltage limits while taking into account the planned and unplanned outages and operating constraints. The expected demand includes the country’s demand plus exports, whereas the supply to meet the demand is all the generation resources licensed by NERSA plus imports, demand-side management resources and rooftop PV. This study provides a statement of the South African power generation adequacy for the calendar years of 2020 to 2024. The adequacy needed to transport and distribute electricity does not form part of the MTSAO.

2. METHODOLOGY 2.1 Adequacy metrics The adequacy metrics used to assess the system’s adequacy of the MTSAO 2019 are shown in Table 1 below. These metrics were chosen after the load- shedding experience in the year 2008, to reflect allowable risk of supply shortages to avoid the unreasonably high cost associated with reducing this risk to a negligible level. The adequacy metrics provide information about the capacity and energy adequacy of the generation system to meet expected demand. The system is deemed adequate if all three of the system adequacy metrics are met.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

Table 1: System Adequacy Metrics Adequacy Metric Thresh old Detail s < 20 GWh per The amount of energy in a year that cannot Unserved energy annum be supplied due to system supply shortages. The combined capacity factor of the OCGT OCGT capacity factor < 6% per annum plant in operation in a year. The combined capacity factor of the Expensive base-load < 50% per annum expensive coal-fired base-load stations in a stations’ capacity factor year.

2.1.1 Capacity Adequacy

Unserved energy and OCGT capacity factors metrics look primarily at the capacity adequacy. Capacity-type contingencies are typically unexpected load increases or co-incident unplanned failure of a number of generating units. These short duration type events (typically hours) are considered capacity contingencies when there is just sufficient total plant capacity to supply the load during high demand hours under the expected supply-and-demand situation. The plant then has insufficient capacity reserve to cater for a capacity-type contingency should it occur. The capacity shortfall would in this case, result in unserved energy for a few hours. Capacity inadequacy will be flagged when the threshold of any one of these metrics is exceeded. The likelihood that the system will be unable to meet the load during a capacity-type contingency then becomes unacceptably high.

2.1.2 Energy Adequacy

The capacity factor of the expensive base-load station looks primarily at the energy adequacy. Energy-type contingency is the occurrence of a significantly higher-than-forecast load growth or the loss of a large source of supply for a prolonged period (weeks/months). The system may be deemed energy inadequate when there is just sufficient base- load plant to supply the load on a continuous basis under the expected supply–and-demand situation. All base-load plant will therefore operate at high capacity factors, resulting in a shortage of energy reserve to cater for an energy-type contingency should it occur. When all base load plant are operating at high capacity factors, the energy shortfall would then require base-load type generation from a peaking plant (mostly OCGTs). This may not be financially sustainable or even possible from a fuel-supply perspective. Energy inadequacy will be flagged when the capacity factors become high, meaning that the system response to an energy-type contingency may not be sustainable.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

2.2 Modelling Framework The MTSAO assesses system risks using Monte Carlo simulation technique, with the PLEXOS ® Simulation Software, in chronological hourly time horizons. A unit commitment and dispatch optimization attempts to balance projected demand using generation capacity resources according to their merit order while taking into account all operational constraints as shown in Figure 1 below. Due to the stochastic nature of the load, renewable energy resources, and plant outages, the model captures uncertainties based on historical data. Variability is controlled using standard error that limits deviation from the mean at any period while changes from one to the next is based on historical autocorrelation. The simulation results are then tested against the adequacy metrics shown in Table 1 above.

Figure 1: MTSAO methodology

3. KEY ASSUMPTIONS This section describes the input assumptions used in the MTSAO 2019. The assumptions are split into demand-side assumptions and supply-side assumptions. The September 2019 year to date information was used for conducting the MTSAO study. The assumptions with the largest impact on system adequacy are the energy forecast, existing and committed RSA generation capacity, as well as the plant performance of the Eskom fleet. Details of the key assumptions are discussed below.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

3.1 Demand Forecast The actual South African energy demand for the calendar year 2018 was 245,77 TWh (StatsSA, 2019) and represents an annual increase of 0,34% from the previous year 2017. The actual historical energy sent out and energy forecast used for this MTSAO study are depicted in Figure 2 below. The first two years of the MTSAO 2019 forecast reflect the current RSA reality of low growth and the last three years are based on the IRP 2019 growth rate. The average annual growth rate for the MTSAO forecast over the study horizon is about 1%. The IRP 2019 energy forecast and the forecast used in the 2018 MTSAO study are shown for comparison.

ENERGY FORECAST vs ACTUALS 290 8 280 7 270 6 260 5 250 240 4 230 3 220 2 210 1 200

ENERGY ENERGY (TWh) 0 190 -1 Y-O-Y % GROWTH 180 170 -2 160 -3 150 -4

2019 MTSAO Y-O-Y GROWTH Actual Y-O-Y GROWTH Actuals 2019 MTSAO 2019 IRP Demand 2018 MTSAO

Figure 2: Energy Forecast and Actuals

Demand is then statistically modelled to cater for the probability of load variability within the hourly median.

3.2 Eskom’s existing and committed supply resources

3.2.1 Eskom’s committed build schedule

The commercial operational dates for the remainder of Medupi and Kusile units are shown in the Table 2 below.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

Table 2: Eskom New Build Dates Medupi Kusile Unit 6 Commercial Unit 1 Commercial Unit 5 Commercial Unit 2 Jun -20 Unit 4 Commercial Unit 3 May -20 Unit 3 Commercial Unit 4 Dec -20 Unit 2 Dec -19 Unit 5 Aug -21 Unit 1 Jun -20 Unit 6 Jun -22

3.2.2 Eskom Fleet Shut Down

The study assumes capacity reductions as Eskom’s coal-fired units reach the end of their 50-year life in alignment with the IRP 2019. These reductions result in the shutdown of a single unit at Arnot in 2021 and all the units of Camden between 2020 and 2023. Duvha Unit 3 was assumed to be unavailable throughout the study horizon; its capacity is netted from the total capacity sent out.

Capacity Shutdown of Eskom Coal Stations

6000

5000

4000

3000

Capacity [MW] Capacity 2000

1000

Month

Arnot Camden Hendrina Grootvlei Komati

Figure 3: Shutdown of Eskom coal-fired station units

The study further assumes that units at Hendrina, Komati and Grootvlei will be shut down when they reach their turbine dead-stop dates and it is no longer economical to carry out the maintenance required in terms of the Occupational Health and Safety Act (OHS Act). Although shutdown brings down a unit to zero power with the potential to be restarted in future if the necessary maintenance work is performed, this study assumes that none of the units return to service during the period under review. Currently 13 units have been shut down at these stations, removing 1 887 MW from the Eskom generation installed base. The cumulative capacity shutdown during the study period is depicted in Figure 3 above.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

3.2.3 Eskom’s existing and committed sent-out capacity

The total Eskom installed capacity consists of coal, nuclear, pumped storage, diesel, hydro, and wind, as shown in Figure 4 below. This capacity takes into account the assumed timing of the commissioning of new build stations and the shutting down of some units from older stations as detailed in Figure 3 above.

Eskom Sent-out Capacity

50000 48000 46000 44000 42000 40000 38000 36000 Capacity [MW] Capacity 34000 32000 30000

Month Coal Nuclear Pumped Storage Hydro Gas Wind

Figure 4: Eskom installed sent-out capacity (MW)

3.2.4 Eskom Plant Performance

3.2.4.1 Planned outages

The MTSAO 2018 study highlighted air-quality compliance as a risk to system adequacy; the MTSAO 2019 study has addressed the impact of air-quality compliance on plant availability. In terms of the National Environmental Management: Air Quality Act, 2004 (Act No. 39 of 2004) (NEMAQA), all of Eskom's coal-fired and liquid fuel-fired power stations are required to meet the Minimum Emission Standards (MES) under GNR 893 dated 22 November 2013, and as amended in GNR 1207 on 31 October 2018, which was promulgated in terms of section 21 of the NEMAQA. Most of the power stations do not comply with one or more of the standards; hence Eskom has committed to retrofitting the non-complying power stations within the MTSAO horizon.

Failure to fulfil these commitments on time if further postponements are not granted threatens Eskom’s licence to operate and is grounds for possible legal action against Eskom that could result in the withdrawal of licences and a directive to stop operating.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

This study hence considered two MES scenarios, namely:

• Minimum Emissions Standards Scenario 1 (MES 1) which incorporates the outages required at the power stations to comply with the emission limits in the licence by delivering the committed retrofit projects on time, as aligned with the dates committed to DEA. • Minimum Emissions Standards Scenario 2 (MES 2) which assumes that if Eskom is not ready to undertake the required retrofit outages on time as aligned with Government commitment, the affected units will be put into extended storage until retrofit outages can be done.

Both MES 1 and MES 2 scenarios increase the planned outages, which effectively reduce the system EAF.

3.2.4.2 Forced outages

Forced outage rate (FOR) includes both UCLF and OCLF. Forced outages were modelled as independent events; that is, one generator failing does not influence the failure of another. In addition, there was no certainty of a forced outage at any particular time; that is, several units may fail simultaneously, or there may be no failures at all. An annual rate of forced outages is defined for all plant and simulated by random occurrences of outages within the probabilistic Monte Carlo scheme. Since the occurrence of forced outages is not predetermined, it is a useful tool for testing the resilience of a power system against sudden unscheduled unavailability of generation resources.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

3.2.4.3 Partial load losses

The System Operator observed a trend of partial load losses typically increasing during peaking hours and decreasing during periods of low demand. Since load losses are categorised as capacity on forced outage in plant performance reporting, an explicit modelling of the pattern of partial load losses improves the assessment of system adequacy during the peak period. Historical data from 2015 to March 2019 was used to predict the pattern of partial load losses used in the study horizon.

3.2.4.4 Energy Availability Factor

The plant performance scenarios considered in the MTSAO 2019 are illustrated in Figure 5, namely MES 1 with high EAF, MES 1 with low EAF, and MES 2 with high EAF. The difference between the high and low EAF scenarios is the forced outage rate , the high EAF scenario has an average FOR of 17% and the low EAF scenario an average FOR of 25%.

The historical trend of the Eskom fleet EAF shows significant improvement from 71% in financial year 2016 to 77% in financial year 2017 and further to 78% in financial year 2018. However, a decreasing trend is observed in the Eskom EAF for financial year 2019, evident from the partial load losses that trend much higher towards the end of the 2018 calendar year. This trend of increasing partial; load losses continues into the beginning of financial year 2019 and is comparable to the trend seen in the same period between years 2016 and 2017. In addition to the above, the trend seen in financial year 2019 is similar to that seen in financial year 2016 when load shedding was inevitable.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

Plant Perfomance 90

70 EAF %

Year

Actuals MES 1 High EAF MES 2 High EAF MES 1 Low EAF Linear (Actuals)

Figure 5: Eskom Fleet Energy Availability Factor

3.3 Renewables from Independent Power Producers All contracted build REIPP capacity, from Bid Window 1 to Bid Window 4, have been considered as committed in the study and is shown in Table 3 below.

Table 3: REIPP Cumulative Capacity (MW) Technology 2019 2020 2021 2022 2023 Wind 1980 2512 3203 3343 3343 PV 1514 2287 2287 2287 2287 CSP 500 500 600 600 600 Hydro 14 19 19 19 19 Landfill 8 8 8 8 8 Other 0 25 25 25 25 Total 4016 5351 6142 6282 6282

Wind and solar PV plants are modelled using historical profiles from year 2013 to 2018, to create a median profile. For plants with history shorter than 2 years, area profiles are used. Using an area profile is in line with the conclusion of (Fraunhofer IWES, 2016) and substantiated by (REDIS, 2018) that wind speed and irradiation is correlated for plants in the same supply area. Eskom’s windfarm, Sere, is similarly modelled using the historical profiles.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

3.4 Existing non-Eskom installed capacity

Installed sent-out capacity of non-Eskom generation in South Africa is shown in Figure 6 below. The DMRE IPP peaking plants at Dedisa and Avon are included in the total for gas. For the purposes of the study, the capacity of 1100 MW from Cahora Bassa (imported hydropower from Mozambique) available to the South African system is based on the average 2018-2019 performance.

The energy produced by the non-Eskom plant, excluding Cahora Bassa and IPP Peakers, was limited to 10,9 TWh 1 per annum throughout the study horizon. Any reduction in this production would negatively affect the system adequacy outlook. The study assumes that non-Eskom generators not contracted to Eskom will continue generating for own use.

Figure 6: Capacities (MW) of non-Eskom plant

Due to the unavailability of data on non-Eskom plant performance, the MTSAO modelled typical plant performance based on plant of similar size and age.

3.5 Small-to-medium scale embedded generation (solar PV)

Rooftop PV installations are not yet licensed by NERSA and hence there’s no official data of the installed capacity. Current indications are that this capacity has been growing due to increasing electricity prices and load shedding. Various publications have been used to derive a reasonable estimate of currently installed and future projected rooftop PV capacity.

According to Power Quality Renewable Services (PQRS) 2016 study, approximately 154 MW of SSEG was installed in the 2015 calendar year. The estimated PV modules sold between January and December 2016 are estimated to about 120 MW. It is assumed that these modules would be installed within South Africa rather than be stored or exported beyond the RSA borders.

1 Source: NERSA 2018 actuals net energy sent out

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

The Green Cape Energy Services (GCES, 2018) Market Intelligence Report references another PQRS study which shows that PV modules sold between November 2016 and November 2017 amounted to approximately 110 MW. Based on the sales of PV modules between January and December 2018, AREP (AREP, 2019) estimates 200 MW of installed SSEG for the 2018 calendar year. This then brings the total amount of rooftop PV installed in South Africa to around 584 MW in 2018. AREP further estimates that between 244 and 300 MW of additional rooftop PV would be installed for the 2019 calendar year, depending on the severity of load shedding. Based on the DMRE position on SSEG, the MTSAO 2019 sets the threshold to 500 MW per annum which will be reached in 2023. Figure 7 below reflects the cumulative rooftop PV capacity assumed in the MTSAO 2019.

Rooftop PV 4000

3500

3000

2500

2000

1500

1000

500 INSTALLED(MW) CAPACITY 0 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 RooftopPV inst 154 274 384 584 884 1234 1634 2084 2584 3084 3584

Figure 7: Estimated Rooftop PV Capacity

3.6 Reserves requirements

Reserves are used to balance the system when unexpected events occur, such as cutomer demand fluctuations, changes in the availability of supply capacity, and generation variations from intermittent plant. The reserve requirements used in the study are depicted in Table 4 below and are as per the Ancillary Services Technical Requirements for the 2019/2020 to 2023/2024 report (SO, 2018). However, not all contracted Eskom coal fired station are able to provide the required reserves.

Demand response refers to loads that can be reduced on the instructions of the System Operator. Contracts for some of the instantaneous demand response required to arrest frequency will be expiring during the study period. Similarly, the supplementary demand response is contracted up to March 2020, but the MTSAO 2019 study assumes these loads will remain available as emergency resources for deployment by the System Operator.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

Table 4: Reserves Requirements (MW) Type Season Period 2019/20 2020/21 2021/22 2022/23 2023/24 Instantaneous Summer Peak 650 650 650 650 650 Off peak 850 850 850 850 850 Winter Peak 650 650 650 650 650 Off peak 850 850 850 850 850 Regulating Summer Peak 450 450 470 500 520 Off peak 450 450 470 500 520 Winter Peak 550 550 570 600 620 Off peak 550 550 570 600 620 Ten minute Summer Peak 1100 1100 1080 1050 1030 Off peak 900 900 880 850 830 Winter Peak 1000 1000 980 950 930 Off peak 800 800 780 750 730 Instantaneous Demand Summer 1000 1000 1000 1000 1000 Response Winter 800 800 800 800 800 Peak Load Supplemental Demand Reduction 300 300 300 300 300 Response Emergency Reserves 1900 1900 1900 1900 1900

4. STUDY CASES

The South African power sector is more susceptible to variations in the availability of Eskom plant, hence the study considered a potentially higher availability and a potentially lower availability. The MTSAO 2019 investigates the sensitivity of system adequacy to variations in plant performance, hence the three scenarios illustrated in Figure 8 below. The study considered three scenarios, namely:

Energy Forecast (1% AAGR) 50 year LOPP with turbine dead stop dates New Build CoD Contracted REIPPP dates Eskom and Non-Eskom Capacity

MES 1 MES 2

MES 1 High MES 1 Low MES 2 High EAF (~72%) EAF (~64%) EAF (~70%)

Figure 8: MTSAO Study Scenarios

• MES 1 with high EAF • MES 1 with low EAF • MES 2 with high EAF

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

Common parameters for all scenarios are the 50-year life of plant for coal-fired power stations, turbine shutdown dates for Grootvlei, Hendrina, and Komati, the forecast commercial operation dates for Eskom’s new build plants, Eskom and non-Eskom capacity, and contracted REIPP capacity and timing.

5. RESULTS

The study outcomes for the base-load capacity factor, unserved energy and the OCGT capacity factor for the period 2020 to 2024 for all three scenarios are depicted in Table 5. The system under the MES 1 high EAF indicate an adequate system except for the unserved energy metric, which is slightly violated. MES 2 high EAF violates the base load and unserved energy metrics while the OCGT usage remain below the threshold. The load shedding risk that the MES 1 and MES 2 high EAF scenarios pose can be managed operationally.

The MES 1 low EAF scenario indicates that if the system EAF deteriorates by increasing plant outages (planned or unplanned), there will be a shortage of energy reserves to cater for an energy-type contingency should it occur, and that the load shedding risk may not be manageable. The increasing energy forecast in the last years of the study indicates that, in order to grow electricity sales, either additional supply-side resources will be required or that the plant performance would have to be improved.

Table 5: Adequacy Metrics Results BASE LOAD CAPACITY 2020 2021 2022 2023 2024 FACTOR, % MES 1 High EAF 41 37 37 37 48 MES 1 Low EAF 64 69 74 75 74 MES 2 High EAF 53 59 59 37 48 50% Threshold 50 50 50 50 50

UNSERVED ENERGY, GWh 2020 2021 2022 2023 2024 MES 1 High EAF 17 56 47 23 168 MES 1 Low EAF 360 385 1 130 1 331 3 246 MES 2 High EAF 100 236 183 23 168 20 GWh Threshold 20 20 20 20 20

OCGT LOAD FACTOR, % 2020 2021 2022 2023 2024 MES 1 High EAF 3 2 2 4 6 MES 1 Low EAF 15 17 29 34 50 MES 2 High EAF 6 8 4 4 6 6% Threshold 6 6 6 6 6

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

6. POSSIBLE RISKS TO SYSTEM ADEQUACY

The risks identified by the MTSAO 2019 are similar to those reported in the MTSAO 2018; therefore, the progress in risk mitigation and control is provided.

• Insufficient plant maintenance due to either funding constraints or the unavailability of space to do maintenance would cause a further deterioration in plant performance beyond the scenarios tested.

o The performance of the base-load capacity continues to deteriorate, leading to load shedding. Though efforts have been made to manage the partial load losses, outage slips and boiler tube leaks, they remain a challenge. o Increase in planned outage requirements and unplanned outages beyond what was considered in the study would impact the system adequacy negatively.

• Earlier shutdown for economic or other reasons of Eskom and non-Eskom units. Such events would further reduce the capacity available to meet the demand.

o The Hendrina, Komati and Grootvlei turbine dead stop dates used in the 2018 MTSAO have been revised to later dates hence adding some capacity during the MTSAO 2019 study period.

• Insufficient primary energy at power stations:

o This risks is being managed. The Eskom 9 Point Plan reports that all power stations are above the Grid Code requirements and none of the stations monitored for wet coal are below 14 days’ strategic dry stock levels. o The load shedding events in 2019 have resulted in a significantly higher than expected OCGTs usage. To mitigate the risk of diesel shortages, the diesel procurement strategy has been developed and will be implemented.

• Failure to comply with air quality standards might lead to load losses or the shutdown of generation units or stations. Currently, some of the Eskom power stations do not comply with one or more of the minimum emission standards required by the National Environmental Management: Air Quality Act 39 of 2004.

o Compliance with minimum emission standards is a priority for Eskom and as such it has been incorporated in the outage plan used for the MTSAO 2019.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

• Switching back to using Eskom’s supply by non-Eskom generators.

o Generation by non-Eskom generators not contracted to Eskom remains the risk as it is not fully understood how and when they will be dispatched in the future. The MTSAO is building a database of historical data in collaboration with NERSA to better model non-Eskom generation in future.

• Eskom and IPP commissioning dates:

o Progress in bringing these plants to commercial operation is monitored continuously and the realistic commercial operation dates are used in the study.

7. CONCLUSION

The MTSAO 2018 concluded that the system would remain adequate at an EAF of 73% and above. The study furthermore indicated that a deteriorating plant performance or an increase in demand would have a negative impact on the system adequacy, which could only be restored by managing sales, adding new capacity, and/or delaying capacity shut downs.

The MTSAO 2019 indicates that the system is adequate at EAF above 72%. This is due to the lower energy forecast used than the one used in 2018, and the delayed shutdown of units at Grootvlei, Hendrina and Komati, which effectively mean a greater installed capacity base in the 2019 study than in the 2018 study. Although the study considers a cone of uncertainty with regard to plant performance and demand, excessive unplanned plant failures and/or an increase in electricity growth would affect the adequacy outlook of the system.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

8. APPENDIX: SYSTEM OPERATOR STATISTICS

Some of the monitored system reliability indices deemed relevant are reported in this section, based on actual System Operator year-to-date data, as at 30 September 2019:

8.1 OCGT Utilisation

Usage of the open-cycle gas turbine (OCGT) capacity factor includes Ankerlig (1 327 MW), Gourikwa (740 MW) and the DMRE OCGTs at Dedisa (335 MW) and Avon (670 MW). Figure 9 totals the calendar year to date OCGT utilisation of 7,2% for the year to date.

OCGT Capacity Factor 2019 YTD

30.00

25.00

20.00

15.00

10.00 Capacity Factor Factor %Capacity 5.00

Month

Avon Dedisa Ankerlig Gourikwa

Figure 9: Year-to-date OCGT utilisation

8.2 Frequency Control

Frequency incidents reveal cases where reserves were deployed. Paragraph 9 of the South African Grid Code: Version 9 stipulates the type (instantaneous, regulating) and capacity in MW required to restore the system, depending on the drop in the level of frequency. The actual incidents in Figure 10 below show that the system had sufficient reserves to recover from the disturbances.

The figure shows that the number of incidents of falling within the 49,5 < f < 49,7 band has increased by at least two-fold compared to the 2018 actuals. Although this band of operation is far from triggering under-frequency load shedding (which occurs when frequency drops to less than 49,2 Hz), an increased number of events may be cause for concern to the System Operator.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

Figure 10: Number of frequency incidents

8.3 Unserved Energy

The RSA interconnected power system (IPS) comprises sufficient installed generation capacity (about two-thirds is base-load capacity) compared to a typical summer demand of ~30 GW. Although Eskom (traditionally) carries out planned maintenance during summer, on any given summer day the supply-and- demand balance should not pose a challenge to the System Operator. However, the power system has had supply-side constraints, mainly due to the underperforming Eskom coal fleet, that resulted in load shedding in 2019.

Figure 11 below shows MWh estimates 2,3 removed from the system as at 30 September 2019. The values include load shedding and load curtailment (also referred to as manual load reduction) but excludes Interruption of Supply (IOS). IOS refers to all contracted and mandatory demand reductions to maintain system frequency and security of supply within acceptable bands.

2 Value are not audited and are meant for indicative purposes only. 3 Estimates of manual load reduction values in October 2019 are not available.

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

Actual Weekly Manual Load Reduction

300 251.21 251.21 250

200 169.95 169.95 158.08 158.08 150

Energy [GWh] Energy 100

50 16.43 ------08-Jul 22-Jul 07-Jan 21-Jan 10-Jun 24-Jun 01-Apr 15-Apr 29-Apr 02-Sep 16-Sep 30-Sep 04-Feb 18-Feb 05-Aug 19-Aug 04-Mar 18-Mar 13-May 27-May

Week

Figure 11: Estimated weekly load shedding

MEDIUM TERM SYSTEM ADEQUACY OUTLOOK 2019

9. REFERENCES

AREP. (2019). estimated growth for the solar PV sector for 2019.

Fraunhofer IWES, C. e. (2016). Wind and Solar PV Resource Aggregation Study for South Africa. CSIR.

GCES. (2018). Market Intelligence Report. Retrieved from https://www.greencape.co.za/assets/Uploads/GreenCape-Energy-Services-MIR-FINAL-08-05-2018- LR.pdf.

IRP. (2018). Draft Integrated Resource Plan. Department of Energy.

NERSA. (2014). The South African Grid Code: System Operator Code, Rev 9.0.

PQRS. (2016). Demystifying the total installed PV capacity for South Africa . Retrieved from https://pqrs.co.za/data/demystifying-the-total-installed-pv-capacity-for-south-africa- nov-2016.

REDIS. (2018). The Renewable Energy Data and Information Service . Retrieved from redis.energy.gov.za.

SO. (2018). System Operator Ancillary Services Technical Requirements .

StatsSA. (2019). Electricity generated and available for distribution. StatsSA.