International Journal of Fluid Machinery and Systems DOI: http://dx.doi.org/10.5293/IJFMS.2017.10.3.296 Vol. 10, No. 3, July-September 2017 ISSN (Online): 1882-9554
Original Paper
How to Avoid Severe Incidents at Hydropower Plants
Masashi Yasuda1 and Satoshi Watanabe2
1 Hydropower Department, JPower 15-1, Ginza 5-Chome, Chuo-ku, Tokyo, Japan [email protected] 2Department of Mechanical Engineering, Kyushu University 744 Motooka, Nishi-ku, Fukuoka, Japan [email protected]
Abstract
Hydropower is now changing its role from the energy generator into the most powerful and reliable tool for stabilizing the electrical network, especially under the increase of intermittent power sources like wind-power and solar-power. Although the hydropower plants are the most robust generating facilities, they are not immune from unexpected severe incidents having long downtime, considerable restoration cost and sometimes fatalities. The present paper provides some study results about severe incidents in the conventional hydropower plants, mainly about the flood, fire and electro- mechanical troubles, except for the incidents of civil facilities. It also provides some possible scenarios which may lead some measures how to avoid such incidents. Finally, it provides some comprehensible recommendations to avoid severe incidents based on experiences.
Keywords: hydropower plant, severe incident, flood, fire, turbine failure, generator failure
1. Introduction Hydropower is now changing its role from the renewable energy generator into the most powerful and reliable tool for stabilizing the electrical network, especially under the increase of intermittent power sources like wind-power and solar-power. Although the hydropower plants are the most robust generating facilities, they are not immune from unexpected severe incidents having long downtime, considerable restoration cost and sometimes fatalities. JPower is the largest wholesale power company in Japan, having a hydropower fleet comprising from 53 conventional hydropower plants (hereinafter it is abbreviated as HPP) of 3,600 MW and 7 pumped storage plants (hereinafter it is abbreviated as PSPP) of 4,970 MW in Japan and participating the operations of one PSPP (720MW) and a few HPPs in overseas. JPower started the hydropower developments from the early 1950’s and provided many technical consulting services for the overseas hydropower projects from 1960’s. Figure 1 shows some pictures of JPower’s hydropower plants.
Note: From left to right of the first row; Sakuma Dam, Kuttari Dam and Oku-tadami HPP
From left to right of the second row; Numappara PSPP, Okukiyotsu PSPP and Kalayaan PSPP (Philippines)
Fig. 1 Some hydropower plants of JPower
Received August 30 2016; accepted for publication April 24 2017: Review conducted by Young-Seok Choi. (Paper number O16029K) Corresponding author: Masashi Yasuda, [email protected]
296 The present paper provides some study results about severe incidents in the conventional hydropower plants, which have been experienced by JPower as well as the collected one from the publicized papers. It also provides some possible scenarios, which may provide some measures how to avoid such incidents. Finally, it provides some comprehensible recommendations and prospects to avoid severe incidents based on the studies. It is an objective of this paper to provide some practical approaches and recommendations to avoid any future and similar incidents. Since the authors are not the expert of civil engineering, so that the topics of dam, gate and waterway are excluded from this paper. In addition, the severe machinery incidents of pumped storage plants had been already publicized in IAHR 2016 [1], so that the subjects of the present paper are mainly limited to the flood, fire and machinery incidents in conventional hydropower plants.
2. General tendencies of incidents in Hydropower plants Figure 2 shows the probability of machine troubles per year and per one unit for each power plant type, which was obtained after analyzing 16,500 incidents, occurred at about 1,800 generating units in Japanese hydropower plants from 1995 to 2004 [2]. For an example, it indicates that the probability ratio of turbine failure in pumped storage type is about 0.33, which means that a pump- turbine has one failure per three (3) year’s operation averagely. The total ratios are around 1 to 2 for each hydropower type, i.e. the average Japanese hydropower plant has one (1) or two (2) machine troubles per year and per unit, therefore it may be said that the Japanese hydropower plants are generally well-maintained with minimal failures. Although this figure shows only machine failures, it indicates that the probabilities of failures are dependent on both the equipment and the hydropower types. The run of river type has the lowest probabilities of failures because of the continuous operation with less frequent start-stops, meanwhile, the pumped storage has the highest ratio due to its frequent start/stops and complicated configurations.
Fig. 2 Probabilities of failures for each power plant type [2]
Table 1 shows the distribution of 985 incidents occurred in around 1,260 Japanese hydropower plants from 2004 to 2012 [3], which are comparatively serious incidents to be reported to the Government due to a law. It is very obvious that a half of incident is related to the flood. Another main cause is the maintenance troubles, mainly due to the degraded machine by the longtime operation. This table has no distinction between the HPPs and PSPPs, so that it shows the general tendency of HPPs due to the majority.
Table 1 Distribution of incidents in Japan from 2004 to 2012 [3] (in percentage) Design Maintenance Other Natural Flood Earthquake Others Total Problem problem Disaster Civil facilities 0.3 4.6 6.9 2.2 1.7 1.7 17.5 Turbine 1.8 8.6 1.2 0.0 0.7 2.2 14.6 Generator 1.8 4.7 1.6 0.0 0.5 1.6 10.3 Main Circuit Eq. 1.5 1.8 16.6 0.4 1.4 1.4 23.2 Station Service Eq. 0.3 1.4 2.9 0.0 0.6 0.8 6.1 Control Eq. 0.2 2.5 18.7 0.2 0.3 1.1 23.0 Others 0.4 1.6 0.6 0.4 0.8 1.4 5.3 Total 6.4 25.3 48.6 3.2 6.1 10.4 100.0
Figure 3 shows the tendency of 1,687 incidents recorded in the same reports [3] for Table 1 from 1988 to 2012. It is clear that the number of incidents due to natural disasters is increasing year by year. The maintenance failures like the degradation of machines are also increasing. Figure 4 shows the breakdown of the natural disasters. It is obvious that the flood is dominant.
297 500 Landslide Storm Snow and Ice and 4% 1% 400 Avalanche 4% Lightning 300 4% Earthquake 200 5% 100 Number of Number Incidents
0 1988-1992 1993-1997 1998-2002 2003-2007 2008-2012 Flood 82% Era
Design Failure Maintenance Failure Natural Disaster Others
Fig. 3 Tendencies of various incidents [3] Fig. 4 Breakdown of natural disasters [3]
Figure 5 shows the incident number for each range of the downtime by analyzing 1,300 incidents occurred in 109 generating units, comprising of 85 generating units in fifty-two (52) HPPs and 24 units in seven (7) PSPPs, from 2004 to 2015. Figure 6 shows the accumulated downtime for each range from the same data of Figure 5. Although the outages longer than one month were merely one percent of all incidents in the number, those severe incidents occupied sixty-two (62) percent of the total downtime. It means that the majority of total downtime is due to the rare but serious incidents, not due to the daily, frequent and minor ones. That is why the severe incidents should be reduced to achieve better maintenance of the plant. Meanwhile, another view may be possible for those figures, like the current good maintenance is keeping almost all the incidents within the minimal outage hours, so that only unexpected incidents sporadically happen and they are inevitable. However, if we want to achieve a higher maintenance level, it is necessary to reduce the severe incidents into minimal level.
10 days to More Less than 3 hour to 1 month than 1 3 hour 1 day 2% month 1% 6% 1 day to 1% 1 day to 10 days 10 days 16% Less than 15% 3 hour More More 41% than 1 than 10 3 hour to month days 1 day 62% 16% 40%
Fig. 5 Breakdown of number of incidents Fig. 6 Breakdown of total downtimes
Table 2 shows the distributions of severe incidents, picking out the serious mechanical breakdown and/or severe incidents having considerable downtime more than 10 days, from the same data of Fig.5 and Fig.6. Table 2(a) shows the records of HPPs and Table 2(b) shows that of PSPPs. The majorities of severe incidents in the conventional hydropower plants are also floods and degraded machine troubles. The typical flood damages were the clogging of tailrace outlet and/or intake with mud and debris. Machine troubles happened by various causes, including the oil leakages from Bulb runner hub, oil leakage and water intrusion in bearing oil reservoir, loosen terminal bolts at circuit breakers and generators, detachment of counter weight of intake gate, etc.
Table 2 Distribution of severe incidents in hydropower plants (in percentage) (a) Conventional Hydropower plants (b) Pumped Storage Plants Design/ Maintenance/ Natural Design/ Maintenance/ Natural Total Total Manufacturing Deterioration Disaster Manufacturing Deterioration Disaster Civil facilities 0.0 8.3 33.3 41.7 0.0 0.0 0.0 0.0 Turbine 0.0 19.4 8.3 27.8 0.0 20.0 0.0 20.0 Generator 2.8 8.3 0.0 11.1 10.0 40.0 0.0 50.0 Main Circuit 2.8 8.3 0.0 11.1 10.0 0.0 10.0 20.0 Station Service Circuit 0.0 5.6 2.8 8.3 0.0 0.0 0.0 0.0 Control Eq. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Pump starting Eq. - - - - 0.0 10.0 0.0 10.0 Others 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total 5.6 50.0 44.4 100.0 20.0 70.0 10.0 100.0
298 From above-mentioned figures and tables, there are some general tendencies for the incidents in Japanese hydropower plants as follows; Ÿ The severe incidents dominate the total outage hours. Ÿ The flood incidents are increasing year by year due to the increase of heavy rainy days. Ÿ The failures of degraded turbine, generator and civil facilities are increasing due to the increase of aged plants. Ÿ Although the fire incidents are minimal at the present, the future increase is concerned due to the increase of aged plants. Ÿ The pumped storage has a unique tendency and it should be distinguished from the conventional hydropower plant.
3. Flood incidents In Japan, there are two heavy rainy seasons, one is a monsoon season in June and July, and another is a typhoon season in September and October. There is a tendency of increase of heavy rainy days as indicated in Fig. 7, which counts the number of heavy rain more than 50mm in one hour observed at around 1,300 robotic rainy stations. Figure 8 graphs the actual flood discharges versus the design flood discharges, which indicates the present flood risk is higher than that of the commissioning time at many existing hydropower plants. Due to the increase of heavy rainy days, Japanese utilities experienced some severe flood incidents of hydropower plants in the last 20 years as shown in Table 3.
⊿N=+21.5/10 years
There are many sites where the actual flood discharges
had been bigger than the
/s) 3 original design values. (m
numbers per year per numbers ofHPPs
Experienced flood Experienced discharge Observatory
Design flood discharge of HPPs (m3/s) Year Fig. 7 Observatory number of heavy rainy days [4] Fig. 8 Relationship between design and actual flood [5]
Table 3 Some big floods in Japan No. Date River system Affected power plants Outline of damage 1 September Kumano river in A run-of river HPP of The plant was completely demolished by “Tsunami” caused by 2011 Nara Pref. 15MW multiple landslides generated by Typhoon12. 2 July Tadami river in 23 HPPs having the total Unprecedented discharge more than 6000m3/s, nearly same with 2011 Fukushima Pref. output of 1400 MW, the design flood discharge, attacked the biggest hydropower and Agano river in including one PSPP center in Japan. Many HPPs were inoperable due to plant Niigata Pref. submergence, powerhouse damage and accumulation of debris at intake and tailrace. 3 September Mimi river in 4 HPPs having the total Those plants were attacked by Typhoon 14 and inoperable due 2005 Miyazaki Pref. output of 224MW to the flooding of plants and accumulation of debris at tailraces. One HPP had a transformer fire as the secondary damage. 4 October Jintsu river in 4 HPPs having the total Typhoon 23 caused severe flood. One HPP was submerged until 2004 Toyama Pref. output of 168MW the top of generator. 5 July Kurobe river in 2 HPPs having the total One underground HPP was completely submerged. The 1995 Toyama Pref. output of 146 MW accumulation of debris at tailrace caused the long outage.
The flood may make the plant inoperable with the submergence of outdoor switchyard equipment and control boards in powerhouse as indicated in Table 1, but also it may cause the closure of intake and tailrace outlet with accumulation of debris, mud and boulders. A JPower’s hydropower plant (See No.2 in Table 3) had a downtime of 3-years due to the difficulty of removal of debris from the tailrace outlet. Another severe risk is the landslide since it sometimes directly damages the waterways, outdoor switchyard and other plant facilities on a ground. It also blocks the main stream of river and makes a natural dam, which may cause an excessive water-rise and may submerge the plant. The multiple landslides are more disastrous since “Tsunami” would be generated. One example (See No.1) was the most devastated one [6]. The mechanism of “Tsunami” is assumed like, one landslide blocked the main river course at the downstream of plant and made a natural reservoir which reached the plant, then the second landslide entered into the reservoir causing a “Tsunami” and the plant was completely washed away by the impact. Some combinations like the earthquake and heavy rain should be considered, typical example is very severe damage of two hydropower plants in Taiwan in 2001 and 2004. As the central part of Taiwan had a big earthquake, Chi-Chi earthquake, having the magnitude 7.6 on the Richter scale in September 1999, many mountainous areas near the epicenter suffered serious ground breaks, hillside collapses and landslides. After the earthquake, many big typhoons attacked those areas and the two plants were heavily damaged by landslides at the tailrace waterways and outdoor switchyards plus the submergence of underground powerhouses [7]. Figure 9 summarizes the possible scenarios of flood. Some measures to prevent or stop the progress are indicated by ①-⑧, but
299 not limited to. This figure is provided for helping an easier understanding of the possible severe incidents. The actual processes to reach the final flood incidents will be far complicated together with more various causes and affected factors which are omitted in the figure, but still this simplified figure will be helpful to get the comprehensive views of flood and to consider some measures to stop the progress. Those figures of following chapters for discussing fire and machinery incidents are also presented for the same purpose.
Primary Causes Vulnerable Parts Secondary Causes Final Incidents Rise of river ① Heavy rain Many openings at Big inflow into water level switchyard, entrance, etc. *1 powerhouse ① Closure of all openings Temporary openings at ② ② Preparation of ample drainage Eq. Natural construction stage Flooding of ③ Use of Hazard map Landslide dam ③ Powerhouse Ensure the Safety of staff Multiple landslides Tsunami ④ Periodical clean-up of drain Loss or short ⑤ Early abandon of operation Floating Drain gutters and ④ Clogging of drainage and closure of gates debris drain holes with debris capability Use of Timeline Chart ⑥ Ensure the back-up power Tailrace Earthquake Continuous ⑤ Closure by mud, ⑦ Higher location of control Eq. operation of outlet & & Volcano boulders, debris Periodical check of drainage pump turbine Intake eruption ⑧ ⑧ Provision of Jet-pump Blackout ⑥ Drainage Eq. like Short-circuit of pump, water level motor and detector, control Submergence of Drainage Eq. ⑦ control Eq. *1 circuits, etc.
Fig. 9 Possible scenarios of flood in hydropower plants
4. Fire incidents Fire is the most critical incident both in lives and property damage. Fortunately, the probability of fire in the hydropower plants is not so high if compared with other industrial plants, but the hydropower plants are not completely immune from fire. There are three main categories for the fire incidents in hydropower plants like a powerhouse fire, a fire of oil-immersed equipment and a generator fire. The generator fire will be discussed in later as the generator incidents. Table 4 shows some examples of powerhouse fire.
Table 4 Examples of powerhouse fire No. Date Affected power plants Outline of damage 1 December A HPP of JPower’s A cable fire occurred in the dam cable shaft and damaged the dam control room. 2015 82MW in Japan The fire source seemed to be a deteriorated cable, which had been already out of use but still in voltage-charged. 2 November A PSPP of 84MW in A 125V DC cable laid under the control room caught a fire that expanded the 2012 USA whole control room. The definite cause was not found. [8] 3 June A HPP of 100MW in The mal-specified lightning surge arrester was installed and caused a grounding 2007 USA fault. The operator carelessly recharged the failed circuit without checking the breakdown and it caused a fire. [9] 4 September A HPP of 175MW in A 480 V power cable had grounding failure due to the rubbing with a grating step. 2002 USA The failure caused a fire in cable shaft between the powerhouse in dam and control building at ground level. Control equipment was severely damaged. [10] 5 1999 A HPP of 140MW in A lighting fixture in the portal building had a fire that spread to the ceiling and Norway walls. [11] 6 June A PSPP of JPower’s A circuit breaker of pony motor had failed to open the circuit due to a looseness 1980 1000MW in Japan of connecting bolt of a conductor. Cubicle fire occurred and expanded to the cables running over the cubicle.
One main cause of powerhouse fire is the deteriorated low-voltage power cable, which has been used for more than 50 years. Those cables may have many cracks, scratches, rubbing and dust accumulation at the sheath and insulation layer, and the insulating property may be seriously degraded under the wet condition. In case of a fire incident in a JPower’s hydropower plant (See No.1 of Table 4), the fire source was a degraded cable, which had been already out of use but still in charged with 200V AC. The most likely scenario of the fire is that the creeping current is generated on the surface of cable insulation layer and it causes a small-carbonized portion under the combination of cracked sheath, deteriorated insulation layer, dust and water. Larger carbonized portion generates bigger creeping current; consequently, the degradation is accelerated by many small arcs at the carbonized area. An ignition of fire must be caused by the electrical fault or overheating, but the fault current is too small and intermittent to activate the protective function of no-fuse-breaker; thus resulting in a fire. A cable fire is easily spread in all directions through the cable trays, especially at the cable concentrated area like the vertical cable shaft and cable treatment room under the control room. The poisonous gas by burning of cables endangers the lives of staffs
300 and workers. The soot containing micro-metal particles will easily damage all facilities in the powerhouse; especially it is fatal to any computer-based equipment. The cubicle fire is also dangerous. Since the fault current is generally big due to the high voltage, the generation of arc can damage the equipment completely and its spattered hot metal fume may cause the fire of nearby cables and other cubicles. Sometimes the human-error and/or inability of protection relay action cause the expansion of fire. The cubicle fire in JPower’s PSPP (See No.6 of Table 4) was potentially the most dangerous incident for us. In this case, both the DC and AC power supply were completely lost and the generator-motor continued its rotation around two-hours under the low-load motoring without means of stop and cooling. Fortunately, there was no severe damage to the generator-motor.
Table 5 shows some examples of oil-immersed equipment fire. Those fires are very dangerous by the potential oil-mist explosion as the secondary damage, especially in the underground powerhouse. The junction box of oil-filled (OF) cable has a high risk of electric failure due to the high voltage stress and mechanical stress by the heat cycle. The step-up transformer has a big risk of explosion due to its big volume of oil. In 1992, a PSPP in Taiwan (See No. 2 of Table 5) had an explosion of a main transformer in the underground transformer cavern. As a lightning directly struck a transmission line tower nearest to the plant and its voltage surge exceeded the insulation strength of winding system, the transformer had an inner fault together with a rupture of tank, resulting in an oil-mist explosion. In old time, there were some severe explosions of high-pressure oil equipment. Although those incidents have been considerably reduced, there are still some examples of oil eruption from the pressure oil tank and piping. If the hot works are conducted nearby, the risk of fire is considerable.
Figure 10 summarizes the possible scenarios of fire in hydropower plants together with the preventive measures like ① - ⑩.
Table 5 Examples of oil-immersed equipment fire No. Date Affected power Outline of damage plants 1 December A HPP of 317MW A junction box of oil-filled (OF) cable was burned out due to the grounding failure. 1995/ April in Indonesia A similar incident occurred at the junction box of OF cable for another unit in 2000. 2000 2 July 1992 A PSPP of 1602MW A generator transformer had exploded by the lightning surge. One fatality, a in Taiwan JPower’s engineer, was counted. 3 March 1989 A HPP of 900MW Cable joints of OF cables in the cable shaft had a fire and it totally destroyed the in Zambia switchyard and administration building. [11] 4 1988 A PSPP of 1000MW A junction box of OF cable with the main transformer was exploded after a in Italy grounding fault. [11] 5 July 1981 A HPP of 4215MW A 525kV OF cable had a fire in the cable tunnel. [12] in USA 6 1973 A HPP of 760MW A junction box of OF cable was destroyed by flash-over and oil mist was filled in in Norway transformer room, generator room and access tunnel, then explosion occurred. Three fatalities were counted. [11]
Primary Causes Vulnerable Parts Secondary Causes Final Incidents ① Periodical check Replacement of degraded cable Improper LV ・ Overheating Delay of Cable Powerhouse connection ① Removal of unused cables Power ・ Short-circuit detection Fire Fire / Contaminants cables ・ Ground fault and trip ⑧ ② Periodical check ③ Periodical check including Secondary ・ Daily Heat cycle Cubicle Resolved gas analysis ・ Mechanical/ Failure of tripping oil-mist ② Switchgears Fire ④ Backup arrester beside the M.Tr Electrical stresses like circuit by operating explosion ⑦ ⑤ breaker mechanism troubles, Preparation of fire extinguishers Design and excess of operating Arc ⑨ Confirmation of safety procedure duties, etc. manufacturing generation ⑥ Ensure the reliability of protection Rupture defects ③ ⑥ ⑦ Ensure the backup protection Oil-filled equipment of oil Inner fault by short ⑧ Close the cable penetration of wall Excessive like M.Tr and OF cable tank ④ circuit, ground fault, with anti-fire materials lightning surge or overheating, etc. over-current Ensure the escape routes ⑨ Provide the anti-blast wall Pressure oil equipment Catch fire with leaked Human error ⑤ ⑩ Loss of Dissipation of blast /Lubricating oil of bearings oil, erupted oil and mist, at hot works lives ⑩ Ensure the escape routes /Painting work vaporized inflammable fume and dust
Fig. 10 Possible scenarios of fire in hydropower plants
301 5. Machine troubles In Japan, severe machine troubles in the conventional hydropower plants are generally very few. In the comprehensive survey [2] from 1995 to 2004, 15 machinery incidents were counted having a downtime more than one month, but only two (2) cases were caused by the pure machine trouble; one was a failure of turbine guide bearing and another was a failure of thrust bearing. The others were all caused by floods. Nevertheless, there are some severe incidents reported for overseas plants, so that it is better to mention such examples for studying the severe machine troubles. 5.1 Turbine troubles Table 6 shows some typical examples of severe turbine incidents collected from publicized papers.
Table 6 Examples of severe incidents of turbine excluding pump-turbine No. Date Affected power plants Outline of damages 1 June Six Bulb turbines of 30MW Many cracks were found at the flange fillet of main shaft at runner side. The cause 2010 per unit in Canada was the Stress Corrosion Fatigue cracking. Similar incidents were reported for a power plant of Rumania and Serbia. [13] 2 August A HPP of 6400MW in The head cover of a turbine, 640MW Francis, broke away and the powerhouse 2009 Russia was totally destroyed together with 75 fatalities. [14] 3 March A Francis turbine of 261MW The runner blades had severe cracks and split portions due to the high vibration 2008 in Canada caused by guide vane cascade failure.[15] 4 February A Francis turbine of 88MW The runner blades had severe damage by hard-hitting with a freed guide vane. The 2008 in USA cause was the dropping of a link pin of the guide vane operating mechanism. [16] 5 2006 Three Francis turbines of Many cracks were found at the trailing edges of turbine blade outlet. The cause 200MW per unit in Iran was the vibration excited by Von Karman Vortex Street. [17] 6 September Three Francis turbine of Severe silt erosions were found at the runner blades, guide vanes (GV), GV bushes 2003 48MW per unit in Nepal and GV facing plates after 1,000-hours operation. 7 June A Francis turbine of 330MW Severe runner cracks were found in two runners. Those cracks were generated by 2000 in China the severe vibration at the start/stops plus the generation of Von Karman Vortex Street. [18] 8 March A Kaplan turbine of 109MW The head cover of a turbine broke away and the generator had a fire by jamming 1992 in Canada of the rotor with stator. [14] 9 May A Francis turbine of 150 A spiral case failed by the excessive pressure rise due to the instant shut-down of 1990 MW in Australia all guide vanes. A steel bar was trapped in the runner and it destroyed all shear pins of guide vanes. [19]
One of typical severe incidents for the turbine is the generation of cracks at runner and stay vane. It is well-known that the Francis turbine has cracks at the trailing edge of runner blade near the junction with crown and band. Those areas have high static and dynamic stresses, especially at the speed-no-load during start-stops (See No.7 of Table 6). Another cause is the generation of Von Karman Vortex Street (See No.5) at the trailing edge of runner blade. We also experienced small cracks for the 18.2 MW Francis turbine runners in Philippines by the Karman Vortex. Since many Francis runners have far thinner blades than the past to achieve higher efficiency and they are requested to conduct far frequent start/stops and low-load operation, those factors seem to be related with the crack generation. Adjustable blade runner of Kaplan, Bulb and Deriaz turbines may have cracks at the runner blade near its stem. The Pelton runner has a weakness at the bucket base due to the alternating impacts from water jets. Each turbine type has characteristic weak points for the crack. The stay vane cracks occur in many large and low-head Francis and Kaplan having tall and slim stay vane [20]. The main cause is attributed to the Von Karman Vortex Street. In Japan, those cracks are not so familiar due to the small machine sizes, meanwhile, old casted stay vanes, which had been manufactured during/post World War 2 with poor quality control having many casting defects, have low-cycle fatigue cracking after a longtime operation. It is one of main reasons for the full refurbishment of old turbines. The horizontal-shaft machine has the stress corrosion fatigue at the main shaft (See No.1) under the stress concentration and wet condition. We also experienced similar cracks in our Bulb turbines, but not so serious ones fortunately. The silt erosion causes severe damage at the runner vanes, guide vanes and their surrounding parts. Some countermeasures like hard or soft coating techniques are effective, but still under the development. JPower provided a technical service for a hydropower plant in Himalayan region (See case No.6), which had many severe erosions in runner and guide vanes. In this case, the inlet valve failed to open due to the large water leakage from guide vanes only after 1,000-hours operation. Those silt problems should be carefully studied from the initial design stage of plant and turbine, including the effective de-silt facilities, smooth flow shapes of turbine parts, application of durable materials, anti-erosion coatings, etc. The guide vane cascade failures are reported from some plants (See No. 3 and 4). We also experienced a failure of turbine guide bearing after a simultaneous closing of two guide-vanes, where some fatigue signs were found at the broken shear pins in a post- failure inspection. The unprecedented failure by the rupture of a head cover occurred in Russia (See No.2). Similar incident seemed to happen in a Canadian plant in 1992 (See case No.8). A careful attention should be paid to any high vibration of machine and integrity of major bolt connection. Many instability phenomena are reported mainly at the large machines in overseas projects, including high vibrations of penstocks, powerhouse floors and sometimes power-swing. Those flow-induced vibrations have many features and symptoms, so to find the root-cause of vibration is necessary. A rupture of steel penstock [21] in a French small plant is reported due to the self-
302 oscillation of spherical valve seal. It is a very rare incident, but the potential risk should be acknowledged. Some social and environmental problems may cause some operational restrictions and considerable downtimes. One example is the oil-spill from the packing of runner blade stem of adjustable-blade turbine. In Japan, it is strictly requested to keep a good river environment, so JPower stopped a Bulb turbine during a considerable period to check the oil leakage when a minimal oil film was found on the water surface of tailrace. The turbine was finally modified into the water-filled hub to eliminate such incident completely. Figure 11 summarizes the possible scenarios of severe incidents of turbine with some preventive measures like ①-⑩.
Primary Causes Vulnerable Parts Secondary Causes Final Incidents ① Good design by CFD analysis Turbine parts, Serious vibration, High flow-induced ① ② ② Good design by CFD and FEM Draft tube swirl building floor, shocks and pulsation & vibration ④ ③ High quality control at manufacturing penstock, generator power swing ④ ④ Air injection High dynamic ② Runner vane ⑤ Francis runner: Crack at trailing edge of blade ⑤ Periodical check including stress / Stay vane Kaplan & Bub runner: Crack at blade near stem Non-Destructive Examination (NDE) Pelton runner: Crack at bucket base Repair at incipient crack Defects at ③ Pelton runner Stay-vane: Crack at joints with stay-ring Apply the shot-peening manufacturing Suitable trailing edge shape of RV/SV Main shaft of horizontal ⑥ Stress corrosion Serious ⑥ Avoidance of stress concentration Alternating stress turbine like Bulb turbine fatigue cracking damages of Avoidance of wet condition runner, guide Periodical inspection and repair Guide vane operating Fatigue cracks Guide vane High vibration ⑧ ⑦ vane and stay ⑦ Periodical check of shear pin mechanism like shear /Loosen bolts cascade vane To provide stopper or brake pins and pins failure ⑧ Survey of vibration causes Improper Periodical check by NDE maintenance Major base and connecting bolts Rupture of head cover, drop of head cover, bottom ring liner, of runner cone, detachment Periodical check of loosen bolts ⑧ runner seals, runner cones, etc. of replaceable seal rings ⑨ Enough capacity of silt settling basin High silt Design of less-erosion shape contaminated Runner vane, Guide ⑩ Serious erosion of leading edge of runner Apply of anti-erosion technology river flow ⑨ vane, inlet valve blade , runner seal, fillets of guide vane, ⑩ Periodical inspection and repair GV bushes, inlet valve seal, etc. Stoppage at high-silt occasion
Fig. 11 Possible scenarios of severe troubles of turbine 5.2 Generator troubles Generator has many troubles in the stator, rotor, bearings, exciter, etc. The serious generator failures are the electric breakdowns of stator and rotor windings as well as the mechanical breakdowns of bearings, stator and rotor components, etc. A generator fire, the most dangerous incident, may be caused as the secondary incidents consequent to those serious failures. However, fortunately in Japan, such incident is minimal due to the renewal of old stator windings with epoxy-resin insulated coils, high reliability of digital protection relay and various periodical inspection of generator. Since we do not have suitable examples for discussing the item, the examples of CIGRE’s survey [22] about the generator fires would be suitable for this topic. Although the details are not available, it shows 64 incidents in total, including 39 stator faults, 8 rotor faults and 5 bearing faults. Table 7 shows some typical examples about the causes and damage of generator fires.
Table 7 Examples of generator fire [22] No. Main causes Outline of causes and damage 1 Stator Fire was caused by defects of stator windings that have developed into phase to phase short- circuit faults. Damage was confined to a section of the bottom end winding, but it took 18-month for the restoration works together with new core and windings. 2 Stator Ground failure due to insulation fault of stator coils damaged the stator windings and the core laminations with the presence of molten copper. 3 Stator Failure occurred in soft soldered joint. Approx. 17 % of the stator winding had to be replaced. 4 Rotor Excitation connections to the generator failed, causing the leads to "flap" free and sheared off a large proportion of the end windings, resulting in a generator fire. The 2/3rd of the generator stator windings were damaged and had to be replaced. 5 Rotor The insulation part of rotor pole was broken. Insulation of stator winding, magnetic core and rotor pole were damaged. 6 Rotor The failure was caused since a pole to pole connection got loose. 7 Rotor/ A circuit breaker failure caused circulating currents in the rotor, the temperature rose and set fire Circuit breaker to the cover plates of fiberglass for air deflection, causing a fire. It damaged the upper stator winding endings and the cover plates. 8 Bearing/ After the oil circulation failure of thrust bearing, the main circuit breaker was not opened even Circuit breaker after relay's high temperature metal operation. 9 Circuit breaker There was circuit breaker failure after several relays operations and the machine operated like Exciter synchronous motor. The failure of the exciter system was consequence of the main circuit breaker failure. The rotor was totally damaged and the stator partially damaged.
303 10 Bearing/ Bolt from lubrication system was broken and fall inside the generator during the testing of Human error generator. 11 Human error An item of steel was left behind in the generator enclosure following routine maintenance. The item caused an electrical fault in the stator, resulting in a generator fire. Damage was minimal, but required a significant clean-up effort. 12 Brake Brake clamp came loose and hit the stator winding resulting in a phase to ground fault.
There are many causes for generator fires including the short-circuit or earth faults of stator coils, failures of circuit breaker, jamming with detached parts from rotor, human error, etc. Anyway, the old stator windings using the thermoplastic insulation system like Bitumen should be treated with a special care for a fire. Another survey of CIGRE [23] shows root-causes such as aging and contamination of windings and internal partial discharges for the insulation failures, fatigue of materials and loosening of rotor parts for mechanical failures, etc. It is worth to remind that the human errors may sometimes cause severe incidents, especially for the generator as like cases of No. 10 and No. 11 in Table 7. There is one similar example in Japan although it is for a PSPP. The incident was caused by the mal- wiring of parallel-in circuit of generator circuit breaker when a renewal work was conducted for the programmable controller (PLC). Because the generator circuit breaker was wrongly closed by a “Open” signal from PLC, the generator-motor in stand-still started its rotation as an induction motor getting a power from the electric power system, so the damper coils of rotor were severely damaged by the over-currents, consequently damaged the stator windings. Figure 12 summarizes the possible scenario of generator fires together with some preventive measures like ①-⑨.
Primary Causes Vulnerable Parts Secondary Causes Final Incidents ① Stator coil system Shrink of coil Looseness of Coil Coil abrasion ① Proper design of coil (especially thermo- insulation coil wedge vibration (corona shield) plastic type) Periodical check of looseness Contamination of winding ② Temperature rise ② Prevention of oil-mist with oil, water and dust /Partial discharge Aging of Collecting carbon-dust windings Daily start/stops, Periodical clean-up heat cycles and Rotor-pole connecting terminals, Loss of stator ④ ③ ③ Diagnosis of coil deterioration vibration exciter lead, pole wedge, etc. parallel circuit Over- Partial discharge analysis /Motoring heating Replace with epoxy-resin coils ・ Overheating and melting Avoidance of high operating temp. Detachment Various stator ・ Looseness of bolts ④ Provision of thermal protection parts (Connecting of parts ・ Low or High-cycle fatigue ⑤ *1 ⑤ Periodical inspection terminals, Core Vibration monitoring end plates, Brake Short-circuit or Provisional Jamming desks, etc.) Item left in ⑥ Confirmation of no left items materials & with stator ground fault of generator ⑥ Check of noise at slow rotation tools stator windings Improper Vibration monitoring ⑦ Damage ⑦ Periodical check of synchronizer maintenance/ Asynchronized Synchronizer of stator Test parallel-in with no-load Human error parallel-in Inability of *1 ⑧ elimination of Periodical test of characteristics Generator circuit ⑧ Failure of circuit continuous fault Replace with new circuit breaker Design problem breaker breaker action at current ⑨ Periodical test of characteristics generator fault Ensure the backup protection Mal-setting of Generator ⑨ No detection of Replace with digital relays protection relay protection relay generator failure Generator Fire Confirmation of no blind spot
Fig. 12 Possible scenarios of severe troubles of generator
6. Lessons learned and some recommendations For a flood incident, it should be acknowledged that the present heavy rain may exceed the past records, so it should be paid higher attention to the risk of submergence at many hydropower plants. In addition to the submergence, the risks of landslide and closure of intake/outlet are also very high. For the submergence, the practical measure is closing of all openings which may permit the water intrusion into the powerhouse, like cable holes in switchyard, as much as possible. The air ventilation openings should be located at safer position from flood and drainage pipe outlet should be provided with check (non-return) valve. Temporary sand bags are effective to keep the water from entering at entrance. The proper functions of drainage facilities should be ensured. Periodical inspection of drainage pumps and its control equipment are indispensable. The motor control center of drainage pumps should be located in a higher floor and water level gauge should be a water-proof type. The increase of drainage capability may be necessary if the leakage inflow to powerhouse is increasing year by year under the recent heavy rain. A jet pump is reliable to ensure the drainage capability even under a blackout. For the landslides, to keep the staff’s safety will have the highest priority. The hazard- map should be utilized to assess the risk of landslides, road blockade, submergence and other perils in advance. When dispatching the staffs to the plant under any heavy rain, two automobiles should be allocated to ensure the mutual support by each other. For the closures of intake/outlet, it is necessary to avoid the excessive accumulation of mud and debris at the intake and tailrace outlet, therefore, it is recommended to abandon the generating operation and to close the gates when the flood level exceeds the
304 predesignated one. Finally, it is also recommended to provide and utilize the timeline chart for reducing flood damage since the preliminary arrangements are possible for the typhoon and heavy rain by consulting with the weather forecast. Fire is the most critical incident both in lives and properties. It is recommended to make periodical inspections of degraded cables, old switchgears and oil-immersed equipment. In old power plants, many poor conditions are remarkable such as no-fire wall of oil-immersed station-service or exciter transformers, water dropping to cable trays and cubicles from ceiling, crowded cables without suitable fixings and partitions, damaged and/or abandoned fixtures without removals, useless but still voltage-charged cables, etc. Fire alarms should be located adequately to detect any incipient fire, especially at the cable crowded areas like cable shaft and cable treatment room under control room, near the pressure oil equipment and oil-immersed transformers, etc. Fire drill should be effective to meet any accidental occurrence of a fire. Since many hydropower plants normally have no operator and maintenance staff, so that the initial fire-fighting work is very difficult. Therefore, a suitable emergency plan should be provided for a fire including the close cooperation with the district fire department. Hydroelectric powerhouse easily becomes a confined space under a fire, so that the escape routes should be ensured, especially from the lower floors and/or from the underground powerhouse. Portable fire extinguishers should be provided besides the working area under any hot works. There are many good recommendations in NFPA 851 [24], so that it should be utilized to improve the fire protection. For the machinery failures, a concept of life-cycle maintenance as shown in Fig. 13 will be useful as one of the practical approaches to study the preventive measures and actions since the causes, features and risky parts of many severe machinery failures are strongly related with the operating time.
Note: The durations of periods and time schedules of overhauls would be dependent on the actual conditions of the machine and the owner’s policies.
Fig. 13 Life-cycle maintenance of hydroelectric generating machines
Until about five years after the commissioning, many initial failures may occur due to defects at design, manufacturing and installation stages. The typical initial failures are the fatigue cracks, loosen bolts, insulation breakdown, galling at moving parts, improper action of protection relays, sequential failures of programmable controller (PLC), mal-setting of speed governor and automatic voltage regulator (AVR), etc. Sometimes the stable operation would be difficult by the flow-induced high vibration and excessive instability like power swing. Therefore, it is recommended to check the machine’s healthiness and integrity thoroughly at the commissioning tests and to make the first intensified inspection of vulnerable parts within a few years after the commissioning. If any severe incident would happen unfortunately, the root-cause analysis should be completely done to avoid any recurrence. During a period from five to twenty years after the commissioning, the plant is generally in a calm state since almost all initial failures had been rectified and the maintenance staffs are familiar to the machines. During the period, the main problems are the cavitation pitting of the runner vanes and sporadic breakdown of various parts. The first overhaul is conducted during the period for the repair of cavitation, replacement of worn-out parts and comprehensive inspections of critical parts. During a period from twenty to forty years after the commissioning, the initial degradation becomes obvious in the stator coil and runner. It is recommended to conduct periodical tests and checks, such as stator winding insulation test and partial discharge analysis for the generator, resolved gas analysis for transformer and junction box of oil-filled (OF) cable, oil property test for bearings, operation time check of circuit breaker and various nondestructive examinations (NDE) at overhaul works. It is important to grasp the longtime tendencies of degradation from those tests in order to plan the future maintenance strategy. If the testing results are not so good, the replacement by new parts should be planned, including runner, stator coils or whole stator, circuit breaker, etc. The short-lived computer-based control equipment, such as speed-governor board, AVR, PLC, protection relays, should be proactively replaced into new one to ensure the high reliability. After forty or fifty years from the commissioning, the degradation by low-cycle fatigues and heat cycles are obvious in many parts of turbine, generator and main transformer, so that the major overhaul or some comprehensive rehabilitation should be considered. The renewal of old power cable is recommended at the rehabilitation works of station service equipment, lighting fixtures, or various electric facilities in the powerhouse. To detect early sign of defective machinery parts will be very beneficial to prevent the severe incidents. One of the most expected future technologies will be the development of advanced vibration monitor using AI technology. Although many vibration monitors have been already used in many hydropower plants, the present on-line monitor is generally to detect the abnormality by the combination of the amplitudes and its duration. This way is not so much reliable and normally disabled at the start/stops due to the high vibration levels, although such period is the most dangerous time to the machine. Another way is an off-line measurement including the human judgment after studying the collected data definitely. It is accurate and usable for analyzing the cause of high
305 vibration, but it is not usable for a continuous monitoring. If the advanced analyzer which can automatically extract the relationship between the measured vibration levels and the machine conditions like the power, head, discharge, rotation speed, tailrace water level, draft pressure pulsation, bearing gaps and temperatures, etc., it will be very useful to detect any abnormality from the very initial stage, when the machine is even in the operation. The verification whether normal or abnormal will be highly improved by comparing the measured values with base-line data obtained under various operation conditions. In addition, the advanced monitor may be usable to collect information about the longtime degradation and remaining lifetime. To stop the machine safely when detecting any abnormality is the most important. Since it is impossible to make all machines immune from any accidental failure, the reliable action of emergency stop should be ensured for limiting the damage as small as possible. The most promised way is to keep the protection relays as much as reliable and to provide some backup system to make sure the successful emergency stop. JPower normally provides backup emergency-stop circuits to trip the generator circuit breaker and to close guide vanes and inlet valve independently even if the normal control circuits lose the functions by fire or other causes. It is also expected to start the international activities about the prevention of severe incidents in hydropower, such as the database collecting actual severe incidents and their analysis, guidelines to avoid severe incidents, standard procedure for repairing the severe damage, etc. Such materials will be helpful to determine the maintenance strategies to avoid any severe incident and to recover from it. The past failures will provide many lessons learned to prevent similar incident, so we hope this paper will be helpful and utilized for ensuring the steady and safe operation of existing and future plants.
References [1] Yasuda, M and Watanabe, S., 2016, “How to Avoid Severe Incidents at Pumped Storage Power Plants,” 28th IAHR. [2] IEEJ, Technical Report No.1144, 2009, “Study on equipment troubles of hydroelectric power plants,” (In Japanese). [3] The Federation of Electric Power Companies of Japan (FEPC), Engineering Department, 1988-2012, “Annual statistics of safety and maintenance activities for electric facilities,” (In Japanese). [4] Japan Meteorological Agency, 2013, “Climate change observation report 2013,” (In Japanese). [5] Japan Electric Power Civil Engineering Association, 2007, “Report for operational control of hydroelectric power against the abnormal weather,” pp29, (In Japanese). [6] Doi, Y., 2015, “Damages and Restoration works of a hydroelectric power station hit by Typhoon 12,” New Energy Foundation, 103rd Practical seminar of middle and small hydropower technologies (In Japanese). [7] Cheng, C.T. et al., 2009, “Evaluation of Landslides and Debris flow hazards for Reconstruction of Chin-shan Hydropower plant in Tachia Watershed after CHI-CHI Earthquake,” An International Conference in Commemoration of the 10th Anniversary of the Chi-Chi Earthquake. [8] Roose, D., 2015, “Investigation, Re-creation Indicates Likely Ignition Point of Thermalito Fire,” HydroWorld.com. [9] Bardy D.M. et al., 2011, “Emergency Response: Lessons Learned during Recovery from a Fire in the Detroit Powerhouse,” HydroWorld.com. [10] U.S. Fire Administration, 2002, “Fire at Watts Bar Hydroelectric plant,” USFA-TR-147. [11] Westberg, O.A., 2015, “Safety conditions in underground power plants: A histrical view of accidents. What did we learn?” Hydro2015, 13.01, Bordeaux, France. [12] Water Power & Dam Construction, September 1981. [13] Halipchuk, P. et al., 2011, “Machining Solutions: Repairing Cracked Turbine Shafts at Jenpeg,” HydroWorld.com. [14] Acker F., 2011, “Fatal failures: Siberia's hydro disaster,” Engineering and Technology Magazine. [15] BC HYDRO, 2008, “GM SHRUM GENERATING STATION G3 RUNNER FAILURE TECHNICAL ANALYSIS AND RECOMMENDATIONS,” Report No. E653. [16] Alvarez, A., 2011, “Solving a Problem Related to Shear Pin Failure,” HydroWorld.com. [17] Aliabadi, A. and Shamekhi, A., 2007. “Von Karman frequency excitation caused cracking of the Karun III Francis runner,” HYDRO2007 6.04. [18] Fisher, R.K. Jr., et al., 2002, “A Case Study in Resonant Hydroelastic Vibration: The causes of Runner Cracks and the Solutions implemented for the Xiaolangdi Hydroelectric Project,” 21st IAHR. [19] Price, J. W.H., 1998, “The failure of the Dartmouth turbine casing,” International Journal of Pressure Vessels and Piping Volume 75, Issue 7, pp 559–566. [20] Gummer, J.H. and Hensman, P.C., 1992, “A review of stayvane cracking in hydraulic turbines,” Water Power & Dam Construction. [21] Combes, G., et al., 2015, “A case of penstock rupture due to spherical valve resonance: investigations and numerical simulation,” HYDRO 2015, 19.06. [22] CIGRE TB454, 2011, “Hydro-generators Fire Protection Update (Questionnaire),” Working Group A1. 04. [23] CIGRE TB392, 2009, “Survey of Hydrogenerator failures,” Working Group A1.10. [24] NFPA 851, 2010, “Recommended Practice for Fire Protection for Hydroelectric Generating Plants”.
306