DYNAMIC PERFORMANCE OF THE EAGLE PASS BACK-TO-BACK HVDC LIGHT TIE. Å Petersson and A Edris ABB Power Systems AB, Sweden and EPRI,USA

Transcription

1 DYNMI PERFORMNE OF THE EGLE PSS K-TO-K HVD LIGHT TIE Å Petersson and Edris Power Systems, Sweden and EPRI,US INTRODUTION Eagle Pass ack-to-ack (t) Tie is a Voltage Source converter (VS) -based tie interconnecting the transmision grid of Texas with the Mexican Power system, at the EP s Eagle Pass Substation in the State of Texas. The Tie was put in operation in July 2. s the VS takes advantage of Gate-Turn-Off type of power semiconductors, the dynamic performance and capability are significantly improved compared to the conventional thyristor based technologies. This paper focuses on highlighting the performance of the VSbased technology, using representative computer simulations and comparable field tests. The presented performance covers dynamic voltage support, operational mode changes and energizing of islanded networks. The influence on the ratings of the VS main components is discussed, as well as the engineering tools used for simulations and validation of design. KEYWORDS: HVD, ack-to-ack, Voltage Source onverters, Pulse Width Modulation, Dynamic Performance. PROJET KGROUND The Eagle Pass substation is located in a peripheral area of the Texas transmission grid, it is operated by EP- PL. Figure illustrates the location on the US Mexico border. The Eagle Pass load area is supplied over two 38 kv transmission lines. The grid structure and the relatively large distance to significant generation, results in weak voltage support to the Eagle Pass area. Eagle Pass also has a 38 kv transmission line that cross-border ties into the Mexican transmission grid, operated by FE. This tie is normally open and is mainly used in emergency conditions to transfer load from the US to the Mexican grid. However, load being transferred needed firstly to be interrupted since it was not possible to interconnect the two asynchronous power systems. With a significant foreseen load growth for the load area it was eminent that voltage stability problems would in the near future be encountered at single contingencies. Studies performed indicated that a facility providing 36 Mvar of dynamic reactive power support at Eagle Pass substation would provide years of relief from single contingency situations. VS type of installation, i.e. a STTOM, would provide needed reactive support instantaneously and could maintain full reactive output at even lower voltage levels than what could result from the single contingency outage. Installation of a VS would be ideal for weak systems whereas the reactive support provided by shunt capacitors is not very effective as it decreases with the square of the voltage. Expanding the installation to a ack-to-ack would enable also uninterrupted bidirectional active power transfer between the US and Mexican grids, enhancing the reliability of the power supply. Thus a two-fold mission is accomplished by the dual VSs. dditionally the tie can be used to energize the Eagle Pass load area from the Mexican grid as the VS technology has an inherent black-start capability. TEHNOLOGY PPLIED simplified one-line diagram of the T tie in Eagle Pass is shown in Figure 2. Eagle Pass Piedras Negras VS VS Figure. Eagle Pass location. Figure 2. T simplified single line diagram.

2 The T scheme comprises two 36 MV Voltage Source onverters coupled to a common D capacitor bus. The VSs are of the NP, Neutral Point lamped, type. The VSs are equipped with IGTs, Insulated Gate ipolar Transistors, operated with PWM, Pulse Width Modulation. The power semiconductor equipment and the D capacitor bank are of indoor design. The IGTs are water cooled, utilizing a closed loop cooling system with water-to-air heat exchangers. Each VS is then on the respective ac terminal connected to phase reactors, which each in turn is connected to a conventional step-up transformer, interfacing the t to respective grid. The ac output voltage from the VSs has a nominal value of 7.9kV, which is stepped up to 38kV. Harmonic filters, tuned to the 2 th and 39 th harmonic respectively, and with high pass characteristic, are shunt connected to the 7.9kV bus. Total filter rating is approximately 6 Mvar on each side. Due to the high frequency PWM switching applied, the filters can be kept small in rating and their respective tuning is not critical. The whole t installation is located on the US side at the Eagle Pass substation. The substation is located in a residential area and specific consideration has been taken to control audible noise produced by some of the components. The layout of the t installation is shown in Figure 3. ooling towers Pump room Pump room Valve hall Phase reactors ontrol room Filters Spare transformer Figure 3. Eagle Pass t layout. To FE Power transformers To EP-PL The t is designed to operate in two control modes. The primary mode is reactive power support for voltage control of the Eagle Pass bus and the secondary mode is active power flow across the tie. The t has the capability, within its MV rating, of providing independent control of the voltages at the US and Mexican grids, as well as the real power transferred between the two grids. In voltage control mode, the t operates with a control slope to supply reactive power for voltage support. To transfer power in either direction the t operates in power control mode, it will continue to transfer power if the voltage at Eagle Pass is within a dead-band. For conditions outside the dead-band the t switches into the Voltage control mode. The dead-band is designed so that local capacitor switching or changes in remote generation which causes slight voltage swings do not cause the t to switch to voltage control mode. VERIFYING DYNMI PERFORMNE The final verification of the dynamic performance characteristics was realized as part of the t tie commissioning. Then using adequate data acquisition equipment, real time recordings are taken of both main circuit parameters and internal tie controller variables. The Eagle Pass tie controller includes a built-in powerful Transient Fault Recording module with submillisecond resolution, which is capable of capturing also the inner control loop dynamics. This functionality also enables stored data to be digitally transferred to a remote user, e.g. the design engineer can have direct just-in-time access to the tie controller. In the design phase of the t installation, extensive digital simulations were performed, as part of the development of the tie controller, and specifically the tuning of its important parameters. These simulations also to some extent verify the configuration of the main power circuit, e.g. the interaction between the main VSs and the harmonic filters can be studied. The simulations have covered a large number of cases and events, out of which only a few could be conducted in the field. When modeling electrical systems involving power semiconductor converters and extensive networks, the user interface becomes a critical factor. User friendly interfaces built around functional block representation is today a proven technique. For the Eagle Pass project a comprehensive model was developed using the EMTD program. The complete model was built up with a number of functional blocks, such as: ac network, stepup transformer, phase reactor, harmonic filters, VS bridge, D capacitor and tie controllers. The ambition with the model was to represent the transmission grids accurately with respect to prospective resonances in the lower frequency range. The circuits within the t was modeled such that magnetic saturation phenomena could be identified. The representation of the IGT converters was accurate with respect to the chosen NP-topology and PWM switching at 26Hz, The digital tie controller was modeled with sampling rates of critical control loops exceeding 7kHz. When the individual models had been validated, the actual simulations were performed with different parameter set-ups for different cases. Using a high performance personal computer, and an on-screen control panel, very effective execution of the simulation can be achieved although the model becomes extensive. The design engineer can observe the behavior of the model as the simulation progresses and initiate actions accordingly. There are typically a large number of dynamic events needing analysis and verification. mong those the following are illustrated below in an attempt to highlight the t tie unique and promising performance:

3 . t power reversal 2. apacitor bank switching 3. lack start capability 4. Nearby fault on the US side. 5. Remote line fault on the US side For some of the cases below, the EMTD simulations are presented together with field recordings with similar set-up.. t Power Reversal The VS converter technology applied in Eagle Pass, allows for the transferred active power to be dynamically controlled over the full available range. s an example of this, Figure 4 illustrates the performance after ordering a power reversal. This case is illustrated in the shape of a plot from an EMTD simulation and the power order is changed from close to 36MW import to US, to 36MW export. lthough locally significant, this transient is absorbed by the respective power systems representations. Looking at the traces of Figure 4, the sub-cycle reversal of active power appears well controlled without any overshoot, an example of the very high performance that can be achieved with PWM controlled VS installations. The active power reversal is executed without any action needed by the voltage control. 2. apacitor ank Switching One possibility to check the transient response of the t was to switch on one of the capacitor banks at the Eagle Pass substation. In the actual case, the reactive power from the T before the switching was equal to 8 Mvar. s the capacitor bank was switched on, the T had to reduce its output by 5 Mvar, corresponding to the capacitor bank rating, down to 3 Mvar. The t controller responded in a well damped manner within a cycle as shown in subplots 2, 3 and 7 of Figure U+ U- PI 29 4;34;23 Uac Primary Sys PI 29 4;34;23 Iac P PI 29 4;34;23 Iac Sys PI 29 4;34;23 Uac Sys PI 29 4;34;23 Uac S PI 29 4;34;23 Udc Sys.8.5 P Q PI 29 4;34;23 PQ Ref Sys Figure 4. Power reversal at Eagle Pass (EMTD simulation) Subplot. EP-PL 38 kv voltages. Subplot 2. EP-PL step-down transformer secondary Subplot 3. ack-to-ack power reference; EP-PL and FE reactive power reference Subplot 4. EP-PL active and reactive power. Subplot 5. FE active and reactive power Figure 5. apacitor bank switched on at Eagle Pass Subplot. EP-PL 38 kv voltages. Subplot 2. EP-PL step-down transformer secondary Subplot 3. EP-PL Phase reactor Subplot 4. EP-PL 7.9 kv voltages. Subplot 5. EP-PL 7.9 kv phase-to-ground Voltages in kv. Subplot 6. D voltages. Subplot 7. EP-PL converter ctive (bottom curve) and Reactive Power. It should be noted in Figure 7 that the switching causes also a ms transient in the active power. lthough not significant, the switching produces a small angular

4 deviation in the bus voltage, relative the VS output voltage. s a consequence a transient in the active power/d voltage appears. On the other hand it is obvious that the t tie completely isolates the Mexican power system from the switching transients. 3. lack Start apability Next, to illustrate the lack Start function, the FE side of the T tie was first energized as during normal energization. On the EP-PL side the line breaker, connecting the t tie to the EP-PL network, was open but was in the control system simulated as if it had been closed. In this way, a small islanded network consisting mainly of the EP-PL transformer was created. ny larger islanded network was not possible to create, as this would have forced a power outage in parts of the city of Eagle Pass. lack Start was then initiated. For comparison Figure 6 shows the result of an EMTD simulation where the black start is performed against a load of approximately 25MW. oth the EMTD simulation and the field recording (in Figure 7) show that the t energizes and picks up load in a controlled way without noticable transients U+ U- P Q PI 29 7;7;56 Uac Primary Sys PI 29 7;7;56 Iac P PI 29 7;7;56 Iac Sys PI 29 7;7;56 Uac Sys PI 29 7;7;56 Uac S PI 29 7;7;56 Udc Sys PI 29 7;7;56 PQ Ref Sys Figure 7. lack start. Subplot. EP-PL 38 kv voltages. Subplot 2. EP-PL step-down transformer secondary currents Subplot 3. EP-PL Phase reactor Subplot 4. EP-PL 7.9 kv voltages Subplot 5. EP-PL 7.9 kv phase-to-ground Voltages Subplot 6. D voltages. Subplot 7. EP-PL converter ctive (top curve) and Reactive Power (bottom curve) Reference. s shown in Figure 7, The EP-PL 38 kv voltage was ramped from kv, at a slower pace than in the digital simulation (Figure 6). The ramp ended at 38 kv although the plot, due to TFR limitations, does not show more than the part up to about 6 kv. On the FE side, the voltages and currents remained undisturbed during the ramping phase. The only load on the EP-PL side consisted of the harmonic filters rated totally 6 Mvar. The system was anyhow observed to perform well. The modulation on the D voltage in Subplot 6 is due to the frequency difference between EP-PL side, having constant frequency, and the FE side. On the FE side, the system operated as normal. Figure 6. lack Start with 25 MW load (EMTD simulation) Subplot. EP-PL 38 kv voltages. Subplot 2. EP-PL step-down transformer secondary Subplot 3. ack-to-ack power reference; EP-PL and FE reactive power reference Subplot 4. EP-PL active and reactive power. Subplot 5. FE active and reactive power 4. Nearby Fault on the US side Near-by -phase to ground faults on the US side was simulated in EMTD. For such a fault the primary voltage to be controlled, drops significantly. In order to avoid raising the voltage at fault clearing, the throughfault capacitive action is inhibited and the t stays at close to no load conditions during the fault. The post fault overvoltage reaches.3 p.u. in one phase, however

5 it is practically eliminated in one cycle. Some current transients are seen in Figure 8 below, these being the result of oscillations excited in the harmonic filters. For this severe unbalanced fault it is worth noting that the transients seen on the Mexican side is practically negligible. The remaining disturbances are caused by the need to stabilize the D voltage against the fluctuating active power caused by the unbalanced fault. Figure 8. Near-by fault on US side (EMTD simulation) Subplot. EP-PL 38 kv voltages. Subplot 2. EP-PL step-down transformer secondary Subplot 3. ack-to-ack power reference; EP-PL and FE reactive power references Subplot 4. EP-PL active and reactive power. Subplot 5. FE active and reactive power 5. Remote Fault on the US side The final case to be highlighted is a remote fault on the US side. Where initially the T was in operation at close to zero active power. Lightning conditions in a remote area caused a voltage dip in the US transmission grid. This case was simulated also in EMTD as is illustrated in Figure 9. The field recording is shown in figure for comparison. Following the initial voltage drop in Eagle Pass, the t current during the fault condition was increased to almost p.u. (capacitive) to support the bus voltage at Eagle Pass. When comparing the digital simulation with the field recording it can be seen that the reactive power transient in the simulation appears more exponential, indicating that the simulated fault creates a truly stepwise change in voltage. Otherwise the simulation and the field recording show good correlation. Figure 9. Remote fault on US side (EMTD simulation) Subplot. EP-PL 38 kv voltages. Subplot 2. EP-PL step-down transformer Subplot 3. ack-to-ack power reference; EP-PL and FE reactive power references Subplot 4. EP-PL active and reactive power. Subplot 5. FE active and reactive power PI 293 7;;9 Uac Primary Sys PI 293 7;;9 Iac P PI 293 7;;9 Iac Sys PI 293 7;;9 Uac Sys PI 293 7;;9 Uac S PI 293 7;;9 Udc Sys PI 293 7;;9 PQ Ref Sys Figure. Remote fault case Subplot. EP-PL 38 kv voltages. Subplot 2. EP-PL step-down transformer secondary currents in mps. U+ U- P Q

6 Subplot 3. EP-PL Phase reactor Subplot 4. EP-PL 7.9 kv voltages. Subplot 5. EP-PL 7.9 kv phase-to-ground Voltages in kv. Subplot 6. D voltages. Subplot 7. EP-PL converter ctive (bottom curve) and Reactive Power (top curve) It is worth noting from figure 9 that the t tie isolates the Mexican transmission grid from the fault, as is indicated by the reactive power reference being stable through the simulation. ONLUSIONS The t HVD Light tie at Eagle Pass is a first-of-itskind installation, where VS technology is used for the dual purpose of transferring active power between two asynchronous transmission systems and providing dynamic voltage support to the respective grids. The dynamic performance and characteristics of the t tie have been illustrated using digital simulations and field recordings. mong the features highlighted are instantaneous voltage support, black start of islanded networks and isolation of the grids with respect to switching transients and faults. State-of-the-art simulation software enables the design engineer to perform efficient simulations of the t tie on a standalone personal computer. cknowledgements EMTD is a simulation software package owned and developed by the Manitoba HVD Research enter, Winnipeg, anada. The Eagle Pass t tie project was a joint effort between EP, EPRI and Power Systems. FE of Mexico also played an important role providing assistance in many engineering matters. Today EP operates the tie and an operation coordination with FE is established.