Gas turbines are expensive. Although peaking units aren’t as costly as baseload units, letting them sit idle is still a waste. Yet that is what peaking units spend nearly all their time doing. Some operators only fire up their units a few days a year. That is like putting money in the bank and collecting interest a few hours at a time, rather than all year round. That strategy only works when the payback is extremely high for those short periods.
The economic model is changing, however, both for good and bad. Peaking turbine owners used to be able to predictably rely on being called into action on summer afternoons. Now utilities are switching to demand curtailment strategies rather than purchases on the spot market. The good news is that increasing use of wind and solar power require peaking to be available at odd times, and that requires dispatchable generating resources to be standing by to make up for sudden shifts in the wind. In addition, the use of renewable power sources and the actions of shutting down generating stations near urban areas increases the need for grid stabilization.
According to the Operational Reliability Analysis Program (ORAP), which contains more than 20 years of data on turbine performance, gas turbines in North America are being started more frequently and being run for shorter periods of time. Strategic Power Systems Inc. (SPS) analyzed the ORAP data and found the average number of starts per unit increased from 77 per year from 1996 to 2000, to 100 per year from 2006 to 2010. During that same period, the service factor (the number of hours the unit was providing power divided by the number of hours the flame was on) dropped from 62% to 37%, so gas turbines are spending more time idling than generating.
"We are seeing market-driven cyclic loads tied to such factors as wind intermittency," said Tom Christiansen, senior vice president of SPS. "Fast start units are needed to keep the grid stable."
Meeting these needs, and increasing the economic return on peaking turbines, requires a different approach to turbogenerator sets. Instead of viewing them as a unit, using a clutch to separate the turbine and the generator in order to increase their flexibility allows them to provide a wider range of services.
The power market includes generation, transmission, and distribution. To operate efficiently and reliably, these three elements must work in close alignment. Prior to deregulation, the utility could decide how best to allocate resources to meet these three aspects and get power to the consumer. With deregulation, responsibility for these functions has been split among different entities. Though market-based pricing has improved efficiencies in some areas, where markets don’t exist for items such as spinning reserve or reactive power, there is a misallocation of resources that didn’t exist prior to deregulation.
In the early days of deregulation, there was a rush to install more peaking turbines and take part in the then extremely lucrative and highly volatile market. But as the market matured, utilities were better able to predict their demand, and we no longer see the wild speculation and market manipulation that led to the electricity crisis in California in 2000 and 2001.
Utilities have also started adopting strategies for curtailing demand when needed. As Ray Dotter, spokesperson for PJM, a regional transmission operator that manages a wholesale power market and transmission system serving 50 million people, stated, "A decrease in usage is just as valid as an increase in generation, and sometimes it is a lot faster, because you can open a circuit a lot faster than you can ramp up a generator."
PJM, therefore, which first implemented a market for spinning reserve in 2002, later modified the market to allow commercial and industrial customers to sign up for demand response programs and participate in the reserve market as well. The market was renamed "synchronous reserve" to reflect that it no longer simply consisted of spinning turbines.
Similarly, the province of Ontario has a three-tiered demand response program for industrial facilities and commercial buildings. Options range from voluntarily reducing power on request to agreeing to let the utility demand that usage be curtailed by a specified amount.
Utilities are also adopting a variety of strategies to curtail peak power usage by residential customers. For decades they have used the power of persuasion, urging people to shut off their appliances on summer afternoons. That approach helped, but there was no economic incentive to cut usage that was comparable to the added cost of providing the peak power. The advent of smart meters has changed that, allowing utilities to use time of day pricing (in limited locales) to give consumers another reason to turn down their air conditioners. A two-year study of 2,500 California residential utility customers found that when prices rose by five times to reflect critical peak situations, usage dropped by 13%, which was more than enough to balance the load.
In Ontario, residential and small business customers can’t participate in the three-tiered program described above, but they can participate in a program called peaksaver, whereby a small device is installed near the water heater or central air conditioner. During times of high power usage, typically between 2 and 6 p.m. on summer weekdays, the device will turn off the water heater or will cycle the AC off for 15 minutes out of every 30. In early July 2008, the Ontario Power Authority had its first province-wide peaksaver demand response call, and the program was estimated to shave 40 MW off the peak power demand.
Demand curtailment programs are useful for reducing consumption during known peak periods, but they are not useful for dealing with volatility. Spinning reserve has always been used when a baseload generator trips offline, but the increasing use of wind power adds a new level of uncertainty to power production.
"The volatility of intermittent renewable resources will have a significant impact on thermal generating units as they will be called upon for quick starts, frequent shutdowns, and short interval restarts," said Dave Hawkins, lead renewables power engineer at the California Independent System Operator (CAISO). "There will also be a need for units that can operate at lower minimum values without major loss of efficiency. Greater flexibility will be a key strategy while maintaining reasonable heat rates and minimum environmental impact."
As the European Union moves toward its goal of achieving 20% renewables generation by 2020, and the U.S. by 2030 (some states have higher targets), the thermal generating units must become flexible enough to overcome wind’s variability.
"It may be politically correct to build solar and wind, but it is electrically necessary to include something that can be dispatched to produce power," said Mark Axford of Axford Turbine Consultants. "I expect to see more power purchase agreements fulfilled by a combination of renewable resources and gas turbine firming."
Stabilizing the Grid
There is one other often-overlooked issue that the deregulation split caused: Reactive power is needed for grid reliability and stability. Reactive power, also known as volt amperes reactive (VAr), is needed to support the active power as it travels along the transmission line. Without adequate reactive power, voltage sags, transmission lines overheat, and less of the generated electricity reaches the customer.
The problem is that although active power can travel long distances, reactive power cannot. Generators and electric motors can both be sources of reactive power. When generation and transmission were owned by the same entity, peaking generators could be located near the load and provide voltage support where needed. With the split, generation owners no longer have an incentive to produce VArs. Instead, transmission and distribution companies have to install expensive devices such as condenser banks or static VAr compensators to preserve power quality.
There is another solution, however, which meets the needs of both peaking power and grid stability. Peaking generators are typically placed near loads, while baseload plants may not be. When not needed for peaking, these turbines can bring the generators up to synchronous speed so they can connect to the grid and start supplying reactive power. By installing a clutch between the generator and the turbine, the turbine can then be shut down until it is needed to start supplying power again, but leaving the generator synchronized to the grid producing VArs . By having the generator already synchronized to the grid, when there is a demand for peaking power, it is much quicker to bring the turbine up to speed than if the turbine also had to accelerate the generator to full speed and synchronize it with the grid.
This approach is working in many parts of the world (see next section). What are missing in many areas are the markets that reward the use of peaking turbines in this fashion. By implementing markets for VAr production, the turbine owners receive additional revenue from their investments, while the transmission operators avoid having to buy additional equipment to provide voltage stability.
Utilities are taking a variety of approaches to economically providing reactive power from existing generation sources. In each of these cases, the company uses a synchronous self-shifting clutch from SSS Clutch Co. of New Haven, Del., or SSS Gears Ltd. of Sunbury-on-Thames, England.
With the clutch engaged, the turbine is used to bring the generator up to synchronous speed, at which point it connects to the grid. If power is needed, the turbine continues to drive the generator. During off-peak times, however, the turbine can be shut down. As the turbine slows below the speed of the generator, the clutch automatically disconnects the turbine and generator and the generator draws just enough power from the grid to keep it synchronized and generating reactive power. Later, when peaking power is needed, the turbine is again brought online, and when it reaches the speed of the generator, the clutch links the two, and the turbine again provides the power needed for generation. A more detailed description of the synchronous condenser operation is found in a recent article in POWER describing the retrofit of an existing combustion turbine.
In Western Australia, power generation has moved away from the population center of Perth, where three-quarters of the population lives, to the coal mines, aluminum plants, and a wind farm hundreds of kilometers from the city. This lowers generation costs, but because some of the local generation facilities were decommissioned, this led to stability problems in the Perth metropolitan area. Western Power Corp. modeled its system requirements and decided to use a peaking station near Perth, Verve Energy’s 576-MW Pinjar Gas Turbine Power Station, as a source of reactive power. GE Frame 6 units at the site were outfitted with clutches, allowing them to operate as synchronous condensers to stabilize the load coming from the more economical, but remote, generating stations, and allowing the transmission line to transport more real power (kW).
Lafayette Utilities System in Lafayette, La., purchases most of its power, but it bought four 50-MW GE LM6000 turbines to provide peaking and emergency power in case there was a problem with the grid. Two of the turbines have clutches, allowing the generators to provide reactive power to stabilize the grid and import more power when the units are not needed to produce electricity.
The same approach can be used with steam turbines. For half the year, BC Hydro operates six steam turbine generators near Vancouver to provide the city with power. The rest of the year the power comes from hydroelectric plants. When the steam turbines are shut down, the generators are used to stabilize the voltage coming from the hydro plants, improving the power factor to the point where BC Hydro could sell more power over interconnects to the U.S.
—Joe Zwers is a freelance writer based in Glendale, Calif., specializing in business and technology.