Effects of Unbalanced Faults on Transient Stability of Cogeneration System

Transcription

1 ffects of Unbalanced Faults on Transient Stability of Cogeneration System WI-NNG CHANG CHIA-HAN HSU Department of lectrical ngineering Chang Gung University 259 Wen-Hwa st Road, Kwei-Shan, Tao-Yuan, Taiwan, ROC Abstract This paper evaluates the effects of unbalanced faults on the transient stability of a real cogeneration plant. First, a brief is given for the structure of the cogeneration system. Use of the electromagnetic transient program (MTP) constructs the cogeneration system. Several fault types including 3-line-to-ground (3LG) fault, double line-to-ground (2LG) fault, line-to-line fault (2LF), and single line-to-ground (SLG) fault occurring in the utility side are assigned respectively for transient stability simulations. The simulation results are listed and evaluated. Finally, the critical clearing time (CCT) curves with different fault residual voltages for different faults are obtained and compared. Keywords CCT curve, Cogeneration plant, MTP, Transient stability, Unbalanced faults. Introduction Due to calculation limitation of simulation programs, balance three-phase faults are usually assigned for stability analysis of power systems. However, most faults frequently occurring in power systems are unbalanced faults. The effects of unbalanced faults on cogeneration system need to be clarified. This helps the field engineers in the cogeneration plant to realize their power system more widely. This also helps then to establish suitable protection system, such as settings of protection relays and relative parameters for necessary grid-disconnection schemes to protect the cogeneration units in the plant from out-of-synchronism. When short-circuit faults occur in the utility side, voltage sags will spread and penetrate into cogeneration plant []. Severe voltage sages with long enough fault intervals may cause cogenerators in the plant out-of-step [2~4]. Before lose of synchronism, the cogeneration plant should be disconnected from the utility grid and operated alone. Use of CCT curves can evaluate the transient stability of the cogenerators to the short-circuit faults. Under voltage relay (27) with directional overcurrent relay (67) can be used together to identify the severity of the faults occurring in the utility [5,6]. Grid-disconnection operation should be executed according to the CCT curves to keep the plant be operated safely. This enhances the fault-ride-through ability of the plant. This paper observes the effects of unbalanced faults on the transient stability of a real cogeneration plant. For the need of unbalanced operations, the MTP program is employed in the paper [7]. For comparisons, transient stability analysis with 3LG fault is firstly conducted as a base case, then simulations with different types of unbalanced fault are performed. Simulation results are given and compared. 2 System Structure Figure shows the studied cogeneration system including four coal-fired cogenerators. Table and Table 2 list the specifications and electricity outputs of these cogeneration units, respectively. G 0A 0A G4 Utility 6kV/ 6,000MVA 6kV 69kV 22kV 22kV 4kV 4kV 22kV 22kV 30MW 45MW 32MW S 3MW Fault site 20MW 0A G2 6MW G3 24MW Fig. Single-line diagram of the cogeneration system 0A ISBN:

2 Table Rated parameters of cogeneration units capacity (MVA) voltage (kv) H (Sec) pf G G G G Table 2 lectricity outputs of cogeneration units utility load conditions (MW) peak load 37 G off-peak load 23 peak load 83 G2 Cogenerator off-peak load 53 units peak load 92.2 G3 off peak load 62.2 peak load 94.5 G4 off-peak load 54.5 The cogeneration plant is an exporting type plant selling electricity to the utility in the peak load period of the utility. The total electricity generations of the plant are 307MW in the peak load period and 93MW in the off-peak load period. The total load demand in the plant is 80MW. Fig. 2 shows the I-AC brushless excitation system and the I-ST2 steam turbine/governor system for all the cogeneration units in the plant, respectively [8, 9]. All parameters of the two subsystems are from the cogenerator manufacturers. V ref r - - stc K A st st B skf st st F R V t A VR min VR max - st S K K D 0 V fd,max fd FX f IN IN I IN KC V (a) I-AC brushless excitation system Δ ref R Fhp ΔP st st rh sr - P Fip st c st co P max P min F lp st ch fd I fd P mech (a) I-ST2 power control system Fig. 2 I standard voltage and power control block diagrams for the cogeneration units The loads in the plant are considered as composite loads including static load and dynamic load. Since the cogeneration plant has powerhouses and chemical processes which use heaters and induction motor loads, dynamic and static loads are modeled for the plant. qu. () shows the rotor dynamics of the induction motor in the dynamic loads. qu. (2) represents the load torque-speed characteristic for all the motor loads in the plant. dr M Dr T TL () dt 3 2 TL( r).47r 3.2r 0.9r 0. (2) In which: M = inertia constant of motor and load, D = damping factor, r = rotor speed, T = induced torque, T L = load torque. Generally, static loads can be represented as (3) and (4) called ZIP load model [0]. All loads in the plant are represented as 80% dynamic load and 20% static load in the simulation. In P(V,f) P[p p( ) p( ) ]( p( )) (3) L Q(V,f) Q[qq( ) q( )]( q( )) (4) L which: P,Q,V,f = operating points, p ~ p, q ~ q = relative coefficients of 4 4 active power and reactive power. 3 MTP Modeling of the System Figure 3 shows the MTP modeling of one of the cogeneration units. The I-AC excitation system and the I-ST2 power control system are modeled and merged into the cogeneration units. Fig. 4 is the MTP modeling of the overall cogeneration plant. The transformers in the plant are also modeled in the MTP according to manufacturers data sheets. The fault site in the utility side is modeled by using a Y-connected branch with neutral impedance to the ground. Different assigned values in the Y branch simulate different types of balance and unbalanced faults. Careful choice of impedance values in the Y branch also controls the degree of fault. With the settings, the cogeneration plant will suffer different types of fault with different fault residual voltages and fault times. 2 ISBN:

3 Fig. 3 MTP modeling of one of the cogeneration units Fig. 4 MTP modeling of the cogeneration plant 4 Transient Stability Analysis Several types of fault events have been assigned in the utility for transient stability simulations. Different fault settings with different fault impedances and fault times are assigned for the transient stability analyses. This obtains the relative critical clearing time (CCT) curves. Generally, larger fault residual voltage and fault time result in longer CCT. The paper considers the following unbalanced faults: 2LG, 2LF, and SLG faults. Different CCT curves with different fault types for peak load and off-peak load periods are evaluated. For comparison, 3LG fault is simulated as a base case. Case : Base Case-3LG Fault Figure 5 shows the system responses for a complete 3LG fault occurring in the utility in the peak load period. The CCT is founded to be 4 cycles. When the fault is cleared, all the cogenerators in the plant go back to their stable operations. When the fault time is increased to 5 cycles, the cogenerator G3 in the plant becomes unstable, as shown in Fig. 6. Fig. 5 Cogenerators responses to a 4 cycles (CCT), 3LG complete fault in the peak load period 3 ISBN:

4 Fig. 6 Cogenerators responses to a 5 (CCT) cycles 3LG complete fault in the peak load period The CCT test is also executed for the off-peak load period in which all cogenerators in the plant decrease their electricity outputs, as shown in Table 2. During the off-peak period, the electricity output to the utility is reduced to zero. Fig. 7 shows the simulation results. The CCT extends to 2 cycles as shown in Fig. 7(a). When the fault clearing time is increased to 22 cycles, the cogenerator G3 becomes unstable, as shown in Fig. 7(b). (a) 2 cycles 3LG fault (CCT) Fig. 8 Cogenerators responses to a 2LG fault with a fault time of 27 cycles (CCT) in the peak load period Fig. 9 Cogenerators responses to a 28 cycles (CCT) 2LG fault in the peak load period Case 3: 2LF Fault Figure 0 shows the simulation results with an 8 cycles 2LF fault occurring at phase b c in the peak load period. All the cogenerators in the plant maintain stable. Further simulation shows that the CCT to the 2LF fault is over 00 cycles for both peak load and off-peak load periods which is long enough for protection devices in the utility to operate to clear the fault and restore the cogeneration system from the fault. The 2LF fault in the utility side does not threat the stability of the plant. (b) 22 cycles 3LG fault (CCTcycle) Fig. 7 Cogenerators responses to 3LG faults occurring in the off-peak load period Case 2: 2LG Fault A complete 2LG fault occurring at phase b-c in the utility side in the peak load period is assigned in the simulation for the CCT test. Fig. 8 shows the simulation results with a fault time of 27 cycles (CCT). Fig. 9 shows the 2LG fault simulation results when the fault time is increased to 28 cycles. The cogenerator G3 becomes unstable when the fault is cleared. The CCT is 27 cycles for a complete 2LG fault occurring in the utility. Fig. 0 Cogenerators responses to an 8 cycles 2LF fault in the peak load period Case 4: SLG Fault Figure shows the simulation results of a SLG fault occurring in the utility side in the peak load period. The fault time is also set at 8 cycles for comparison. The cogeneration system maintains stable when the fault is cleared. Increasing the fault time to 00 cycles does not make any of the cogenerators unstable. The 4 ISBN:

5 simulation using the off-peak load data shows similar result. It is concluded that SLG fault in the utility has not significant effect on the transient stability of the plant. Initial Point (40, 0.738) (a) 3LG fault Initial Point (40, 0.738) Fig. Simulation result of 8 cycles SLG fault Figure 2 shows the torque versus rotor angle characteristics of the G3 for different types of complete fault in the peak load period. For comparison, the fault intervals are all set at 4 cycles. In the pre-fault situation, the magnetic torque of the G3 is pu and the rotor angle is 40 degrees. Fig. 2(a) is the response to 3LG fault which shows a typical oscillation phenomenon that can be widely seem in many articles describing the equal-area criteria. The fault clearing angle is 00 degrees. In the post-fault interval, the maximum rotor angle swing is 40 degrees. Fig. 2(b) shows the response to 2LG fault. During the fault interval, the fault residual voltage is about 0.6 pu, which partially sustains the power transfer ability of the cogenerator, the cogenerator still has power output during the fault interval. This decreases the acceleration area. Hence, the fault clearing angle is not so large as compared to Fig. 2(a). This leads to larger deceleration area to be used for extending the CCT. The similar phenomena can also be observed in Fig. 2(c) for SLG fault. It is evidenced that SLG fault hardly causes the cogenerator to be out-of-step. The simulation results in Fig. 2 explain why the 3LG fault is the most severe fault for the transient stability performance. Furthermore, unbalanced faults generate negative-sequence voltages at the terminal of the cogenerator. The negative-sequence voltage induces braking torque which helps stabilizing the cogenerator from lost-of-synchronism []. Hence, this also helps increasing the CCT. Initial Point (40, 0.738) (b) 2LG fault (c) SLG fault Fig. 2 Torque versus rotor angle characteristics of G3 for different types of fault Figure 3 summarizes the calculated CCT curves versus fault residual voltages for the cogeneration plant for different types of faults in the peak load and off-peak load periods. The CCT curves show that the cogeneration plant is most vulnerable to different degrees of 3LG fault occurring in the peak load period. The CCT curve of the 2LG fault in the peak load period lies between the 3LG faults occurring in the peak load and off-peak load periods. Since the CCTs of the SLG and 2LF faults are larger than 00 ms which can be cleared in time by protection devices in the utility, these two types of unbalanced fault are not so 5 ISBN:

6 significant to the transient stability of the plant. Fig. 3 also recommends a low voltage tripping line for the grid-disconnection scheme of the cogeneration plant. When the fault residual voltage touches the line, the cogeneration plant should be disconnected from the utility and switched to islanding operation mode to protect these cogenerators in the plant. Critical Clearing Time (Sec) () 3LG peak load (2) 3LG off-peak load (3) 2LG peak load (2) Fault Residual Voltage (pu) Fig. 3 The CCT curves for different types of fault occurring in the utility (3) () Low voltage tripping (27) 5 Conclusion This paper observes the effects of unbalanced faults on the transient stability of a real exporting type cogeneration plant by using the MTP program. From the simulation results and the CCT curves in Fig. 3, the following conclusions are listed:. The most severe fault to the transient stability of the in-plant cogeneration units is 3LG faults occurring in the peak load period for an exporting type cogeneration system. 2. When the output power of the cogeneration unit is decreased during the off-peak period, the maximum allowable swing angles of the cogenerators are increased due to the equal-area criteria. The obtained CCT curve of the cogeneration plant is higher (larger) then that at the peak load period. 3. In the peak load operation period, the CCT curve to 2LG faults is near twice of the CCT curve of 3LG faults. In the off-peak period, the CCTs of the 2LG faults are long enough to be used for protection devices in the utility to clear the faults. 4. The SLG and 2LF faults occurring in the utility is not so easy to cause the cogenerators out-of-step. 5. The rotor inertia constant of cogenerator also influences the stability of the cogeneration unit. Cogenerator with larger rotor inertia constant will slow down the acceleration of the rotor during the fault interval, hence the time to critical fault clearing angle is extended. This increases the CCT. References: [] M. F. McGranaghan, D. R. Mueller and M. J. Samotyj, Voltage sags in industrial systems, I Trans. Industry Applications, 993, Vol. 29, No. 2, pp [2] J. C. Das, ffects of momentary voltage dips on the operation of induction and synchronous motors, I Trans. Industry Applications, 990, Vol. 26, No. 4, pp [3] P. Kundur, Power System Stability and Control, Mcgraw-Hill, Inc., 994. [4] P. M. Anderson and A. A. Fouad, Power System Control and Stability, I Press, 993. [5] W. S. Zimmermann, S. Hopp, M. Bondeur and D. N. Chen, Transient stability study of the Hsin Yu Co-Generation Plant in Hsin-Chu Science Based Industrial Park in Taiwan, Proceedings of I Power ngineering Society Winter Meeting, 2000, Vol., pp [6] C. T. Hsu, Cogeneration system design for a high-tech science-based industrial park, I Trans. Industry Applications, 2003, Vol. 39, No. 5, pp [7] lectromagnetic Transients Program (MTP), MTP-V3 Rule Book I, Development Coordination Group of MTP, 996. [8] I Standard 42.5, I Recommended Practice for xcitation System Models for Power System Stability Studies, [9] I Committee Report, Dynamic models for steam and hydro turbines in power system studies, I Trans. Power Apparatus and Systems, 973, Vol. 92, pp [0] Y. Li, H. D. Chiang, B. K. Choi, Y. T. Chen, D. H. Huang and M. G. Lauby, Representative static load models for transient stability analysis: Development and xamination, IT Gener. Transm. Distrib., 2007,, (3), pp [] T. M. M. O Flaherty and A. S. Aldred, Synchronous-Machine stability under asymmetrical faults, Proceedings of the I - Part A: Power ngineering, 962, Vol. 09, pp ISBN: