A new asynchronous transmission link will interconnect the U.S. and Mexico grids. The link will be between American Electric Power (AEP) and Mexico's Comisión Federal de Electricidad (CFE) and will be located in Laredo, Texas, U.S. This project is one of several reliability-must-run (RMR) exit-strategy projects for the Laredo Power Station and is endorsed by the Electric Reliability Council of Texas (ERCOT) board of directors for the purpose of maintaining reliable power supply to the Laredo area.
AEP selected the 100-MW variable frequency transformer (VFT) from GE Energy for this project. The VFT is a controllable, bidirectional transmission device that allows power transfer between two networks that might not be synchronized. Functionally, the VFT is similar to conventional high-voltage direct current (HVDC) converters and voltage source converters (VSC) that are arranged back-to-back (BTB) to provide power-transfer capability across asynchronous grids.
In the past, local generation supported the Laredo area, but that changed in 2002 when the generation owner notified ERCOT of its intention to shut down the Laredo Power Station due to market conditions. ERCOT responded by entering into an RMR agreement with the generation owner to keep the units operating until power-delivery system improvements could be constructed. AEP then began studying short-term solutions to reduce the unit run times until a new 345-kV transmission line could be brought on-line in 2010. Part of the short-term plan required that a fifth 138-kV source flow into the Laredo area. Given the presence of a strong 138-kV CFE network just across the border, and cross-border 138-kV transmission infrastructure already in place, an asynchronous tie to CFE made sense as an expedient short-term solution to provide this fifth source.
After careful analysis of alternative technologies, AEP selected the VFT, which will add the capability for import of up to 100 MW of real power to ERCOT from CFE, thus allowing more flexibility to serve the Laredo area load while meeting AEP, ERCOT and North American Electric Reliability Council (NERC) stability and reliability requirements.
The VFT project initial design and system study began in early 2005 and is scheduled for commercial operation in early 2007. A one-line system diagram of the VFT project is provided in Fig. 2.
The VFT is essentially a continuously variable phase-shifting transformer that can operate at any adjustable phase angle. The core technology of the VFT is a rotary transformer with three-phase windings on both the rotor and stator sides (Fig. 3).
The collector system conducts current between the three-phase rotor winding and its stationary bus work. In the case of the Laredo VFT system, the rotor side of the VFT is connected to the CFE grid and the stator side of the VFT is connected to the ERCOT grid. This arrangement is arbitrary and could have been configured the other way just as easily.
Power flow is proportional to the magnitude and direction of the torque applied to the rotor. This torque is applied to the rotor by a drive motor, which is controlled by a variable-speed drive system. If torque is applied in one direction, then power flows from the stator windings to the rotor windings. If torque is applied in the opposite direction, then power flows from the rotor windings to the stator windings. If no torque is applied, then no real power flows through the rotary transformer.
A closed-loop power regulator maintains power transfer according to the operator setpoint. The regulator compares measured power with the setpoint, and adjusts motor torque as a function of power error. The power regulator will respond quickly to network disturbances and maintain stable power transfer.
Regardless of power flow, the rotor inherently orients itself to follow the phase angle imposed by the two asynchronous systems, and will rotate continuously if the grids are at different frequencies. The motor and drive system are designed to continuously produce torque while at a standstill. If the power grid on one side experiences a disturbance that causes a frequency excursion, the VFT will rotate at a speed proportional to the difference in frequency between the two power grids. During such a disturbance, if the VFT is transferring power, it will continue without interruption and at full-expected power. The VFT is designed to continuously regulate power flow with drifting frequencies on both grids.
Reactive power flow through the VFT follows conventional ac circuit rules. It is determined by the series impedance of the rotary transformer and the difference in voltage magnitude on the two sides. And, unlike power-electronic alternatives, the VFT produces no harmonics and cannot cause undesirable interactions with neighboring generators or other equipment on the grid.
Because the Laredo area is vulnerable to dynamic voltage collapse, particularly during summer peak-load conditions, the studies performed to evaluate the prospective asynchronous devices consisted of power-flow analysis and dynamic-stability analysis. In the dynamic studies, detailed modeling of the ERCOT transmission network and connected generation was included in its entirety. Existing flexible alternating current transmission systems (FACTS) devices — including the Laredo and Military Highway ±150-MVAR STATCOM and the Eagle Pass 36-MW VSC BTB tie — were modeled. Alternate representations of the CFE system also were considered but did not affect results.
Special consideration was given to the modeling of the South Texas area load. A primary objective was to simulate the dynamic behavior of a heavy concentration of air-conditioning load that would be typical of a summer demand peak. To accomplish this, an aggregate load model was derived from estimated load class percentages, typical summer peak composition data, and typical load device modeling data. The resulting model consisted of approximately two-thirds of the load represented as dynamic induction machines and one-third as a static polynomial-type load model. This special load model is applicable to transient voltage-collapse studies and was applied at each South Texas area load bus.
Evaluation of the asynchronous interconnection centered on the following three different asynchronous tie device types: Conventional HVDC BTB, VSC BTB and VFT. Dynamic models from individual vendors were integrated separately into the base study case, thereby providing a basis for comparative performance evaluations among the various asynchronous devices.
A set of N minus one (N - 1) contingency events involving three-phase faults and nonfault-initiated line tripping on the 138-kV network supplying Laredo and the vicinity was simulated. Due to induction machine stall tendencies, the fault cases proved to be the limiting contingencies for power import. In every case, the system post-fault response exhibited an acute transient voltage-recovery problem, even with the added reactive injection supplied by the Laredo STATCOM. If the initial power import into Laredo was high enough, voltages did not recover. Thus, the power import limitation into the Laredo area is defined by transient voltage collapse. (Figures 4 and 5 illustrate a typical fault case simulation with the VFT in service. Note the delay in post-fault voltage recovery in Fig. 5.)
Successive load year cases spanning 2007 to 2010 were established based on forecasted load growth with the objective being to determine how each of the asynchronous devices would support that growth. The study time horizon of 2010 corresponds to a maximum Laredo area load of 539 MW. The stability study results are summarized in Fig. 6. The best performing of the conventional HVDC BTB device is exhibited here. This indicates the magnitude of load that may be stabilized by injection of power from CFE through each asynchronous interconnection technology. The area to the right and below each plot is the unstable region. Increasing area load requires increasing steady-state real power imports to maintain stable operation.
Also in Fig. 6, the VFT and VSC BTB devices show an increasingly substantial benefit as area load is increased compared to the conventional HVDC BTB. The VFT at 100 MW is observed to have a small but consistent advantage over the 150-MW VSC BTB. While the 150-MW VSC BTB transient injection of reactive power helps to stabilize system voltage, the 100-MW VFT combined transient flow of real and reactive power is more beneficial even though the overall reactive injection is significantly less than that of the 150-MW VSC BTB. The benefit of supplying an immediate real power boost into the faulted system is that it helps to keep the Laredo area voltage angle from lagging further behind the ERCOT system. This real injection occurs just after the fault is cleared (Fig. 4). A smaller ERCOT-Laredo angular separation translates into a smaller voltage dip during the post-fault swing as the induction machines reaccelerate.
Another noteworthy result in Fig. 6 is that the VFT can supply a stability benefit with zero-scheduled steady-state power flow at the 2007 peak load (469 MW). This is a benefit of the inherent characteristic of the VFT that makes it look more like a phase-shifting transformer than does a conventional HVDC BTB or a VSC BTB. As a result, a transient real and reactive power boost through the VFT is attainable at zero steady-state flow. The VSC BTB also can remain in service and supply reactive power at zero-scheduled steady-state power flow. However, a conventional HVDC BTB must be turned off completely.
The ability of the asynchronous device to ride through severe voltage depressions is essential for this application. An interruption of the device transmitting capability during a post-fault voltage-recovery period could result in a system collapse. The AEP specification included several scenarios to address this requirement. This also had been a requirement for the Langlois VFT. GE Energy demonstrated this capability via extensive simulator testing, and there have been field events that substantiate the expected behavior from the initial VFT installation.
Steady-state system requirements are related to the thermal capacity of the 138-kV transmission system connecting Laredo with the rest of ERCOT. The asynchronous device import capability offsets the need to import power from ERCOT, relieving the 138-kV network loading. And so the steady-state power-import requirements are a simple function of the asynchronous device capacity.
The conventional HVDC BTB and VSC BTB devices, rated at 150 MW, were obviously better in this regard in comparison to a single 100-MW channel VFT. However, further power-flow studies showed that a 100-MW injection was still adequate to achieve the required power import relief within the study time horizon. Therefore, the steady-state system requirements were not a key factor in the selection process.
Continued operation of the Laredo Power Station, while offering much toward stabilizing system voltage-collapse tendencies, is dependent on favorable market conditions and economic factors that cannot be assured. While both the VFT and the VSC BTB were able to achieve necessary stabilization without running the Laredo Plant throughout the study period, the VFT was selected because it offered better overall performance with regard to the stability requirements. This is indicated in Fig. 6 as the VFT and VSC BTB graphs exceed the 539-MW Laredo area load level at steady-state power injections still lower than their respective rated values. The best-performing conventional HVDC BTB device was unable to achieve system stabilization beyond 500 MW of total area load, thus running the Laredo Plant would be necessary in the latter years of the study period.
ERCOT determined the cost of the RMR contract to the market and requested AEP find a transmission solution to eliminate the RMR contract. A new 345-kV transmission line is required for the long-term support of the Laredo area, but cannot be placed in service before 2010. The solution for the interim period from 2007 to 2010 includes the asynchronous tie to CFE for which the VFT option was chosen. The VFT will operate continuously at 0 MW for reliability. This operation will substantially decrease the need for the RMR unit, which will still be retained and utilized if needed for reliability.
This approach will allow AEP to increase the transfer limit into the Laredo area up to 450 MW. The VFT also will allow even higher transfer levels into the area without the two Laredo 35-MW units required to operate and with reduced run time on the 110-MW unit throughout the year. If the 110-MW unit is down for maintenance or trips, then ERCOT has the flexibility to schedule energy across the tie from CFE. At high load periods prior to completion of the 345-kV transmission line, ERCOT has the ability to preemptively schedule energy from CFE in preparation of the worst single contingency for the Laredo area.
Rob O'Keefe received bachelor's and master's degrees in electric power engineering from Purdue University. Between 1983 and 1990 he was employed with General Electric's Power Systems Engineering department. Since 1990, he has been with American Electric Power Service Corp.'s Transmission Planning Group in Gahanna, Ohio, U.S. firstname.lastname@example.org
David Kidd received his BSEE degree from New Mexico State University in 1987. After working at El Paso Electric for 10 years in system protection, he joined the Texas Transmission Planning group at American Electric Power in Tulsa, Oklahoma, U.S. email@example.com
|Condition||HVDC BTB||VSC BTB||VFT|
|Stability performance beyond N-1||150 MW only adequate through 2008||150 MW adequate through 2010||100 MW adequate beyond 2010|
|Blackstart capability||Requires a synchronous condenser||Untested||Demonstrated the ability|
|SSTI* potential with adjacent generators||Known to be capable of SSTI||Known to be capable of SSTI||Inherently avoids SSTI|
|Harmonics||Requires a large amount of filtering (changes with power level)||Requires less constant filtering||No harmonics generated so no filters required|
|Coordination with future plans||Conventional but not adequate||Adequate||Adequate|
|STATCOM support 2007-2009||Required in 2009||Not required||Not required|
|*Subsynchronous torsional interaction (SSTI).|