Thermal Study for Power Uprating of 400 MW Generator Using Numerical Simulation

Transcription

1 Theral Study for Power Uprating of 400 MW Generator Using Nuerical Siulation Hilan Syaeful Ala, Ia Djunaedi, and Dei Soetraprawata casings or hydrogen seals [3]. Therefore, theral analysis is needed to deterine the effect on the increasing of hydrogen pressure in the cooling syste. In addition, teperature distribution analysis is highly necessary to ensure there is no heat concentration point or hotspot that ight cause daage to the insulation on the winding or deagnetization. Since the ability and speed of odern coputers has increased rapidly and allows to run the calculations in a relatively short tie, the theral analysis of electric achines has becoe an interesting research field for both industry and investigators to iprove the cooling perforance and to enhance the efficiency of electric achine. In order to predict precisely the theral characteristics of electric achines, theral analysis can be divided into two ethods: Luped Paraeter Theral Model (LPTM) and nuerical ethod. The LPTM is the ain tool for a fast yet accurate theral analysis. However, a coon drawback to this approach can ake it useless to a designer, because there is a need for experiental data in order to tune paraeters that are very iportant in order to obtain accurate results [4]. It is of coon practice in theral odeling of electric achine to use the luped theral ethod, however if higher level of detail is required, it needs to use the ore refined ethods like nuerical ethod based on finite eleent or finite difference. These types of ethods can give a highly accurate and clear view of the teperature distribution within the electric achine [5]. Based on these advantages, nuerical ethod has been applied to assist in analyzing the theral characteristics of soe electric achine: icro gas turbine generator [6], switched reluctance generator [7], large ring otor for ining industry [8], Peranent agnet generator for large direct-drive wind turbines [9], and axial-flux peranent agnet generator [10]. This paper considers the theral analysis on a 400 MW hydrogen cooled generator using nuerical ethod based on finite eleent. The study was conducted on the generator in SURALAYA Coal Fired Power Plant, in Indonesia as one of the iportant aspects required in the planning for uprating or increasing the power output of generator. Theral analysis is focused to deterine its influence on the increasing of hydrogen pressure in the generator cooling syste. In addition, the teperature distribution on the generator was studied to ensure there is no heat concentration point or hotspot that ight cause daage to the insulation or deagnetization. Abstract One way to get ore power out of the generator in a power plant is by ipleent a power uprating, however it is required design review especially for theral characteristic of generator in order to prevent over heat and heat concentration point which can lead to the failure of insulation systes. In this study, theral characteristics of 400 MW Generator due to power uprating had been analyzed using nuerical siulation based on finite eleent. Case study was conducted on the generator of SURALAYA Coal Fired Power Plant in Indonesia based on the three conditions, i.e. existing condition, 5% and 7% uprating. Based on siulation results, the teperature distribution in the generator was able to be represented in detail in every segent where there was no indication of heat concentration point or hotspot on the generator. The increasing teperatures on winding due to the 5% and 7% uprating were respectively 29.9 C and 31.1 C by the largest error ±3.33%. Index Ters Power output, teperature, stress, generator insulation. I. INTRODUCTION One of the considerations in any power plants is how they get ore power out of their existing equipent. In any cases, soe equipent ay be capable of a 5% increase in power output or uprating, because the argin already exists [1]. For generators, one way to uprating is to the ake full use of any argin that exists to raise the field current and thereby increase the power output, but the consequences will raise the teperature and stress on probably all of the generator coponents, and priarily on the copper winding and insulations. According to Azizi et al. [2], the troubles initiated in the insulation are one of the priary root causes of failures of electric achines because one-third of the forced outages of large generators in generating stations and industrial plants are the caused by the failure of insulation systes. In soe cases, ost of generators ay be able to handle the increasing teperature by odifying the cooling syste. For large generator with a cooling syste using hydrogen, the odification can be done by increasing the hydrogen pressure inside the achine, but the ethod is liited by the strength of other coponents for exaple Manuscript received May 19, 2015; revised June 3, This work was otivated by a eorandu of understanding (MoU) which has been ipleented by the Indonesian Institute of Sciences (LIPI) and PT. Indonesia Power in the fields of research and developent of electricity on Noveber 7, 2011 with the MoU Nuber: 8.MOU/2/IP/2011 and 10/KS/LIPI/IX/2011. Hilan Syaeful Ala and Dei Soetraprawata are with the Technical Ipleentation Unit for Instruentation Developent, Indonesian Institute of Sciences, Jl. Sangkuriang Kop. LIPI Gd. 30, Bandung, 40135, Indonesia (e-ail: hil003@lipi.go.id, deistp@gail.co). Ia Djunaedi is with the Research Center of Physics, Indonesian Institute of Sciences, Jl. Sangkuriang Kop. LIPI Gd. 80, Bandung, 40135, Indonesia (e-ail: ia_djunaedi@yahoo.co). doi: /ij II. MATERIALS AND METHODS A. Generator Main Characteristics Generator Unit 3 in Suralaya Power Plant is a Hydrogen 191

2 Pwind 2Dr Len x Cooled, 3-phase synchronous Generator, 23kV, 0.85PF, 3 phase, 50 hertz, 3000 rp with anufacturing year in The ain characteristics of generator are shown in Table I. where Dr = outer diaeter of the rotor () dan Len = length of stator laination (). TABLE I: MAIN CHARACTERISTICS OF GENERATOR Paraeter Sy. Value Unit Power Noinal Voltage Nonial Frequency Nuber of Phase Noinal Speed Pole Nuber Hydrogen Cooled Pressure Stator Outer Diaeter Stator Inner Diaeter Rotor Outer Diaeter Nuber of Slot Stator Sinusoidal Flux density Winding Resistance per Phase Phase Current Weight of the rotor Length of stator laination P E f n p Hpr Do Di Dr Ss Bp Ra I Wr Len MW V Hz Coil Width of stator Coil Height of stator bs hs C. Heat Transfer Mechanis The source of heat in generator is dissipated to all parts of the achine following the pattern of radial heat flow or in the sae direction as the generator radial axis and circupherential heat flow or upright to the generator radial axis as shown in Fig. 1. For a 2-D theral analysis, axial heat flow which goes towards the sae direction as the axis of generator shaft can be neglected. rp bar Tesla Oh A kg B. Losses Calculations According to Ohyaa et al. [11], heat in generator is produced by total losses which coes fro copper or winding losses (Pcu), core losses (Pcore) and rotational losses (Prot). Those losses are then dissipated to all parts of the achine. At the generator, winding/copper losses, Pcu are generated due to resistances of the coil or winding that generates heat effects as the conversion of electric energy. Winding losses can be entioned using the following equation [12]: Pcu Ra I 2 Fig. 1. Heat transfer echanis on generator in 2D. In general the equation for heat transfer is based on the principle of energy conservation postulating that the network generated equals to the su of heat produced and energy alteration stored in the syste. Matheatically it can be written as follows [13]: (1) q Q where = phase nuber, Ra = winding resistance per phase (Ω), and I = phase current (A). Core losses, Pcore are the su of hysteresis loss and eddy current loss. Core losses can be represented by the following equation [11]: Pcore Ch fb p a bbp Ce f 2 B p (6) q k T 2 (2) (7) Therefore, heat transfer on a solid aterial can be stated in a partial differential equation as follows [13]: k T Q C p T 0 t (8) where k is theral conductivity, Cp is specific heat capacity, ρ is density, Q is rate of the heat which is generated per volue, and T is distribution of the predeterined teperature. In addition to conduction, heat can be transferred through convection and radiation. Coputation of convection heat transfer is ore coplicated due to involveent of fluid oveent and conductive heat. Newton gives a theory that heat transfer that passing thru a certain area is in accordance with the teperature difference of solid fluid. Teperature difference norally occurs through fluid borders located close to the solid surface. The Newton equation can be stated as follows [13]-[15]: (3) Friction loss is coputed using equation: Pfric k fbwr n 10 3 e t where q is the heat conduction, Q is the heat generated in the syste, δe/δt the change in internal energy. q is derived fro Fourier law as heat conduction, naely: where Ch = coefficient of hysteresis loss = 5, Ce = coefficient of eddy current loss = e-5, a and b are constants which depend on the stator aterial (for soft agnetic/silicon steel, a = and b = ). f = frequency, and Bp is sinusoidal peak of the flux density. Rotational losses Prot consist of friction and windage losses, the rotational losses can be calculated using the following equation [11]: Prot Pfric Pwind (5) (4) where kfb is the constant of bearing friction with the approxiation of 1-3, Wr = weight of the rotor (kg) and n = rotational velocity (rp). Meanwhile, wind losses Pwind are coputed using equation: Qc hc A T TA 192 (9)

3 where hc is coefficient of the convection heat transfer, A is area of the heat transfer, and TA is roo teperature. For natural convection, coefficient of the convection fro the air is approxiately between 6-30 W/2 C and for force convection, it is around W/2 C [16]. For heat transfer by radiation, it occurs fro or to the surface surrounded by parallel air with convection between the air and surface. Therefore, the total heat transfer is deterined by adding the contribution of those two heat transfer echaniss. Heat transfer by radiation can be coputed using following equation [13]-[15]: Qr t A T 4 TA 4 used as one of FEM software based on open source. Specifically, FEMM 4.2 discretizes the proble doain using triangular eleents as shown in Fig. 2. Over each eleent, the solution is approxiated by a linear interpolation of the values of potential at the three vertices of the triangle. The linear algebra proble is fored by iniizing a easure of the error between the exact differential equation and the approxiate differential equation as written in ters of the linear functions. The siulation procedures using FEMM 4.2 consist of geoetric odeling, aterial definition, setting of boundary condition, eshing or discretization and presenting the results. In geoetric odelling, geoetry of the generator is odelled into a quarter of segent of the 2D-generator, because the other parts of segents are syetrical. The types of aterials is previously defined in Table II are then correlated to the accordingly segents as depicted in aterial boundary. There are two boundary conditions that are used in siulation. The first is the heat flux as a source of heat generated fro the calculation of losses in the generator (Prot, Pcu, and Pcore) and the second is the convection boundary conditions that are applied to the area or boundaries related to the cooling teperature. The boundary conditions, geoetry and aterial boundary are shown in Fig. 3. (10) Fro the equation above, coefficient of the radiation heat transfer, hr, can be deterined as follows: hr t T 4 TA T TA (11) As the result, the total of heat losses in the generator due to convection and radiation, Qcr is as follows: Qcr hc hr A T TA (12) During theral analysis using finite eleent, heat transfer effects caused by radiation is siply ignored due to the high eissions of copper aterial and silicon steel inside the stator generator, which respectively are 0.63 and 0.7 [13]. Then, aterial properties of the generator coponents associated with theral analysis can be seen in Table II. TABLE II: MATERIAL PROPERTIES OF GENERATOR COMPONENTS Segent Material Specific Heat Theral Cap. Cond. (MJ/3K) (W/K) Stator Silicon Steel Winding Copper Frae Al-Alloy Shaft and rotor Forged Steel Ventilation hole Air D. Finite Eleent Siulation Fig. 3. Geoetry and boundary condition. Fig. 2. Discretization of doain using triangular eleent in FEMM 4.2 [15]. Theral analysis and teperature distribution on the generator is carried out using nuerical ethod based on the finite eleent (FEM). This ethod is able to give the approxiation of unknown values on soe discrete points of a continuu by dividing the continuu into saller eleents interconnected to each other at a certain coon point of two or ore eleents (nodal points or nodes) with liited nuber (finite eleent) fro a sipler geoetry than the previous continuu [18], [19]. In this study, FEMM 4.2 was Fig. 4. Result of eshing process. The next step in siulation is eshing. In eshing, the 193

4 saller the eleent, the ore accurate result of siulation will be. The capability in defining the nuber of eleents is greatly influenced by the capability/specification of not only the coputer but also the software in use. In this study, eshing process establishes 1164 nodes and 2211 eleents which are shown in Fig. 4. conditions consist of a heat source generator based on losses (windage, core and rotational) and cooling syste. Assuing an average abient teperature of the power plant at 35 C, the teperature distribution on the generators of finite eleent siulation results without cooling or natural convection to all three condition can be shown in Fig. 5. Teperature distribution represented in a 2-diensions for for call now generator segent and it can be seen that no heat concentration or hotspot exists for both conditions. Every segent receives generators are relatively the sae teperature flow (evenly). E. Heat Reoval Calculation The cooling of the arature winding is dependent on a nuber of factors: cooling ediu (air, hydrogen, water); insulation thickness; and overall electrical losses. Table III illustrates a relative heat reoval capability iproves fro air to hydrogen, with increased hydrogen pressure, and even ore significant with the use of water cooling [20]. In this study, theral analysis on the generator using finite eleent siulation based approach using air cooling syste. Then the siulation results calculated linearly based on the data of cooling using hydrogen at a predeterined pressure. Finite eleent siulations perfored on the existing condition, 5% and 7% uprating, where each condition were analyzed using air cooling systes and hydrogen. TABLE III: GENERATOR HEAT REMOVAL CAPABILITIES OF VARIOUS FLUIDS [20] Fluid Relative Relative Approx. rel. heat specific density reoval heat Air Hydrogen 2.07 bar Hydrogen 3.10 bar Water (a) III. RESULTS AND DISCUSSION The ain source of heat in generator is produced fro the windings located at stator s laination, and dissipated to all parts of the achine. By adding the influence of the other losses, the losses calculation results according to eq. (1) to (5) based on existing conditions, 5% and 7% uprating can be shown in TABLE. Fro the calculation results for the all three condition, only winding losses are increased according to the increasing of the current phase. The calculation results of heat flux adjusted to the location of the boundary condition where the area for winding losses, core losses and rotational losses are respectively , , (b) TABLE IV: LOSSES CALCULATION RESULTS FOR EXISTING, 5% UPRATING AND 7% UPRATING Looses Existing 5% 7% Cond. Uprating Uprating Winding looses, kw Core looses, kw Friction Looses, kw Windage Looses, kw Rotational Looses, kw Heat flux of wind., W/ Heat flux of core, W/2 Heat flux of rot., W/2 (c) Fig. 5. Teperature distribution of generator with natural air convection for (a) Existing condition, (b) 5% Uprating and (c) 7% Uprating. The results of theral siulation results using air cooling syste is used as input for the analysis of heat reoval using actual hydrogen cooling syste. Based on Table IV, the approxiate relative heat reoval using hydrogen cooling is 3.10 bar which has a cooling capacity of as uch as four ties better when copared by air cooling. Therefore, if the calculations are carried out according to data fro the theral siulation results, the ratio of heat reoval in the stator winding for all the three cases using soe type of cooling syste can be shown in Table V. To validate the nuerical siulations using actual conditions, the teperature of the coponents in particular generator coil is then copared by actual conditions in accordance by The results of losses calculation is then used as an input data for finite eleent siulation. Theral siulation was perfored for two conditions, i.e. without cooling (natural air convection) and air force convection. Then the calculation results of heat reoval in the generator cooling are correlated by actual cooling using hydrogen. The applied boundary 194

5 generator operating data on existing conditions and axiu power. Based on the operating data fro the control roo of the power plant, there are 6 therocouple sensors ounted on 6 locations of the coil / winding, where the average winding teperature has a value 1.47 C or 0.42% lower than the coil teperature siulation results on the sae conditions or existing. However, the axiu teperature that occurs in the sixth position of the coil (Coil 1) is 11.7 C or 3.33% higher than the siulation results. In order to eliinate the risk of insulation failure, the axiu actual teperature on the existing condition is used as a reference for existing coil teperature, and the estiation of the raising teperature associated with 5% and 7% uprating with the sae cooling conditions at hydrogen pressure of 3.10 bar is based on the siulation results by the axiu error of ± 3:33%. Therefore it can be concluded that the raising teperature on winding due to the 5% and 7% uprating is 29.9 C and 31.1 C respectively by the largest error ±3:33%. [6] [7] [8] [9] [10] [11] [12] [13] TABLE V: THE APPROXIMATE HEAT REMOVAL FOR EXISTING AND UPRATING CONDITIONS Condition Paraeter Existing 5% 7% Uprating Uprating Natural air T. Coil, C cooled Force air cooled T. Coil, C Heat re., C C Hydrogen cooled T. Coil, bar Heat re., C Hydrogen cooled T. Coil, C bar Heat re., C [14] [15] [16] [17] [18] [19] IV. CONCLUSION [20] Theral analysis using finite eleent ethods shows teperature distribution in generator accurately and detail in every segent of the generator. Based on the siulation results, there was no indication of heat concentration point or hotspot on the generator. The increasing teperatures on winding due to the 5% and 7% uprating were respectively 29.9 C and 31.1 C by the largest error ± 3.33%. Theral studies using a coputational fluid dynaic (CFD) technique is recoended for further study in an attept to obtain ore accurate results, therefore the interaction between the cooling fluids and heat sources on the generator can be analyzed dynaically. Hilan Syaeful Ala received his B.S. degree in echanical engineering fro Jenderal Achad Yani University, Indonesia in 2003 and M.S. degree in echanical engineering fro Institut Teknologi Bandung, Indonesia in 2010, He worked as a researcher at Indonesian Institute of Sciences (LIPI) fro 2006 until now. His research areas are renewable energy, energy conversion, engineering design, intelligent aterial & structures, instruentation & control, finite eleent and coputational echanics. Ia Djunaedi received his B.S. degree in echanical engineering fro Institut Teknologi Bandung, Indonesia in He had been research as a Ph.D research student at Curtin University of Technology, Australia fro 1993 to He worked as a researcher at Indonesian Institute of Sciences (LIPI) fro 1986 until now. His research areas are renewable energy, optiization, echatronics, theral engineering and power plant. REFERENCES [1] [2] [3] [4] [5] J. S. Park, S. Park, K. M. Ki, B. S. Choi, and H. H. Cho, Effect of the theral insulation on generator and icro gas turbine Syste, Energy, vol. 59, pp , B. Ganji, M. Heidarian, and J. Faiz, Modeling and analysis of switched reluctance generator using finite eleent ethod, Ain Shas Eng J, S. B. Andersen, I. F. Santos, and A. Fuerst, Multi-physics odeling of large ring otor for ining industry Cobining electroagnetis, fluid echanics, ass and heat transfer in engineering design, Applied Matheatical Modelling, Y. Alexandrova, R. S. Seken, and J. Pyrhönen, Peranent agnet synchronous generator design solution for large direct-drive wind turbines: Theral behavior of the LC DD-PMSG, Applied Theral Engineering, vol. 65, pp , M. Polikarpova, P. Lindh, C. Gerada, M. Rilla, V. Nauanen, and J. Pyrhonen, Theral effects of stator potting in an axial-flux peranent agnet synchronous generator, Applied Theral Engineering, vol. 75, pp , K. Ohyaa, M. Naguib, F. Nashed, K. Aso, H. Fujii, and H. Uehara, Design using finite eleent analysis of a switched reluctance otor for electric vehicle, JPE, vol. 6, no. 2, pp , Z. Xiaochen, L. Weili, C. Shukang, C. Junci, and Z. Chunbo, Theral analysis of solid rotor in PMSM used for EV, in Proc. the Vehicle Power and Propulsion Conference, 2009, pp A. Boglietti, A. Cavagnino, D. Staton, M. Shanel, M. Mueller, and C. Mejuto, Evolution and odern approaches for theral analysis of electrical achines, IEEE Transactions on Industrial Electronics. vol. 56, no. 3, pp , G. Mahalinga and A. Keyhani, Technical Report, Departent of Electrical Engineering - The Ohio State University, I. Boldea and A. N. Syed, The Induction Machine Handbook, Electronic Edition, CRC Press LLC, F. Kreith, Principles of Heat Transfer: Third Edition, Harper & Row Publisher, Inc, S. S. Rao, The Finite Eleent Method in Engineering, Elsevier Science and Technology Books, 2004, p. 3. D.V. Hutton, Fundaental of Finite Eleent Analysis, The Mc Graw Hill Copany, 2004, p. 4. D. Meeker, Finite Eleent Method Magnetics Version 4.2: User s Manual, Basic AC Electrical Generator, Aerican Society of Power Engineers. (2015). [Online]. Available: Optiized Maintenance of Generator Rotors, Electric Power Research Institute (EPRI), Palo Alto, CA, , D. Azizi and A. Gholai, Optiization of seiconductive coating and groundwall insulation layers in stator slot of HV generator, Electrical Power and Energy Systes, vol. 57, pp , P. Irasari, H. S. Ala, and M. Kasi, Theral analysis on radial flux peranent agnet generator (PMG) using finite eleent ethod, The Journal for Technology and Science, vol. 22-2, pp , J. A. Malubres, M. Satrustegui, I. Elosegui, J. C. Raos, and M. M. Iturralde, Analysis of relevant aspects of theral and hydraulic odeling of electric achines: Application in an open self ventilated achine, Applied Theral Engineering, vol. 75, pp , Z. Zhanga, S. M. Muyeenb, A. Al-Durrab, R. Nilssena, and A. Nysveena, Multiphysics 3D odelling of ironless peranent agnet generators, Energy Procedia, vol. 53, pp , Dei Soetraprawata received his B.S. degree in physic fro Nasional University, Indonesia in 1996 and M.S. degree degree in Instruentation and Control fro Institut Teknologi Bandung, Indonesia in 2014, He worked as a researcher at Indonesian Institute of Sciences (LIPI) fro 1983 until now. His research areas are renewable energy, instruentation and control. 195