Experiences on Using Ageing Models in Lifetime Estimation of Hydroelectric Generators. Voitto I.J. Kokko

Transcription

1 Experiences on Using Ageing Models in Lifetime Estimation of Hydroelectric Generators Voitto I.J. Kokko Fortum Power and Heat Oy, FINLAND

2 I INTRODUCTION Most of the existing hydropower production in Europe was constructed in the 194s, 5s and 6s [1]. Therefore, the age and condition of units used by hydropower production utilities can vary greatly. The role of hydropower has changed several times during its 1-year history. Until the 196s, hydropower was used mainly for base load production. A strong increase in thermal and nuclear power energy in the 196s and 7s led to an increase in the start-stop cycles and power regulation of hydropower units. De-regulation of the Nordic electrical energy market in the mid-199s led again to an increasing number of start-stop cycles and regulation in this area [2]. Experimental studies from areas with a relatively high amount of wind power show that an increase in wind power will once again increase regulation and probably the start-stop cycles of hydropower units [3]. A major failure of a hydro generator causes the shutdown of a production unit until the failed component has been repaired or renewed. This can lead to high unplanned production losses. Therefore, an estimation of the remaining lifetime of generators is of primary interest in the asset management of hydropower production. Most hydro generators reach their full lifetime of several decades with high reliability. Therefore the number of old units can be relatively high in the fleet of hydro generation companies. Generator refurbishment plans are usually made after about 1 years and preliminary plans even 2 to 3 years ahead. When the lifetime estimation of windings covers decades various factors affecting the total lifetime need to be taken into account. In addition, the lifetime estimation should be performed even though there are not yet any clear measurable signs of ageing. Therefore, the use of ageing models in lifetime estimation is needed. In this study the hydro generator has been divided into two main systems: stator and rotating mechanical systems. Lifetime estimation is performed separately for these systems. A stator includes the foundation, frame, core and winding. A rotating mechanical system includes rotor components: shaft, frame, rim, ring and poles, and bearing brackets and bearings. The failure of a stator winding causes production shutdown of a unit until the failure has been repaired. Figure 1 shows an asphaltic-mica multi-turn stator winding that failed due to a turn-to-turn short circuit. The stator frame and core usually have a longer lifetime than the stator windings. The repair time for a stator re-winding after a failure is shorter than renewal of both stator core and winding. In addition, the investment cost of stator core and winding renewal is about double in comparison to stator re-winding only. These facts have often led to the stator being re-wound instead of renewed, especially if the stator core has been in good condition.

3 Fig. 1 Failed stator winding due to turn-to-turn short circuit of multi-turn stator winding There are several factors which can restrict the technical lifetime of mechanical rotating systems. Mechanical vibration problems can be caused by hydraulic fluctuations in waterways, static and/or dynamic eccentricity, poor shaft straightness and vibrations caused by a hydro turbine. Vibrations can lead to fatigue and wearing of the shaft, rotor constructions, bearing brackets and bearings. High rotor eccentricity can lead to rubbing between rotor poles and stator core leading to stator failure. Cracks in the shaft are a serious risk, but fortunately they occur very seldom. Poor condition of bearings and rotor pole windings have often been improved or repaired with maintenance projects and therefore they have not usually restricted the technical lifetime of the mechanical rotating system. The lifetime estimations should be performed with systematic procedures to ensure repeatability of the analysis processes. The objective of this paper is to present the lifetime estimation procedures of hydro generators and present experiences of lifetime estimations. II AGEING MODELS Based on statistical failure studies, there are several ageing mechanisms which are to be taken into account [4], [6]. These can be divided between electrical ageing, thermal ageing, thermo-mechanical ageing or thermal cycling, mechanical ageing and environmental ageing. These will be briefly presented below using ageing models. A more detailed explanation of the models can be found in the references. In this study an electrical ageing model (1) has been used to estimate if the high electrical stress limits the expected lifetime of a stator winding [5]. The electric field stress is determined by dividing the winding phase voltage by the ground insulation thickness (kv/mm).

4 Error! Bookmark not defined. t (1) E = t Er E E w r n where t E is the expected lifetime, t Er is the reference lifetime in operation hours for the reference electric field stress E r, E w is the electric field stress of the reviewed stator winding and n is the calculated constant matching the ageing model to the operation history of the stator winding database. The following electro-thermo-mechanical ageing model was used to estimate the lifetime consumption by start-stop cycles and power regulation cycles [7]. t s E m l n w e = tr E re l (2) rm where, t s is the lifetime consumption (h/start-stop), t r is the reference value for lifetime consumption with reference mechanical stress, E re is the reference value for electric field stress, E w is the electric field stress of the reviewed stator winding, l e is the maximum relative thermal expansion or bar/coil during the start-up cycle, l rm is the reference value of mechanical expansion stress, and m and n are the calculated constants for electrical and thermo-mechanical stresses matching the ageing model to the operation history of the stator winding database. To estimate the environmental ageing no separate ageing model was used, but the effects of high ambient temperature and contamination are taken into account in the following so-called electro-thermal ageing model (3) since they both lead to an increase in the stator winding temperature [8]. This ageing model also takes into account ageing caused by high load leading to a stator winding temperature increase. n T T TI w E w Tb = ET tti k Ere t (3) where t ET is the expected lifetime of the stator winding with combined electrical and thermal stress, t TI is the lifetime in temperature of Temperature Index, k is a constant (.5), T TI is the temperature of the Temperature Index of the insulation system, T w is the winding average temperature, T b is the base difference temperature (1 C), E re is the reference value for electric field strength, E w is the electric field strength of the reviewed stator and n is the calculated constant for electrical stress matching the combined electrothermal ageing model with the lifetimes of stator windings in the operation history of the generator database. An expression used to estimate mechanical ageing is based on the mechanical ageing model (4).

5 t t M = M M m (4) where M is the mechanical stress, M is the value of mechanical stress below which the mechanical ageing can be neglected and failure under multi-stress conditions is the consequence of ageing produced by the other stresses, t is the life of M = M and m is the mechanical endurance coefficient. III LIFETIME ESTIMATION The target of the lifetime estimation used in this approach is not to forecast accurate breakdown times. The target is rather to maintain enough risk tolerance to avoid major failures which will result in high production losses. Lifetime estimation is performed on the basis of ageing models until clear changes in the diagnostic measurements are found. The schematic lifetime and failure probability estimation principle is presented in Figure 2 [4]. The expected lifetime is calculated for the two main components of a hydroelectric generator: the stator and the mechanical rotating system. Failure of a hydro generator stator winding causes the production shutdown of a unit until the failure has been repaired. Therefore, the status of a stator winding often determines the technical lifetime of the generator stator system. Failure probability 1% Lifetime estimation with ageing models Lifetime estimation and failure probability estimation with diagnostic measurements Repair action or refurbishment Allowed failure risk Probability level % Degradation mechanism or failure mechanism identification % Mortality failure period Random failure period, low failure probability, no clear changes in the measurement trends Increased failure probability period, clear changes in the measurement t Fig. 2 Schematic lifetime and failure probability estimation principles

6 Stator winding ageing is dominant for stators and therefore the lifetime estimation mainly concentrates on the stator winding. A schematic flowchart for the lifetime estimation of hydroelectric generator stator windings is presented in Figure 3 [5]. Factors to be taken into account include: design and manufacturing data, operation and failure history, operation mode and operation conditions, performed maintenance improvement actions and the results of both condition inspections and diagnostic measurements. Design, manufacturer and manufacturing year Manufacturing and installation Electrical ageing Used Operation Time Electro-Thermal ageing Electro-Thermo-Mechanical ageing / start-stop cycles Electro-Thermo-Mechanical ageing / regulation cycles Operation conditions Failure history Maintenance improvements Inspections and diagnostic tests Future operation mode Expected remaining lifetime in operation hours and operation years Fig. 3 Schematic flowchart for lifetime estimation of stator windings The hydro generator operation mode has changed over the decades. The number of both yearly start-stop cycles and regulation cycles has in many cases increased significantly. For example, some units which operated originally with 1 yearly start-stop cycles now have 15 to 2 yearly start-stops. Therefore, statistical lifetime estimation using operation history data does not any longer give adequate results: the lifetime effect of start-stop cycles and regulation cycles should be taken into account more precisely.

7 Lifetime estimation is performed in two phases: first, the consumed lifetime until the review date is calculated and after that the expected remaining lifetime is determined by taking into consideration the future operation mode. This allows the possibility to also simulate various future operation mode scenarios. The simulation can test the effects of the operation mode, such as the number of yearly start-stop cycles and regulation cycles on the lifetime estimation when planning changes to the operation mode. Figure 4 shows the simplified procedure used for rotating mechanical systems. The mechanical ageing model has been used to estimate the lifetime consumption by startstop cycles and regulation cycles. For ageing caused by vibration/mechanical problems, the analysis and opinions of generator experts and vibration measurement experts were used. Design, manufacturer and manufacturing year Manufacturing and installation Used Operation Time Mechanical ageing / Start-stop cycles Mechanical ageing / Regulation cycles Vibration/Mechanical problems caused ageing Maintenance improvements Inspections and diagnostic tests Expected remaining lifetime in operation hours and operation years Fig. 4 Lifetime estimation procedure used for rotating mechanical system IV LIFETIME STATISTICS OF HYDRO GENERATOR MAIN COMPONENTS The following lifetime statistics are based on an analysis of about 19 hydroelectric generators operating in Scandinavia. The output of the analyzed units is between 6 MVA and 135 MVA. Figure 5 presents performed refurbishments of rotating mechanical systems in five-year periods, including the primary reasons for the refurbishments. Based on the results, the lifetimes of rotating mechanical systems have been relatively long and the primary reasons for refurbishments have been generator up-ratings and

8 refurbishments. Therefore the timing of refurbishments has only seldom been the same as the technical ends-of-lifetime of rotating mechanical systems Number of rotors Refurbishments in 5-year periods Generator up-rate Generator refurbishment Fig. 5 Refurbishments of rotating mechanical systems in five-year periods Figure 6 shows stator re-windings and renewals performed in five-year periods, including the primary reasons for stator re-winding or renewal. Based on the results, the lifetimes have varied considerably and the primary reasons for refurbishments have been poor condition / frequent failures or generator up-ratings and conditions. In many cases, the re-winding or renewal timing has also been the same as the technical end-of-lifetime. Stator windings are sensitive to fire and flood in the power plant and therefore such cases have led to several winding failures Number of stator windings Re-windings and renewals in 5-year periods Up-rate and condition Poor condition/failure Others (flood, fire) Fig. 6 Re-windings and renewals of stator windings in five-year periods

9 Figure 7 presents performed stator core renewals in five-year periods, with the primary reasons for the stator renewals. Based on the results, the lifetimes of stator cores have been relatively long and the primary reason for renewals have been generator up-ratings and poor condition of the stator winding and core. Therefore, the renewal timing in many cases has been at least near of the technical end-of-lifetime of the cores Number 8 of stator cores Renewals in 5-year periods generator up-rate poor winding and core condition Fig. 7 Renewals of stator cores in five-year periods It can be summarized that the lifetime distribution of rotating mechanical systems and stator cores are quite similar. For stator windings the lifetime distribution is much wider and there is a need to understand in more detail the reasons for various lifetimes. Stator winding failures always cause production breakdowns until they have been repaired. Therefore the following section has concentrated in more detail on the remaining lifetime estimation of stator windings. V EXPERIENCES WITH STATOR WINDING LIFETIME ESTIMATION Lifetime estimation for more than 2 generators was performed based on the procedure presented in Figure 3. In the first phase of the analysis the remaining lifetime estimation was made based on ageing models. With the electrical ageing model whether the high electrical stress limits the expected lifetime of winding was estimated. The effects of high ambient temperature and contamination were taken into account in the electro-thermal ageing model since they both lead to an increase in the winding temperature. Finally, the developed electro-thermo-mechanical ageing model was used to estimate the lifetime consumption by start-stop cycles and regulation cycles. In those cases where a degradation mechanism was identified, the results of inspections and diagnostic measurements were used to update the estimation. Lifetime estimation was performed for 73 asphaltic-mica stator winding systems. The output of the analyzed units was between 8 MVA and 7 MVA, and the winding voltage

10 was 6 kv to 16 kv. By using the electro-thermo-mechanical ageing model the average value for all windings was 9.4 hours per start-stop cycle, which was nearly the same as in another study [2] of asphaltic-mica stator winding systems. The distribution of the results is presented in Figure 8. It can be observed that there are significant differences between various units Number of asphaltic-mica stator windings Lifetime consumption [h/start-stop cycle] Fig. 8 Distribution of lifetime consumption (hours/start-stop cycle) for asphaltic-mica windings The most common stator winding insulation systems today are based on epoxy-mica insulation. Lifetime estimation was performed for 9 epoxy-mica winding systems. The output of the analyzed units was between 8 MVA and 135 MVA, and the winding voltage was 6 kv to 2 kv. By using the electro-thermo-mechanical ageing model the average value was about 4.2 hours per start-stop cycle. The distribution of the results for epoxymica stator windings is presented in Figure 9. Based on this figure, it can be stated that there are again significant differences between various units. Unfortunately, there is not enough experimental data from epoxy-mica stator windings which have already reached the end of their lifetime to verify these results. By comparison of average values and Figures 8 and 9, it can be seen that the lifetime effect of start-stop cycles is significantly smaller with epoxy-mica than with asphalticmica winding systems.

11 25 2 Number of epoxy-mica stator windings < Lifetime consumption [h/start-stop] Fig. 9 Distribution of lifetime consumption (hours/start-stop cycle) for epoxy-mica windings There are different opinions about whether the lifetime estimation should be made based on operation age, operation hours or operation hours taking into account ageing factors caused by the operation mode. In the fleet being studied there are dozens of generators with asphaltic-mica multi-turn stator windings and about one dozen with asphaltic-mica single-turn stator windings from the same manufacturer. The oldest of them were manufactured at the end of the 193s and the newest at the beginning of the 196s. Most of them have already reached their end-of-lifetime and therefore they were used to verify differences between various lifetime estimation methods. Cases in which the reason for the renewal was an external reason, such as a fire or flood in the power plant, were excluded. The "operation-mode-corrected lifetime estimation" procedure was used by subsequently calculating the reached end-of-lifetime for tens of units that had already been renewed. Cases in which the reason for the renewal was an external reason, such as fire or flood in the power plant, were excluded. Some results from these calculations are presented in Figure 1. Lifetime consumptions have been divided into the following factors: operation hours, ageing caused by start-stop cycles, ageing caused by regulation cycles, electro-thermal ageing and electrical ageing. It can be seen that there are significant differences between the reached lifetimes. In some cases the weak design is the reason for a shorter lifetime, and in some cases the reason for the renewal was that the relatively high turbine upgrade potential would increase the profitability of the renewal project. In these cases the generators were also renewed even though there was still a remaining lifetime of several years. In Figure 1,

12 the winding numbers 1 to 36 are multi-turn constructions and numbers 37 to 49 are single-turn constructions. In many cases the lifetime effect of start-stop cycles was significant (as in winding numbers 3, 4, 1, 11, 18, 21, 43, 44 and 45 in Figure 13). Electro-thermal ageing has only a slight effect on these lifetimes because cooling water is relatively cool throughout the year and therefore the winding temperature stays quite low. 6 5 End-of-lifetime [operation-mode-corrected operation hours] Asphaltic-mica stator windings Operation hours Start-stops Regulation Electro-thermal ageing Fig. 1 Reached end-of lifetimes presented with operation-mode-corrected operation hours for asphaltic-mica stator windings To estimate the effectiveness of lifetime estimation methods and the differences between lifetimes of various winding constructions, the windings were divided into multi-turn and single-turn asphaltic-mica stator windings and the ends-of-lifetime were estimated afterwards with three different methods. Figure 11 shows the end-of-lifetime distribution for lifetimes based on operation years, Figure 12 is based on operation hours and Figure 13 is based on operation-mode-corrected operation hours, taking into account start-stop cycles, power regulation cycles and electro-thermal ageing. Instead of using a fixed value for start-stop cycles and regulation cycles, their effects were determined separately with the electro-thermo-mechanical ageing model.

13 12 1 Number of Asphaltic-mica stator windings End-of-lifetime [operation years] Multi-turn Single turn Fig 11 End-of lifetimes in operation years for multi-turn asphaltic-mica stator windings Number of Asphaltic-mica stator windings End-of-lifetime [operation hours * 1] Multi-turn Single turn Fig 12 End-of lifetimes in operation hours for multi-turn asphaltic-mica stator windings

14 12 1 Number of Asphaltic-mica stator windings End-of-lifetime [operation mode corrected operation hours * 1] Multi-turn Single turn Fig 13 End-of lifetimes in operation-mode-corrected operation hours for multi-turn asphaltic-mica stator windings When comparing of the effectiveness of the various lifetime estimation methods, it can be assumed that a suitable method gives a narrow lifetime distribution for a similar stator winding system. Table 1 presents key statistical figures for various methods by using the end-of-lifetime results of multi-turn asphaltic-mica stator windings from the same manufacturer. It can be found that the method using "Operation-mode-corrected operation hours" has the smallest deviations in comparison to average values and it can be concluded that it is the most suitable method. Deviations using "Operation years" are 4 to 5% higher and deviations with "Operation hours" are more than double in comparison to "Operation-mode-corrected operation hours". Table 1 Comparison of lifetime estimation methods using statistical figures with various lifetime estimation methods for multi-turn asphaltic-mica windings Statistical figures for multi-turn asphaltic-mica stator windings (total 36) Method Operation years Operation hours Operation-modecorrected operation hours Average lifetime 5.1 years Max lifetime 74 years 532 h h Max deviation+ 46.7% 31.3% 21.3% Min lifetime 34 years 164 h h Max deviation 32.2% 42.9% 32.1% 9% deviation+ 23.7% 28.4% 17.% 9% deviation 24.2% 4.9% 15.9% Standard deviation+ 11.8% 14.2% 8.5% Standard deviation 12.1% 2.5% 8.%

15 The drawback of using the average operation years in lifetime estimation is that lifetime estimates are too low for generators having low yearly operation times. This is not a risk for the availability of the generators but can lead to early renewal. By regularly monitoring the condition of stator windings this drawback can be handled and the renewal decision is made just after the condition is noticed to be poor. The benefit of this method is that it is simple and in many cases gives reasonable results. Especially in cases when there is no access to operation history and operation modes, average operation years can be used for the rough remaining lifetime estimations. The drawback of using the average operation hours for lifetime estimation is that the operation mode, including a high number of yearly start-stops and/or regulation cycles, is not taken into account. In these cases the expected lifetime is too long. This can lead to unplanned winding breakdown and therefore this method is not recommended to be used with hydro generators. The method of using operation-mode-corrected operation hours takes into account operation mode and operation conditions, including start-stop cycles, regulation cycles and various ageing effects. It survives well in all cases and therefore it can be concluded that it is the most suitable method for lifetime estimations. This method requires information about the operation history and operation mode. However, this is not a problem for power plant operators. Often it is assumed that the lifetime of the multi-turn windings is shorter than that of single-turn windings. This property can also be found in the statistical key figures for end-of-lifetime distribution of multi-turn and single-turn windings presented in Tables 2 and 3. In Table 2, figures for operation years are presented and in Table 3 those for operation-mode-corrected operation hours. In both tables the summary results for multiturn and single-turn windings have also been presented. Deviations are smaller separately for multi-turn and single-turn windings than for the summary results including both winding types. Based on these results it can be also concluded that lifetime estimation should be made separately for each winding construction. Deviations with the operation-mode-corrected lifetime estimation method are smaller than the estimation with operation years.

16 Table 2 Statistical figures for multi-turn and single-turn asphaltic-mica windings using lifetime estimation method of operation years Statistical figures for multi-turn and single turn asphaltic-mica stator windings - Operation years method Method Multi-turn windings (36) Single-turn windings (13) Multi-turn & singleturn windings (49) Average lifetime 5.1 years 56.3 years 51.8 years Max deviation+ 46.7% 29.6% 42.9% Max deviation 32.2% 39.6% 34.3% 9% deviation+ 23.7% 26.1% 37.1% 9% deviation 24.2% 29.% 28.5% Standard deviation+ 11.8% 13.% 18.6% Standard deviation 12.1% 14.5% 14.3% Table 3 Statistical figures for multi-turn and single-turn asphaltic-mica windings using lifetime estimation method of operation-mode corrected operation hours Statistical figures for multi-turn and single turn asphaltic-mica stator windings - Operation-mode-corrected operation hours Method Multi-turn windings (36) Single-turn windings (13) Multi-turn & singleturn windings (49) Average lifetime h h h Max deviation+ 21.3% 21.% 41.9% Max deviation 32.1% 28.4% 34.6% 9% deviation+ 17.% 11.% 21.8% 9% deviation 15.9% 28.4% 21.4% Standard deviation+ 8.5% 5.5% 1.9% Standard deviation 8.% 14.2% 1.7% There are still many original asphaltic-mica windings in operation. Figure 14 presents the estimated lifetime consumptions for 24 stator windings still in operation. The lifetime figures have been divided into the following factors: operation hours, electrical ageing, ageing caused by start-stop cycles, ageing caused by regulation cycles and electrothermal ageing. It can be seen that in some cases the consumed lifetime is still quite low because of the low yearly operation time. In some cases the effect of start-stops is significant (6, 15). Electrical ageing will probably limit the lifetime in two cases (6, 1). The effect of electro-thermal ageing is again low.

17 6 5 Lifetime [Operation-modecorrected operation hours] Asphaltic-mica windings in operation Operation hours Start-stop cycles Regulation cycles Electro-thermal ageing Electrical ageing Fig. 14 Current lifetime consumptions of asphaltic-mica stator windings in operation Figure 15 presents the estimated remaining lifetime for the same stator windings presented in Figure 14. It can be seen that the estimated remaining lifetimes of two units have already been passed. Both of these generators have been added to the refurbishment program as urgent cases. In addition, it can be seen that a considerable number of these windings will reach their end-of-lifetime over the next five years. These cases have also been included in the renewal plan Estimated remaining lifetime [year] Asphaltic-mica windings in operation Fig. 15 Estimated remaining lifetimes of asphaltic-mica stator windings in operation

18 VI CONCLUSIONS Based on the presented lifetime statistics, the stator windings have often had shorter lifetimes than stator cores or rotating mechanical systems. This has led to re-winding or stator renewal. Up-rating of turbines and generators has most often been the primary reason for the refurbishment of mechanical rotating systems and stator cores. Lifetime estimation of stator windings based on average operation lifetime in years is suitable for cases when there is no good access to the operation history and operation mode. With this method the rough lifetime estimation can be obtained and reasonable results reached. The drawback is that the method gives too low lifetime figures for generators having a low yearly operation time rate and a relatively low start-stop rate. Lifetime estimation of stator windings based on operation hours gives unreliable results, especially for hydro generators, because it does not take into account the operation mode and therefore it is not recommended to be used. Lifetime estimation of stator windings based on operation-mode-corrected operation hours gives the most reliable results. It is recommended that the effect of start-stop cycles and regulation cycles are determined separately for each winding with the electro-thermo-mechanical ageing model instead of using one fixed value for all windings. This method requires information about the operation history and operation mode. However, this should not be a problem for power plant operators. It is recommended that in lifetime estimation the winding construction is also taken into account because, based on the experimental results, the lifetime of single-turn stator windings is longer than with multi-turn windings. In summary, lifetime estimation procedures are a useful tool in the maintenance and investment planning of hydro generators. With a lifetime estimation tool, whether the operation mode or other ageing modes will shorten expected lifetime significantly can be estimated. In addition, it is possible to use the lifetime estimation tool for power plant operation planning by simulating the effects of future operation mode changes. Both of these properties are useful, particularly in hydropower production. VII REFERENCES [1] J. Sierra and I.J. Pérez-Arriaga, "Unity for All," IEEE Power and Energy Magazine, vol. 7, no. 5, pp , September/October 29. [2] A. Karlsson and T. Karlsson, "Estimating Lifetimes for Stator Windings in Hydropower Generators," 9th International Conference on Probabilistic Methods Applied to Power Systems, KTH, Stockholm Sweden, June 26. [3] D. Corbus, D. Lew, G. Jordan, W. Winters, F. Van Hull, J. Manobianco and B. Zavadil, "Up With Wind," IEEE Power and Energy Mag. vol. 7, no 6, pp , 29. [4] V. Kokko, "Failure Risk Estimation for Hydroelectric Generator Stator Windings Using Condition Index," ViennaHydro Conference, Vienna Austria, 21. [5] V. Kokko, "Electrical Ageing in Lifetime Estimation of Hydroelectric generator stator windings," ICEM 21 Conference, Rome Italy, September 21.

19 [6] CIGRE, "Hydrogenerator Failures - Results of the Survey," CIGRE Study Committee SC11, EG11.2, 22. [7] V. Kokko, "Ageing Due to Thermal Cycling by Start and Stop Cycles in Lifetime Estimation of Hydroelectric Generator Stator Windings," International Electric Machines and Drives Conference", May 211, Canada Niagara falls. [8] V. Kokko, "Electro-Thermal Ageing in Lifetime Estimation of Hydroelectric Generator Stator Windings," SDEMPED 211 Symposium, Bologna Italy, September 211. BIBLIOGRAPHY Voitto Kokko was born (27 August 1956) in Rovaniemi, Finland. He received his B.Sc degree in electrical engineering in 1979 and joined Fortum Power and Heat in He studied electrical engineering at the University of Oulu and received his M.Sc. degree in 1996 and D.Tech degree in 23. From 1979 to 1997, he worked as a commission engineer in thermal power and hydropower plants. Between 1998 and 22, he worked as a development manager focusing on the analysis of electrical condition measurements. Between 22 and 28, he worked as head of the condition management centre. From 28, he has worked as the leading expert in the asset development of hydro power in Finland and Sweden. His main technical responsibility area is generator lifetime management.