PHOTOVOLTAICS AS A LONG-TERM SOLUTION TO POWER OUTAGES
CASE STUDY: THE GREAT 1996 WSCC POWER OUTAGE
copyright American Solar Energy Society
Richard Perez and Robert Seals ASRC, The University at Albany
Howard Wenger and Tom Hoff, Pacific Energy Group
Christy Herig, NREL
ABSTRACT
The Western System Coordinating Council (WSCC) experienced two major power outages during the summer of 1996. Both outages occurred during extreme heat waves and virtually cloud-free conditions. Focusing on the August 10 outage, we present evidence that dispersed photovoltaic power generation has the potential to mitigate the effects of massive power outages.
The authors have previously shown that PV output is highly correlated with heat-wave-driven loads (A/C driven loads). The following analysis indicates that the circumstances of the August 10 event are no exception. PV availability was at a maximum and the high loads on the system were directly attributed to the heat wave.
The paper presents PV deployment strategies that would serve to mitigate the effects of the outage, initially from the perspective of direct consumer relief. However, as dispersed PV generation saturates the electrical system over the long term, the high load demand match relief could prevent the cascading trips of large generating plants, thereby providing system relief
1. THE POWER OUTAGE
1.1Background
On Saturday, August 10, 1996, at 3:48 PM Local Pacific Time, a massive cascading power outage interrupted electric service to over 7 million customers throughout the Western Systems Coordinating Council (WSCC) service area [1]. These customers lost power for as long as 6 hours, resulting in large financial losses, customer inconvenience, and an undetermined number of health impacts. Fig. 1 shows the extent of the WSCC service area and the region of events originating the disturbance. The WSCC coordinates operation of 91 utility systems with the goal of providing a reliable source of electricity to about 60 million people in 14 states, two Canadian provinces, and part of Mexico.
1.2Causes
The first domino fell in Oregon. A combination of high ambient temperatures and power demand overheated three 500,000-volt Bonneville Power Administration (BPA) transmission lines, causing the lines to sag into tall trees and automatically shut down. This occurred during the hour or so before the beginning of the massive outage. Power had to be shifted to other transmission lines to compensate for the loss of the 500,000-volt lines. This caused other BPA lines
Figure 1. Map of WSCC service area
to sag into trees and trip off-line, and caused the BPA transmission and generation system to become unstable and at 3:48 p.m. the intertie between California and Oregon
was separated. As a consequence, there was a tremendous instability in voltage and frequency that resulted in four separate "electrical islands" and resulted in a massive outage affecting millions of customers across the entire WSCC.
High temperatures that drive high demand from air conditioning systems have always been the bane of utilities since they cause peak demands that strain the electric network. The WSCC outage was no exception. Fig. 2 shows how temperatures crested in the Pacific Northwest on the day of the outage, and a general trend that temperatures (and air conditioning demand) were sharply rising throughout the entire WSCC.
Figure 2: Average daily temperature departure from monthly mean in the WSCC service territory
The bottom line is that if demand could have been curtailed, or if local distributed generation were deployed, the chain of events would probably not have led to the massive outage.
1.3 Future Power Outages
Many experts believe that power outages will increase in frequency in the future, for a number of reasons. First, the slow phasing in of deregulation has caused a "prolonged paralysis in capital spending by utilities and independent power producers" which has resulted in the deterioration of the electric network [2]. "The western electrical grid is sensitive to major equipment outages and improvements are not being made to prevent cascading system collapses" [2]. Second, the demand for electricity continues to grow 2-3 percent annually, which means margins are tightening and the stress on the system is growing. And third, it will become more difficult to keep the regional transmission system in balance because larger power transfers will occur, and utilities and power generators will become even less inclined to cooperate because of competition. "Greater volumes and more complex patterns of exchange are likely as the industry moves toward competition", observed then Energy Secretary Hazel O'Leary in a letter to President Clinton on August 2. "The institutions charged with maintaining the reliability…will have to evolve at a pace commensurate with the changes in the industry [3]".
In addition, strong heat waves are likely to continue as a byproduct of observed greenhouse effect tendencies-- the ten hottest years of the century occurred since 1980 [4]. This will increase the frequency of high demand and intense power transfer conditions within the WSCC area as well as in other US reliability councils.
2. PV AS A LONG TERM SOLUTION
2.1 PV offers a Solid Match to Heat-Wave Driven Loads
In previous studies, we have shown that the summer heat waves that tend to increase peak loads and promote long distance power transfers are also characterized by high photovoltaic load carrying capabilities (ELCC) [5,6] -- the ELCC is a measure of a power plant’s capacity credit. In fact, we observe that summer PV’s ELCC typically exceeds 60% over much of the US, and 70% within much of the WSCC area [7]; the map in Fig. 3, showing the US distribution of summer PV ELCC, is a preview of our current research on the seasonal variations of PV capacity.
Figure 3 US distribution of summer PV ELCC
2.2 Case Study: PV Availability on the Day of the Power Outage
The availability of PV during the August 10 heat wave, and during days that preceded it, fully confirms our contention that PV offers a solid match to summer peak loads.
Figure 4: GOES-9 view of the US on 8/9 at 1 PM Local Pacific Time
Figure 5: GOES-9 view of the US on 8/10 at 11 AM Local Pacific Time
Figure 6: GOES-9 view of the US on 8/10 at 3 PM Local Pacific Time
The GOES 9 satellite pictures in Fig. 4-6 show that the WSCC region was characterized by very clear conditions extending from the Mexican border, well into British Columbia, at the time of the outage and during the hours and days that preceded it. On the afternoon of the outage, we only note a slight degree of cloudiness over mountain ranges and in the southeastern portion of the WSCC area which is typically affected by summer monsoon from the gulf of Mexico. Conditions were extremely clear in northern California and Oregon, within the zone of outage triggering events.
PV output integrated over the entire WSCC region was simulated from hourly GOES 8 and GOES 9 images [8] for August 9 and 10. The PV output profiles are shown in Fig. 7 for fixed and tracking 1 kW nominal summer condition-rated PV arrays. We observe that PV generation on both days would have been close to its theoretical maximum, still exceeding 80% of nominal capacity at the onset of the outage. Note that the slight afternoon output degradation is caused by temperature effect, as we have assumed the PV arrays to be of crystalline silicon type. Note also that PV output in northern California/Oregon, where the outage was triggered, would have been even higher than shown in Fig. 7, because (a) the region lies at the west of the WSCC, hence receives its peak solar energy later than the region as a whole, and (b) conditions were clearest in this region.
Hence, a sizable distributed PV resource in the western US would have fed large amounts of power in the concerned grids, well in phase with local peak demands, in both time and place where it was needed most. See, for instance, a typical Saturday demand profile for PG&E in Fig. 8, and the availability of sub-transmission capacity caused by the compounded effects of high demand and high temperature [9] in Fig. 9.
As a consequence, power transfers could have been reduced, greatly diminishing the risk of the catastrophic domino effect that led to the outage.
Figure 7: Satellite-simulated PV output on August 9 and 10 for fixed a two-axis tracking 1 kW PV arrays.
Figure 8: Typical summer Saturday load profile for Pacific gas and Electric
2.3 Dispersed PV Solutions
There are essentially three possible ways that PV can be deployed to increase the reliability of electricity supply. Fig. 10 illustrates these PV solutions. We call these solutions (A) Proactive; (B) Reactive; and (C) Disconnect. The proactive strategy calls for the deployment of PV in vulnerable capacity constrained locations on the utility's transmission and distribution system, either via larger ground-mounted systems or residential and commercial customer rooftop/building integrated systems. This solution requires the utility to be proactive by choosing the right locations for these PV systems in order to prevent outages before they occur. The second strategy (B) is for customers to have their own PV systems with storage so they are able to react to the outage after it occurs with a reliable source of back-up power. Finally, the third solution (C) is for consumers to simply disconnect from the utility grid altogether and have a PV power supply that meets their daily electricity needs. These customers are indifferent to outages since they live independent of the electric grid and its power lines.
The reality of the situation is that it would require tens, if not hundreds, of MW of ground-mounted PV systems in the right locations (solution A) to prevent a massive power outage like the great WSCC outage. This magnitude of PV penetration by proactive utilities will probably not occur in the next 10 years for economic reasons. Thousands of individual customer-sited PV systems can in aggregate, however, have the same impact of a few very large ground-mounted PV systems (also solution A). This points to new opportunities for utilities to engage in energy service businesses that offer distributed PV systems to their customers in key locations that provide a new source of revenue and simultaneously benefit the utility's grid.
Individual grid-connected systems are competitive today in customer-owned niche markets such as in Hawaii (solution B). Electric utility restructuring, in states like California, will provide for added customer incentives (such as rebates and low-cost financing) that will make individual customer-sited PV systems competitive with grid power and even more widely available. It is likely that these types of systems will come with a UPS battery back-up option to improve reliability. Consumer-based PV systems with UPS back-up (solution B) appear to be the most promising near-term PV solution for reliability enhancement.
Figure 10: Contrasting three dispersed PV deployment solutions
Off-grid systems are generally cost-effective in remote locations for those willing to adapt to commensurate lifestyle changes (solution C). This solution will do little for the grid since these customers' loads do not match the customers that create high loads during heat waves, not to mention that a relatively small number of customers are willing to live in remote areas in the first place. One can envision, however, that once PV and storage become very economical and establish a track record of high reliability, many customers will choose to simply disconnect from the grid and have their own power system (also solution C).
Other evolving distributed generation technologies, such as clean natural gas-fired engines and fuel cells, could make this a reality sooner than later.
CONCLUSIONS
The evidence presents a clear set of conditions during the outage that indicate that dispersed PV generation deployment could lessen the effects of the outage. The WSCC high system demand during the outage is attributed to hot weather loads, and PV capacity was at a maximum. Furthermore, the large emergency power purchases required to supplement the large generating unit outages, which extended for several days after the initial event could have been lessened.
The three dispersed PV strategies -- utility pro-active, customer reactive and lifestyle change -- all contribute to system load relief. The paper encourages a long-term PV deployment strategy. The level of system relief depends on the saturation of dispersed PV generation. Initial relief will be at the level of the customer and local electrical system, but over the long term, dispersed PV generation could prevent the cascading outage affecting a large area of the electrical system.
The media response related the outage to electrical system reliability issues arising from the restructuring -- and eventual deregulation -- of the electric utility industry. The policies required to address the reliability issues also provide the opportunity to include PV deployment strategy. Many policies are emerging through recent legislation to ensure continued deployment and market development of renewable technologies for the public good. We conclude that beyond the public benefits of environmental management and technical leadership, policies that promote PV deployment also address the reliability issues of deregulation.
ACKOWLEDGEMENTS
This project was undertaken as a byproduct of NREL contracts No. XR-1-11168-1 and XAX-6-16817-01.
REFERENCE