Power Quality Assessment of Large Motor Starting and Loading for the Integrated Steel-Making Cogeneration Facility

Transcription

1 Power Quality ssessment of Large Motor Starting and Loading for the Integrated Steel-Making Cogeneration Facility Cheng-Ting Hsu, Member, IEEE Department of Electrical Engineering Southern Taiwan University of Technology Tainan, Taiwan bstract This paper presents the power quality assessment of large synchronous motor starting and loading in the integrated steel-making cogeneration facility. To execute the transient stability analysis, the proper mathematical models and the accurate parameters of the cogeneration units, excitation systems, governor systems, load and static var compensators (SVC) are investigated in detail. Four study cases with or without considering the connection of the power grid, the installation of autotransformer starter and SVC are performed to demonstrate the dynamic responses of system frequency, voltage and cogeneration units due to motor starting and loading. lso, the voltage sag ride-through curve of sensitivity load has been included and a power quality index (PQI) due to voltage variation in the assessment period has been proposed to find the impact of motor starting and loading on the power quality of the cogeneration system. It is concluded that the system dynamic responses and PQI values have better performances if the autotransformer starter is applied with either the regulation of SVC system or connecting to the bulk power grid. Keywords- Power Quality; Cogeneration; Motor Starting I. INTRODUCTION The cogeneration facility to be studied in this paper is a large integrated steel-making plant, which will be built in the near future [1]. It will produce steel with iron-making, steelmaking and rolling for converting raw iron into many kinds of products. From the economic and security viewpoints, the energy efficiency and power system reliability can be greatly improved if the cogeneration units are installed. The process by-product gases such as coke oven gas, blast furnace gas and basic oxygen furnace gas are mixed with coal or oil as multiple fuels to provide boilers to generate steam for the cogeneration units. The whole project plans to fulfill the plant expansion through four stages of construction. lthough the cogeneration facility can provide electric power to its own load demand, the electric power system has to be connected to the utility so that the backup power can be obtained in case of outage or maintenance of the cogeneration units. The load demand of the steel plant may vary dramatically and irregularly due to the stochastic load characteristics. The rapid and frequent changes of load such as rolling mills may result in serious voltage and frequency fluctuations phenomena. In addition, the large motors starting can draw several times of their full load current to result in the significant voltage sags Hui-Jen Chuang Department of Electrical Engineering Kao-Yuan Institute of Technology Kaohsiung, Taiwan hjchuang@cc.kyit.edu.tw which may cause the magnetic contactors to drop out and disrupt sensitive equipment [1,]. This paper discusses a large synchronous motor starting and loading by executing the transient stability analysis. To alleviate the motor starting impact on the cogeneration system, the suitable tap autotransformer starter can be used to reduce the starting current and the voltage sag. lso, the static var compensator (SVC) systems have widely used in the steel plant to reduce voltage fluctuation by quick response of reactive power compensation with power electronics devices [3]. It is also important for the system planners to know the effectiveness of the power grid on the motor starting and loading. The large motors are planned to be started and loaded under various operation scenarios as described previously by executing transient stability analysis. The dynamic responses of system frequency, voltage and cogeneration units are examined carefully for each case study. esides, a power quality index based on the root-mean-square (rms) voltage level has been proposed and applied to evaluate the impact of voltage variation due to motor starting and loading. y comparing the dynamic response and power quality index for each case study, the most suitable operation case is therefore selected and adopted by the cogeneration facility. II. VOLTGE VRITIONS ND POWER QULITY INDEX The IEEE Standard 1159 [4] provides the categories and typical characteristics of the power system events. Table I and II give the classification of short term and long term voltage variation by duration and voltage magnitude. In general, interruptions are caused by the system faults, while voltage sags are usually associated with system faults, the heavy loading and large motor starting. On the other hand, the voltage magnitude variation between 0.9pu to 1.1pu can be considered as a voltage fluctuation or flicker which may result largely from the continuous and rapid variations of load [5]. ny kinds of voltage variations described above may affect the normal operation of equipment and deteriorate the power quality. lthough interruptions cause the most serious impact on the load, the occurrence probability of flickers and sags are much larger than interruptions. Flickers may dim light and voltage sags can cause the tripping of sensitive equipment such as adjustable-speed drives, magnetic contactors, etc /05/$0.00 (C) 005 IEEE

2 TLE I. TYPICL DURTION ND VOLTGE MGNITUDE FOR SHORT-TERM VOLTGE VRITION Voltage Magnitude Instantaneous Interruption < 0.1 Sag Swell Momentary Interruption < 0.1 Sag Swell Temporary Interruption 3-60 < 0.1 Sag Swell TLE II. TYPICL DURTION ND VOLTGE MGNITUDE FOR LONG-TERM VOLTGE VRITION Voltage Magnitude Interruption > 60 0 Under voltage > Over voltage > More and more industrial equipments become very sensitive to voltage sags [6]. In general, equipment sensitivity to voltage sags can be presented in the form of ride-through capability curves. The ride through capability curves proposed by the Information Technology Industry Council (ITIC) [7] and the semiconductor industry (SEMI-F47) [8] are both considered in this paper. Table III and IV give the duration and voltage magnitude specified by the ITIC and SEMI-F47. The ITIC and the SEMI F47 also define the over voltage and swell regions of ride-through curves. y integrating the ITIC curve with the SEMI F47 curve, a wider range of voltage sag ride through capability curve as shown in Fig. 1 has been adopted in this paper. The horizontal scale in Fig. 1 is logarithmic in seconds. ny voltage sag events have the magnitude and duration below the ride-through curve (the dotted line in Fig. 1) may result in the tripping of sensitivity load. Fig. 1 The proposed voltage sag ride-through capability curve and PQI diagram where N is the event numbers occurred in the assessment period. The PQI values of X t and E tn can be further expressed as X t 100 () N E tn 100 (3) N where is the average rms voltage in the assessment period and N is the minimal voltage magnitude of the N-th event. s seen from Fig. 1, it is quite obvious that the and N are the voltage deviations from the nominal value to the ride-through curve of the equipment. y applying the proposed power quality definition, the PQI value will be zero if the rms voltage variation can be always maintained at the nominal value as shown in the Fig. 1. In addition, the PQI value is defined as 100 if the points of the events are dropped just on the ridethrough curve of the equipment. TLE III. VOLTGE SG DURTION ND VOLTGE MGNITUDE FOR THE ITIC RIDE-THROUGH CPILITY < > 1000 Voltage magnitude not specified not specified TLE IV. VOLTGE SG DURTION ND VOLTGE MGNITUDE FOR THE SEMI F47 RIDE-THROUGH CPILITY < > 1.0 Voltage magnitude not specified not specified Furthermore, any kinds of voltage variations can affect the power quality even if points of the events are located above the ride-through curve. power quality index based on the rms voltage level is therefore proposed to evaluate the impact of voltage variation on the cogeneration system. In the t- second assessment duration, the power quality index (PQI t ) due to the rms voltage variation is defined as X E E... E PQI t + t1 + t + + tn t (1) N + 1 III. SYSTEM DESCRIPTION ND MODELING Fig. shows the one line diagram of the cogeneration system at the phase I construction. The cogeneration system will be connected to the power grid of Taiwan Power Company (TPC) through two 161kV circuits. In the cogeneration facility, two steam cogeneration units (TG1 and TG) with 10MW rating capacity each are connected to bus 1 through the 16/161kV transformers to supply the power demand of the steel plant. The TRT1 cogeneration unit with a rating capacity of 17.8MW uses the high-pressure exhaust gas to produce electricity will be installed and connected to bus 3. Two 33kV substations at buses 10 and 11 receive electric power from the 161kV bus through two 161/33kV on load tap changer transformers. The total average and peak load demand are 135MW and 19MW respectively. lso, a total amount of 36Mvar shunt capacitor bank is applied at this stage. In the oxygen plant (bus 3), a large synchronous motor with rating capacity of 16.1MW will be installed to drive the air compressor. Furthermore, the autotransformer (T) and the SVC system are both installed to improve the transient response and enhance the power quality of the cogeneration system for the large motor starting and loading.

3 Fig. 3 The equivalent model of the induction machines TLE V. PRMETERS OF THE SYNCHRONOUS MOTOR R s X s X m R r X r H Fig. One-line diagram of the study power system To simulate the transient behavior of large motor starting and loading in the cogeneration facility, the modeling of the apparatus must be derived carefully [9]. First, the transient and sub-transient impedances of the steam turbine cogeneration units are considered in the computer simulation so that more accurate transient stability analysis can be obtained. The excitation systems of all cogeneration units correspond to IEEE type and 3 models. Furthermore, the simplified governor models are applied to represent the dynamic behavior of the output mechanical power of the turbine. Detailed parameters of all the cogeneration units, excitation systems and governor systems can be found in reference [1]. The models of load and SVC system applied in this paper are given below in a very detailed manner.. Load Models oth of the dynamic and static load models [10,11] are applied in this paper. static load model with the combination of constant power, constant current and constant impedance is adopted to represent the load behavior on voltage magnitude and frequency as shown in (4) and (5). P Q V V f0 V 0 V 0 P C + C + C [ 1 + N ( f )] (4) V V f0 V 0 V 0 Q C + C + C [ 1 + N ( f )] (5) In this study, the constant power load is adopted when the bus voltage is greater than 0.7pu. When the bus voltage is dropped below 0.7pu, the constant current load will be considered. The frequency deviation factor, N 1 and N, are assumed to be value of. On the other hand, a conventional dynamic induction model as shown in Fig. 3 is applied for the analysis of motor starting and loading to investigate the dynamic responses of voltage and frequency by considering the slip and inertia constant. Table V gives the parameters of the motor to be studied in per unit with base values of 16.1MV and 11.5 kv.. SVC Model SVC system with fixed capacitor-thyristor controlled reactor type is assumed to be installed at bus 11 of the cogeneration system via a step up transformer as shown in Fig. 4. The susceptance of the reactor, L, can be regulated according to the variation of voltage magnitude at controlled bus (bus 11) by adjusting the firing angles of the thyristors. Fig. 5 shows the basic control block diagram of the SVC system [1]. The measurement variable, V mea, is taken from the terminal voltage (V t ) of controlled bus by a potential transformer with /D conversion and filtering process. The control circuit adopts the voltage deviation as input signal to regulate the reactive power output of the reactor. The gain of the voltage regulator, K R, is set to be 10 and the time constant, T R, is usually between 0ms and 150ms. The lead-lag terms (T 1 and T ) are set to be zero in this study. The adjustable susceptance range of the reactor is set to be between the 0.3 and 0.9, and the susceptance of fixed capacitor is set to be 0.6 on the 100MV base. Table VI gives the parameters of the SVC system. Fig. 4 SVC system connected to the bus 11 of the cogeneration facility Fig. 5 Control block diagram of the SVC system TLE VI. PRMETERS OF SVC SYSTEM T m K R T R T 1 T T Lmin T Lmax C

4 IV. TRNSIENT NLYSIS FOR THE MOTOR STRTING ND LODING To investigate the impact of large motor starting and loading on the power system of the steel cogeneration facility, four different operation conditions are considered by executing the transient stability analysis. Without connecting to the bulk TPC system, the large synchronous motor in the oxygen plant may be started with the full voltage, the T starter and the regulation of SVC system, which are represented as case study of, and C respectively. For the case study D, the cogeneration system is connected to the TPC system and the synchronous motor will be started by using the T starter. Since the exciter of the synchronous motor will not be applied during the motor starting, it is therefore to model the synchronous motor as an induction motor as shown in Fig. 3 with the corresponding parameters. Furthermore, the exciter will be applied to control the power factor of the synchronous machine to near unity when the synchronous speed is reached. esides, the synchronous motor is assumed to be started with lightly loading and the load of 11.3MW will be applied after the motor reaches the synchronous speed. In the cogeneration facility, the TG1 unit is the only power supplier with a 5MW output to provide the load demand at powerhouse plant. Case In the case, the cogeneration system is disconnected with the TPC and the synchronous motor is starting with the full voltage. Fig. 6 shows the active and reactive power consumptions of the synchronous motor. Fig. 7 shows the voltage responses of buses 1, 11, 3 and cogeneration unit TG1 respectively. It is found that the motor can reach its synchronous speed at about 13.9 second after the motor starting. The initial active and reactive power consumptions are 9.3MW and 34.4Mvar respectively. The severe voltage sags at different buses have been introduced due to the large reactive power demand by the motor. t the instant of motor starting, it can be seen that the voltages at buses 1, 11, 3 and TG1 drops to the minimum values of 0.883pu, 0.833pu, 0.659pu and 0.891pu respectively. The voltage levels at many buses have suddenly increased to exceed the nominal voltage when the motor has approached the synchronous speed. fter that, the excitation system and 11.3MW load are then applied at about 40 second. y adjusting the synchronous motor at unity power factor, the bus voltages are therefore increased since the reactive power consumption of the synchronous motor has been reduced to be zero. Fig. 7 Voltage variations of buses 1, 11, 3 and TG1 for case study Fig. 8 gives the frequency response of the isolated cogeneration system. The frequency variation is mainly introduced by the active power consumption of the synchronous motor. The frequency drops to the minimal value of 59.5Hz at about 13.4 second and then restores to the value of 59.95Hz. Fig. 9 shows the electrical active and reactive power outputs, the mechanical power input of the TG1 cogeneration unit. The cogeneration units will adjust their power outputs by regulating the excitation system and governor controller according to the deviations of the bus voltage and system frequency respectively. It is found that the electrical power responses are very similar to the power requirement of the synchronous motor. The mechanical power is regulated by the governor system to match the electrical active power variation of the TG1 cogeneration unit. The serious voltage sags in this case study may result in the tripping of more sensitive equipments even if the frequency variation of the isolated system and the power variation of the cogeneration unit are both acceptable. Fig. 8 System frequency response for case study Fig. 6 ctive and reactive power variations of the synchronous motor for case study Fig. 9 Electrical and mechanical power responses of the TG1 generator for case study

5 Case To alleviate the voltage sags during the motor starting, a conventional voltage step-down method by adjusting the suitable tap of the autotransformer is considered in this case study. The T starter is to regulate the bus voltage applied to the motor as 6.9kV with 60% tap position. Fig. 10 shows the consumption of active and reactive power profiles of the synchronous motor. It is found that the motor can reach its synchronous speed at about 3.4 second after the motor starting. Furthermore, the initial active and reactive power consumptions are reduced to 4.8MW and 0.4Mvar respectively. Fig. 11 gives the voltage responses of buses 1, 11, 3 and TG1. t the instant of motor starting, the voltages at buses 1, 11, 3 and TG1 have been dropped to the minimum values of 0.966pu, 0.93pu, 0.853pu and 0.95pu respectively. Most of the time, the voltages at buses 1, 11, 3 and TG1 are kept approximately at 1.0pu, 0.98pu, 0.9pu and 1.0pu respectively. It is observed that the voltages at 161kV and 33kV buses are maintained well due to the quick response of the TG1 excitation system by increasing its reactive power output. fter that, the nominal voltage of the synchronous motor is applied at about 35 second. It is the reason why the power consumption of the motor is suddenly increased and the voltage magnitude at each bus is decreased abruptly. Finally, the exciter of the synchronous motor and the load of 11.3MW are then applied at about 40 second. found that the variations of the electrical and mechanical powers are reduced more significantly as compared to the case. ccording to the discussion above, it is concluded that the action of adjusting the autotransformer at 60% tap can solve the problem of large motor starting and loading without deteriorating the stability of the isolated cogeneration facility. Fig. 1 System frequency response for case study Fig. 13 Electrical and mechanical power responses of the TG1 generator for case study Fig. 10 ctive and reactive power variations of the synchronous motor for case study Fig. 11 Voltage variations of buses 1, 11, 3 and TG1 for case study Fig. 1 illustrates the frequency response of the isolated cogeneration system for case. The minimum frequency of 59.75Hz has been observed after applying the 11.3MW load. Fig. 13 shows the electrical active and reactive power outputs, the mechanical power input of the TG1 cogeneration unit. It is Case C To improve the voltage variation of the cogeneration system further, the SVC system as shown in Fig. 4 has been installed at bus 11. In this case, the cogeneration system is still disconnected with the TPC system and the motor is started with the T starter. Fig. 14 shows the voltage responses of buses 1, 11, 3 and TG1. t the instant of motor starting, the voltages at buses 1, 11, 3 and TG1 have been dropped to the minimum values of 0.991pu, 0.953pu, 0.873pu and 0.963pu respectively. The voltage variation at each bus has been improved significantly as compared to cases and due to the very quick reaction of reactive power compensation by SVC system. Fig. 15 shows the electrical active and reactive power outputs, the mechanical power input of the TG1 cogeneration unit. The variations of the active power and mechanical power are similar to case. However, the reactive power variation of TG1 has been reduced greatly because of the regulation of the SVC. On the other hand, the power consumption of the motor and the frequency variation of the isolated cogeneration system are almost the same as case. Case D In this study, the cogeneration facility is connected with the TPC and the synchronous motor is starting with the T starter. Fig. 16 shows the voltage responses of buses 1, 11, 3

6 Fig. 14 Voltage variations of buses 1, 11, 3 and TG1 for case study C Fig. 17 System frequency response for case study D Fig. 15 Electrical and mechanical power responses of the TG1 generator for case study C and TG1. t the instant of motor starting, the voltages at buses 1, 11, 3 and TG1 have been dropped to the minimum values of 0.996pu, 0.981pu, 0.898pu and 0.966pu respectively. Furthermore, the voltage variation at each bus has been greatly improved. Fig. 17 shows the frequency response of the cogeneration system. It can be found that very little frequency variation has been introduced by the motor starting and loading. Fig. 18 shows the electrical active and reactive power outputs, the mechanical power input of the TG1 cogeneration unit. The electrical and mechanical powers of TG1 have the smallest variation among all study cases. With the strong support of bulk TPC system, the power quality of the cogeneration facility can be improved effectively according to the dynamic responses of the voltage, frequency and cogeneration unit. Fig. 16 Voltage variations of buses 1, 11, 3 and TG1 for case study D Fig. 18 Electrical and mechanical power responses of the TG1 generator for case study D V. POWER QULITY SSESSMENT DUE TO MOTOR STRTING ND LODING y applying the PQI definition in section II and observing the results of transient analysis, the PQI values at various buses during different assessment periods for each case study can be calculated. Table VII gives the PQI values of bus 11 in a 0 second assessment period for four cases. The PQI value is calculated as 69.6 for case. In this case, there are three events (E t1 ~E t3 ) occurred to result in the lowest voltage magnitudes of 0.83, 0.9 and 0.986pu with corresponding duration of 1.067, 1.15 and 0.68 seconds respectively. The points of three events are all above or just on the proposed voltage sag ride-through curve as shown in Fig. 19. y using (6)-(8), the PQI values for the three events are calculated as 85, 100 and 7 respectively. esides, the PQI value of X t is obtained by (9) because the average rms voltage in the 0 second assessment period is 0.954pu. The PQI value of case is therefore calculated as 69.6 by applying (10). y the same way, the PQI values can be calculated as 48.5, and for the cases, C and D respectively. Table VIII gives the PQI values of bus 11 in a 40 second assessment period for four study cases. The PQI values are 66.36, 38.50, and for the cases,, C and D respectively. The quantity analysis of power quality at bus 11 has a little improvement in the 40 second assessment period. It can also be found that the case C has resulted in the best power quality performance due to motor starting and loading. Furthermore, the power qualities at bus 11 are acceptable for all study cases because the PQI values are all less than the critical value of 100. Table IX and X give the PQI values of bus 3 for the four study

7 cases in the 0 and 40 second assessment periods respectively. It is observed that the power qualities of bus 3 are worse than bus 11 due to the motor starting and loading. esides, the power qualities are acceptable except case which has the PQI values over 100. TLE VII. THE PQI t0s VLUES OF US 11 FOR THE FOUR STUDY CSES Study cases C D E E E Event no. Occurred starting time Lowest voltage rms value Events PQI E t E t E t X t0s E t Cases PQI t0s X t0s E t X t0s E t X t0s Fig. 19 Points of events in the PQI diagram (6) 0. t (7) 0.1 t (8) 0. t X t 0s (9) 0.1 PQIt 0s 4 X t + E t1 + E t + E t3 N (10) TLE VIII. THE PQI t40s VLUES OF US 11 FOR THE FOUR STUDY CSES Study cases C D Event no. Occurred starting time Lowest voltage rms value Events PQI E t E t E t X t40s E t E t X t40s E t E t X t40s E t Cases PQI t40s X t40s TLE IX. THE PQI t0s VLUES OF US 3 FOR THE FOUR STUDY CSES Study cases C D Event no. Occurred starting time Lowest voltage rms value Events PQI E t E t E t E t E t X t0s E t Cases PQI t0s X t0s E t X t0s E t X t0s TLE X. THE PQI t40s VLUES OF US 3 FOR THE FOUR STUDY CSES Study cases C D Event no. Occurred starting time Lowest voltage rms value Events PQI E t E t E t E t E t X t40s E t E t E t X t40s E t E t E t X t40s E t E t X t40s Cases PQI t40s

8 VI. CONCLUSIONS This paper has investigated the impact of large motor starting and loading on the phase I configuration of the selected integrated steel making cogeneration facility. large synchronous motor with a rating capacity of 16.1MW has been started under four different operation scenarios. ccording to the transient stability analysis, the dynamic responses of system frequency and cogeneration unit are acceptable for all study cases and significant voltage sag has been resulted when the motor is started directly. lso, the power quality quantity of the cogeneration facility for each operation case has been solved by applying the voltage sag ride-through capability of sensitivity load and the proposed power quality index in the 0 and 40 seconds assessment period. To enhance the power quality of the cogeneration facility due to motor starting and loading, both of the autotransformer starter and the SVC system should be applied when the cogeneration system has been disconnected from the bulk power grid. On the other hand, the autotransformer starter provides the significant improvement in power quality when the cogeneration facility has tied with the utility power system. H. J. Chuang received the.s. and M.S. degree in Electrical Engineering from National Taiwan University of Science and Technology in 1990 and 199 respectively, and Ph. D. degree in Electrical Engineering from National Sun Yat-Sen University in 00. He is presently an ssociate Professor at Kao Yuan Institute of Technology, Lu Chu, Taiwan. His research interest is in the area of load flow and power system analysis of mass rapid system. REFERENCES [ 1] C. T. Hsu, "Transient Stability Study of the Large Synchronous Motors Starting and Operating for the Isolated Integrated Steel-Making Facility", IEEE Trans. on Industrial pplications, vol. 39, no. 5, pp , Sep./Oct., 003. [ ] R. T. Dyer, D. R. Mccue, S. L. Williams, S. J. Swencki and P. O. Thoits Static Var System for Starting and Operating Large Synchronous Compressor Motors, IEEE Trans. on Industry pplications, vol. 3, pp , Jan./Feb [ 3] C. S. Chen, H. J. Chuang and S. M. Tseng, "Mitigation of Voltage Fluctuation for an Industrial Customer with rc Furnace ", Proceeding of the IEEE PES Summer Meeting, pp , July, 001. [ 4] IEEE Recommended Practice for Monitoring Electric Power Quality, IEEE Std , Nov., [ 5] R. C. Dugan, M. F. McGranghan and H. W. eaty, Electrical Power Systems Quality, New York: McGraw-Hill, [ 6] M. H. J. ollen, Understanding Power Quality Problems: Voltage Sags and Interruptions, New Jersey: IEEE Press, 000. [ 7] "ITI (CEM) Curve pplication Notes", Information Technology Industry Council, Washington, DC, 000. [ 8] "SEMI F Specifications for Semiconductor Processing Equipment Voltage Sags Immunity", Semiconductor Equipment and Materials International, Mountain View, C, 000. [ 9] P. M. nderson and.. Fouad, Power System Control and Stability, New York: IEEE Press, [10] W. J. Lee, M. S. Chen and L.. Williams, "Load Model for Stability Studies", IEEE Transactions on Industry pplications, vol. 3, pp , Jan./Feb [11] IEEE Task Force on Load Representation for Dynamic Performance, "Load Representation for Dynamic Performance nalysis", IEEE Trans. on Power Systems, vol. 8, pp , May [1] IEEE Special Stability Controls Working Group, "Static Var Compensators Models for Power Flow and Dynamic Performance Simulation", IEEE Trans. on Power Systems, vol. 9, pp. 9-40, Feb., IOGRPHIES Cheng-Ting Hsu was born in Taiwan in He received the.s., M.S, and Ph.D. degrees in electrical engineering from the National Sun Yat-Sen University, Taiwan in 1986, 1988, and 1995, respectively. From 1990 to 199, he was with Phoenixtec Power Company Limited as a power electronics engineer, developing UPS equipment. He is currently a Professor of Electrical Engineering, Southern Taiwan University of Technology, Tainan, Taiwan. Dr. Hsu is a member of IEEE.