A Rigorous Method for Power Quality Evaluation of High-speed Railway Using Electrical Transient Analyzer Program

Transcription

1 Qingshan XU 1, Qiuqi ZHU 1, Haixiang ZANG 1, Xiaodong YUAN Jiangsu Provincial Key Laboratory of Smart Grid Technology & Equiment,,Southeast University, China (1), Electrical Power Research nstitute (EPR) of Jiangsu, China (). A Rigorous Method for Power Quality Evaluation of High-seed Railway Using Electrical Transient Analyzer Program Abstract. A rigorous method is introduced to evaluate the overall imact of high-seed railway (HSR) traction loads on a ower system using the electrical transient analyzer rogram (ETAP) latform (Oeration Technology, nc, USA) in this aer. y establishing the detailed models of network elements and the general traction substation model with the owerful simulation software, this method is caable of handling the large-scale ower systems and the various transformer connection schemes. As the interaction between the ower system and the traction ower-suly system is considered, the negative sequence comonents and harmonics effects can be estimated in detail both at the oints of common couling (PCCs) and in ublic grid. Shanghai-Nanjing intercity railway is taken for an examle, in which four ower quality indices are simulated and evaluated in different oerating conditions. y the comarison of simulation data and measured data, the roosed method is verified to be credible and feasible for the ower quality evaluation in ractice. Streszczenie. W artykule oisano sosób ścisłej oceny wływu trakcji elektrycznej kolei dużych rędkości na system energetyczny. W badaniach symulacyjnych wykorzystano latformę z analizatorem stanów rzejściowych sieci (ang. Electrical Transient Analyzer Program). (Metoda ścisłej oceny jakości energii w kolejach dużych rędkości analizatorem ETAP) Keywords: high-seed railway (HSR); negative sequence comonents; harmonics; electrical transient analyzer rogram (ETAP). Słowa kluczowe: Kolej dużych rędkości, składowa rzeciwna, harmoniczne, analizator ETAP. 1 ntroduction Recently, high-seed railway (HSR) has raidly develoed in China and the ower quality issue is becoming increasingly serious. Poor ower quality may significantly affect the oeration of the ower system and other connected equiments. As usually using the single-hase ower suly, the railway system generates lots of negative sequence comonents to the ublic grid. These negative sequence comonents may lead to additional loss and overheating of motors, undesired triing of relays, etc. [1,]. A large amount of harmonics are also generated by traction drivers into the grid, which may lead to torque ulses and torque dros of rotational machines, extra temerature rise of equiment, vibrations of ower caacitor banks and rotectors, etc. [3,4]. Comared to the conventional electrified railway, the high-seed electrified railway has a higher traffic density, which requires more reliable and larger traction ower [5]. Furthermore, its effects have new characteristics, such as the greater negative sequence comonents and the wider harmonic sectrum. As the HSR traction substations along a railway route are usually fed by the high-voltage network directly, its effects on the ublic grid need to be aid secial attention. Many works have been undertaken for ower quality evaluation of traction loads [6-8]. Although these methods are quick and straightforward, they are somewhat not rigorous. These methods are limited to estimate the effects at the oints of common couling (PCCs) without considering the enetration in the ublic grid. n addition, the imact of traction loads cannot be decided by these simle methods alone for it is caused not only by traction loads but also by other disturbance sources in grid. Furthermore, the interactions between traction loads of different substations are also significant, esecially for the neighbouring substations. To estimate the ower quality effects, the comrehensive simulation of the whole system that is comosed of the ower system and the traction ower-suly system is indisensable. And the interaction between these two systems should be taken into consideration. Thus, the detailed comonent models, including the models of network elements and the traction substation model, must be established aroriately. Moreover, a advanced system analysis caability is needed for the rigorous study of the effects, thus, a owerful simulation tool should be used to satisfy the requirements of the rigorous system analysis. As a result, the detailed comonent models and the suitable simulation tool are redominantly concerned in the evaluation work. The effects of the negative sequence comonents and harmonics are estimated in detail both at the PCCs and in the ublic grid in this aer. The electrical transient analyzer rogram (ETAP) software (Oeration Technology, nc., USA) is chosen as the analysis tool since it is owerful in the field of ower system analysis, esecially for the unbalance analysis and the harmonic analysis. All network elements, including generators, transmission lines, transformers and loads, are modelled in detail in this aer, and, a general traction substation model based on the rincile of equivalent transformation is adoted. These make the roosed method be caable of handling largescale ower systems and various transformer connection schemes. n this aer, Shanghai-Nanjing intercity railway was surveyed and taken for case study. The evaluation work in this aer covers four ower quality indices, including unbalanced three-hase voltages, negative sequence current injections into nearby generators, harmonic current injections and harmonic voltages at the PCCs. Additionally, the simulation results are comared with the ractical measured data to verify the efficiency of the roosed evaluation method. System model As the fundamental information of the evaluation, the three-hase ower flow results, such as voltages on each bus and currents of each branch in the ower system must be firstly calculated, then, the further assessment can be undertaken. Therefore, a detailed model of the whole system that is comosed of the ublic ower system and the traction ower-suly system should be established. n this section, the ower system model is firstly introduced, followed by the traction ower-suly system model..1 Power system model For the unbalance analysis and the harmonic analysis of a high-voltage network with HSR traction loads, the asymmetry, nonlinearity and frequency characteristics of the ower system comonents must be recognized. All detailed network comonent models, including generators, 48 PRZEGLĄD ELEKTROTECHNCZNY (Electrical Review), SSN , R. 88 NR 11a/01

2 transmission lines, transformers and loads, are comletely rovided on the ETAP latform. Given the required data for the unbalance analysis and the harmonic analysis, these existing models can be directly alied in this study. The rimary considerations for the data requirements of network comonent models are briefly resented as following 1) Generators: n the generator model, the internal sequence imedances are adoted to describe the inherent generator hase unbalance resulting from the system unbalance. At the harmonic frequencies, the equivalent imedance of a generator aroaches its negative sequence imedance. Thus, the sequence imedance data are required for the generator model. ) Transmission lines: Transmission line is reresented by a multihase couled model in which the data of ositive, negative and zero sequence imedance and admittance are needed. Harmonic imedance of a transmission line is determined by the imedance at the fundamental frequency considering the adjustments cause by harmonic frequencies. 3) Transformers: -winding and 3-winding transformers are included in this aer. The coer and core losses, the off-nominal taing, the hase shift and so on are considered in the models. And, the ositive and zero short circuit imedance, the winding connections and the grounding tye are included in the concerned transformer data. Also, the resistance and inductance comonents of the transformer short circuit imedance are frequencydeendent. 4) Loads: The balanced loads in grid which are comosed of constant imedance, constant current and constant ower comonents are reresented by the olynomial models. f not roducing the harmonics, the equivalent harmonic imedance of the load can be determined by its basic load.. Traction ower-suly system model Several transformer connection schemes are commonly used in HSR substations, such as single-hase, V-V, Scott, Le-blanc, Wood-bridge, and YNd11. Each connection scheme has its own roerties, which result in different effects on the ower system. oth accurate and simlified models for the traction substations have been comletely develoed and been used for the evaluation [9,10]. n order to accommodate to all transformer connection schemes, a general method based on the rincile of equivalent transformation is utilized here [11]. The traction transformer is considered as a multi-ort network as shown in Fig. 1. Each traction arm of the substation is viewed as one ort. Figure 1. Traction transformer ort model C U U C A n this model, the transformer is assumed to be ideal. Therefore, no imedance is included in the transformer and, U A U U as a result, no voltage dro and loss are incurred. esides, the three-hase system at the rimary side is assumed to be symmetrical. The a-hase voltage is selected as the reference voltage. The voltage and the current of the ort (=α, β) of the secondary side are shown as following j A j( ) e (1) U U e 3U K e () j where, U is the voltage of ort ; is the load current of ort ; K is the ratio of traction ort voltage and the rimary line voltage; ψ, which is defined as the connection angle of ort, is the lagging hase angle of U to U A ; φ is the ower factor angle of current lagging to voltage U at ort. When a ort at the secondary side works indeendently, no zero sequence is included at the rimary side. Thus, three-hase currents caused by ort satisfy (3) 0 A C Consider the ower balance equation for the two sides of the traction transformer (4) U U U U * * * * A A C C Combining equation (1) - (4), A,, C can be solved easily. y the suerosition rincile, when all orts work, the total three-hase currents at the rimary side can be achieved as (5) A 1 1 j a a K e 3, C 0 a a j K e, where, a equals to e jπ/3. Then, the currents in equation (5) are transformed to three-hase ower load to simulate the unbalance characteristics of traction loads. Multilying the harmonic current content ratio of locomotives by the currents in equation (5), the traction substation can be constructed as a harmonic current source. The traction substation model above based on the rincile of general equivalent transformation can be alied to all transformer connection schemes. According to the connection angles, the commonly used traction transformers can be classified into two grous. For some connection schemes, such as V-V and YNd11, the connection angle difference of two traction orts is 10. While for other connection schemes, such as Scott, Leblanc and Wood-bridge, the connection angle difference is 90.When the connection angle is determined, the general model is secified and can be alied in this study. 3 Case study A case study using Shanghai-Nanjing intercity railway system is resented in this section to demonstrate the efficiency of the roosed method in estimating the overall imacts on the ower system with HSR loads. Shanghai- Nanjing intercity railway is the first high-seed electrified railway directly connected to 0kV voltage level in Jiangsu Province. 5 traction substations are included along the railway route, namely, aohuashan, Danyang, Changzhou, Wuxi East and Kunshan. V-V connection transformer is adoted by all these substations and each substation is sulied by two indeendent incoming feeders from 0kV PRZEGLĄD ELEKTROTECHNCZNY (Electrical Review), SSN , R. 88 NR 11a/01 49

3 high-voltage grid. 9 generators, 93 transformers and 1345 transmission lines are contained in the 0kV network of Jiangsu Province. The degree of ower quality effects caused by HSR traction loads deends on the electric locomotive tye, the traction transformer connection schemes, the trains movements, etc.. Under different conditions, the load currents of two traction arms are very different. n this section, three tyical scenes are reviewed for the ower quality issue of the railway as following 1) Scene 1: currents of heavy load arm and light load arm are calculated by RMS. ) Scene : currents of heavy load arm and light load arm are resectively calculated by 95% RMS and RMS. 3) Scene 3: currents of heavy load arm and light load arm are resectively calculated by maximum and RMS. Namely, the normal, 95% robability and maximum oeration conditions of the railway are resectively taken into consideration. And, the load currents of traction substations are obtained by the field measurement. Four ower quality indices of the evaluation of negative sequence comonents and harmonics are resectively discussed in this section. 3.1 Unbalanced three-hase voltage at the PCCs n order to reduce the negative sequence comonents effects caused by HSR loads, the arrangement of the exchange hase between traction substations (i.e., substations in succession are fed on different hases alternatively) are usually adoted by the railway deartments. Thus, the traction load is almost equally distributed in the three-hase. According to G/T , the unbalance degree of three-hase voltage is ermitted to be no more than % in normal situation. The simulation values and measured values of voltage unbalance degree at the PCCs are listed in Table 1. Table 1.Three-hase voltage unbalance degree at the PCCs [%] PCC Simulation value Measured value Scene 1 Scene Scene 3 Average 95% Dongyang Yuheng Dongqing Xiangnan Shiai As can be seen from the above table, three-hase voltage unbalance degree at each PCC is within the safe limit. The greatest value of unbalance degree aears in scene 3, followed successively by scene and scene 1. n scene 1, the greatest voltage unbalance degree, u to 0.341%, occurs at Xiangnan caused by Wuxi East substation. n the other two scenes, the voltage unbalance degree at Yuheng caused by Danyang substation is greatest, which reaches to 0.369% and 0.43% resectively. The oerating condition of scene is consistent with that of 95% robability measurement level. As shown in table 1, these two sets of data are very close excet at Dongqing affected by Changzhou substation, which confirms the validation of the roosed evaluation method. 3. Negative sequence current injections into nearby generators n general, HSR traction loads are connected to highvoltage level on the ublic ower system to increase the short-circuit caacity at the PCCs and therefore reduce the unbalance effect. However, this makes substations so electrically close to the generators in ower system that more concerns are attracted for the excessive negative sequence current injections into nearby generators [1]. The maximum negative sequence current allowed of synchronous generator is rovided in G n continuous oeration, the negative sequence current 1 should be less than 8% of rated current N. Several ower lants nearby the HSR route are selected and the equivalent negative sequence current injections are comuted. Table shows the simulation values under the severest circumstance (scene 3). Table. Maximum negative sequence current injections into nearby ower lants Power Plant Caacity [MW] Equivalent Equivalent N [A] 1 / N [%] Jinlingranmei Jinlingranji * Zhenchang *137.5+* Zhenchang * Jianbi 4* Jianbi * Changzhou * Ligang *600+* Wangting * Wangting * Changshu 3* As can be seen from Table, the ercentages of the equivalent negative sequence current are all less than 1% in scene 3. Wangting whose ercentage is u to 3.853% is affected most severely. t should be noted that the maximum load condition at each PCC does not certainly occur at the same time. Consequently, the simulation results shown in Table are comaratively conservative. Generally seaking, the ability of a generator to withstand the negative sequence current is roortional to its caacity. esides, the closer the electrical distance between traction substation and generator is, the severer the negative sequence effect is. For those PCCs whose unbalance degree is very low due to high short-circuit caacity (i.e., very electrically close to some generators), secial attention should be aid to the ossible excessive negative sequence current injections into nearby generators. 3.3 Harmonic current injections into the ower grid According to G/T , the harmonic current injections into PCCs from all users should be restricted within allowed limits, however, the allowed values can t be used for assessment directly. Firstly, they should be converted by the ratio of the actual short-circuit caacity and the reference short-circuit caacity. Secondly, they should be distributed by the ratio of the user s rotocol caacity and the ower suly caacity at that oint. Fig. shows the harmonic current injections into ower system from Wuxi East substation in 3 scenes. The simulation results and the allowed values of several characteristic harmonic currents are listed in Table 3. t can be seen from Fig. and Table 3 that the harmonic currents in scene 3 are highest, but still not exceeding the allowed value. Comaring the harmonic currents of all orders, 3 rd, 5 th, 7 th, 11 th, 13 th and 37 th ones are relatively redominant. Since ulse width modulation (PWM) technology has been alied in Shanghai-Nanjing intercity railway, its harmonic characteristics differs from that of conventional electrified railway in which the AC-DC tye electric locomotives are generally used. On one hand, loworder harmonic comonents are greatly reduced; on the other hand, the harmonic sectrum range becomes wider and the high-order harmonic comonents increase obviously. 50 PRZEGLĄD ELEKTROTECHNCZNY (Electrical Review), SSN , R. 88 NR 11a/01

4 5 harmonic current [A] 4 Scene 1 Scene Scene harmonic order Figure. Harmonic current injections from Wuxi East substation Table 3. Harmonic currents injected from Wuxi East substation [A] Harmonic order Limit Scene Scene Scene voltage distortion [%] Scene 1 Scene Scene harmonic order Figure 3. voltage HD on Xiangnan bus Table 5. voltage HD and THD on Xiangnan bus [%] Harmonic order THD Scene Simulation Value Scene Scene Measured Value Average % Harmonic voltages at the PCCs According to G/T , the limits of harmonic voltage distortion in ublic grid are listed in Table 4. t should be noted that the standard of 0kV voltage level alies to the one of 110kV level. Table 4. Public grid harmonic voltage limits Standard HD [%] THD [%] voltage [kv] Odd Even y harmonic analysis, the voltage individual harmonic distortion (HD) and the total harmonic distortion (THD) at the PCCs can be achieved easily. Fig. 3 shows the voltage HD on Xiangnan bus connected to Wuxi East station. esides, simulation data and ractical measured data of several characteristic harmonics are comared in Table 5. Obviously, the voltage THD at each PCC is less than %, the voltage HD of each odd order is below 1.6% and the HD of each even order is below 0.8%. Thus, the national standard requirements of harmonic distortion indices are fulfilled. Although the harmonic values of all orders increase from scene 1 to scene 3, they are very close in each scene esecially at low-orders. And, the voltage HDs of oddorders, such as 3 rd, 5 th, 11 th, 13 th and 37 th are high. t can also be observed that the harmonic voltage sectrum is wide and high-order harmonic comonents are great, which coincides with the conclusions drawn above. The simulation data in scene is in accordance with the measured data of 95% robability at the characteristic orders such as 3 rd, 5 th, 11 th, 13 th, 3 rd and 5 th, which verifies the efficiency of the roosed evaluation method again. 4 Conclusion A method was resented in this aer, caable of evaluating the overall imacts on a ower system with HSR traction loads. ased on the owerful simulation caability of ETAP latform, the method can handle the large-scale ower system with numerous network elements. esides, it is ossible to deal with the various transformer connection schemes by adoting the general model of traction PRZEGLĄD ELEKTROTECHNCZNY (Electrical Review), SSN , R. 88 NR 11a/01 51

5 substation. y using this method, both the effects at the PCCs and the enetration in the ower system can be evaluated with sufficient accuracy. With the significantly increasing demand for ower of HSR traction systems, it is critical that an in-deth imact evaluation for these secial loads be erformed by ower comanies. A feasible solution to this roblem is resented in this aer, and it has a significance of being referenced and imroved for further study. REFERENCES [1] Antonio D.A., Maria M., Vito G.M., Pericle Z., New Power Quality Assessment Criteria for Suly Systems Under Unbalanced and Nonsinusoidal Conditions, EEE Transactions on Delivery, vol.19, no.3, July 004, [] Golovanov N., Lazaroiu G.C., Roscia M.; Zaninelli D, Voltage Unbalance Vulnerability Areas in Power Systems Sulying High Seed Railway, Power Engineering Society General Meeting, San Francisco, CA, USA, June 005 [3] Peechin T., Pohchiang L., Donaldgrahame H., Otimal medance Termination of 5-kV Electrified Railway Systems for mroved Power Quality. EEE Transactions on Power Delivery, 005, 0, [4] Petere S., Marek W., Markf M., System macts Evaluation of a Single-hase Traction Load on a 115-kV Transmission System. EEE Transactions on Power Delivery, 006, 1, [5] Yuquan L., Guoei W., Huangsheng H., Li W., Research for the Effects of High-seed Electrified Railway Traction Load on Power Quality. nternational Conference on Electric Utility Deregulation and Restructuring and Power Technologies, Weihai, China, July 011 [6] Shaofeng X., Qunzhan L., A Practical Method for Assessment of Harmonic Emission of Electrified Railway, EEE ndustrial Electronics Society, France, November 006 [7] Wenshyan C., Jyhcherng G., inkwie C., Sanyi L., New Criteria for Estimating Voltage Unbalance Due to Secial Connected Transformers in High Seed Railway Systems, nternational Journal of Emerging Electric Power Systems, vol.4, no.1, 005 [8] Haiqun W., Yingjie T., Qinchang G., Evaluation of Negative Sequence Current njecting into the Public Grid from Different Traction Substation in Electrical Railways. nternational Conference and Exhibition on Electricity Distribution, Prague, Czech Reublic, June 009 [9] Chenging H., Chijui W., Yungsung C., Shihkai P., Jungliang Y., Minghong H., Loading Characteristics Analysis of Secially Connected Transformers Using Various Power Factor Definitions. EEE Transactions on Power Delivery, vol.1,no.3, July 006, [10] Mohsen K., Mohammad J.S., Siamak F., Seyed S.F., Modeling and Comarison of Tranction Transformers ased on Utilization Factor Definitions, nternational Review on Modelling and Simulations, vol.4, no.1, February 01, [11] Qunzhan L., Power Suly and Comensation Analysis for Traction Substation, eijing: China Railway Press, 006 [1] Shilin C., Rayjong L., Paohsiang H., Traction System Unbalance Problem Analysis Methodologies, EEE Transactions on Power Delivery, vol.19, no.4, October 004, Authors: Dr. Qingshan Xu works in School of Electrical xuqingshan@seu.edu.cn; Qiuqi Zhu studies in School of Electrical autqisky@16.com; Haixiang Zang studies in School of Electrical zanghaixiang@seu.edu.cn; Xiaodong Yuan works in Electrical Power Research nstitute of Jiangsu, Nanjing, China, 11103, lannyyuan@hotmail.com. 5 PRZEGLĄD ELEKTROTECHNCZNY (Electrical Review), SSN , R. 88 NR 11a/01