The largest grid operator in the US just released an innovative study that quantifies the reliability of possible future energy mixes. It found that portfolios with very large amounts of wind energy, dozens of times greater than the current mix, scored among the highest for reliability and resilience.
PJM, which operates the power system across all or part of 13 Great Lakes and Mid-Atlantic states as well as the District of Columbia, has previously found large amounts of renewable energy can be reliably integrated. For example, its 2014 study found no reliability problems from obtaining 30 percent of the region’s electricity from renewable energy. That finding has been confirmed by many other studies, as well as grid operating experience in the U.S. and around the world.
PJM’s latest study shows that even more renewable energy can be reliably integrated. In many of the scenarios PJM evaluated, wind and solar energy reliably provided the majority of electricity.
It should be noted that many readers may be initially confused by the way PJM’s results are presented.
Due to the way PJM’s study accounts for capacity, wind should be multiplied by around five and solar by around 1.2 to convert the PJM scenarios into the share of energy that would be provided by those resources in a particular generation mix (see footnote 68 of the Appendix). For example, when the report discusses solar providing 20 percent of “unforced capacity,” that actually corresponds to solar providing around a quarter of all electricity generated on the PJM system.
Wind was able to go much higher, reliably providing the vast majority of PJM’s electricity in some scenarios. In fact, PJM found no maximum on the amount of wind it could accommodate, as noted on page 34 in appendix. As shown on page 28 of the report, scenarios in which wind provided nearly all of the generation on PJM’s system (15-25 percent of unforced capacity) maintain reliability at levels comparable to those on today’s power system.
Wind makes the power system more resilient
Perhaps the most innovative aspect of PJM’s study was testing the resilience of future energy mixes to extreme events like the 2014 polar vortex. As we’ve explained previously, wind performed quite well during that event, while many conventional power plants failed due to the cold weather.
Only around one-third of the reliable energy portfolios PJM analyzed passed the resiliency test. Portfolios with a large amount of wind energy tended to be more resilient because, as PJM noted, wind energy possesses the unique benefit that “unavailability rates for wind are likely to decrease” under a polar vortex event.
Said another way, wind energy output tends to be above average when extreme weather causes output from nearly all other energy sources to fall below expectations. That type of negative correlation with the availability of other energy sources is the key to using portfolio diversity to make the power system more resilient.
Interestingly, PJM’s results show wind energy contributing to resilience in a way that is comparable to the contributions of coal and nuclear power plants. PJM found wind energy played a large role in almost all of the scenarios that maintained resilience while retiring many coal and nuclear power plants. As seen in the following chart showing the portfolios that passed the resiliency test, wind steps in in almost perfect lock-step when coal and nuclear are not available. For the two-thirds of scenarios that did not pass the resiliency test (those that appear in Figure 16 of the Appendix but not below), many had much lower amounts of wind energy.
PJM’s report discusses other extreme weather events that have affected a large share of generation, but does not quantitatively analyze resilience to them. Examples include droughts that have limited conventional power plants’ access to cooling water. Droughts and sustained high temperatures in various parts of the U.S. have forced fossil and nuclear plants to operate at reduced output or even go offline, while the recent drought in California greatly reduced the state’s hydroelectric output. Wind energy and solar photovoltaics continued to generate as expected during these events, as they require no water to operate.
Wind provides grid reliability services
The report notes that technological advances enable wind energy to provide many of the reliability services that conventional power plants provide today.
In fact, in many cases wind plants exceed the ability of conventional resources to provide these reliability services. As a result, in the high renewable scenario, PJM found wind could grow to provide 25.5 percent of PJM’s flexibility needs, up from 1.3 percent today. Partially as a result, PJM found a high renewable power system was about 50 percent more flexible than today’s power system (Appendix, page 29).
However, there appear to be some flaws in PJM‘s quantification of energy sources’ reliability services contributions. (See pages 17-20 of PJM’s report for the definitions of the following reliability services) As one example, because PJM’s calculation of reactive power contribution is based on historical data, it does not account for the increased reactive capability required of new wind and solar plants under a 2016 Federal Energy Regulatory Commission order.
PJM also incorrectly de-rated wind’s reactive power capability by 87 percent, due to the capacity accounting methods discussed above, while solar’s was reduced by 62 percent. While this derating of reactive capability makes sense for conventional generators, it is inappropriate for wind and solar plants that use power electronics to provide reactive power when they are operating at partial, or even zero, real power output. As a result, new wind plants are capable of providing fast and accurate voltage and reactive power control the vast majority of the time.
Similarly, PJM’s assumptions regarding the provision of frequency regulation service are based on current operating procedures, even though those will change as the resource mix changes. For example, Xcel Energy already uses wind plants to provide frequency regulation service, and has found wind provides frequency regulation that is faster and more accurate than that provided by conventional power plants. In the Appendix PJM acknowledges that at higher penetrations wind is likely to provide the same primary frequency response contribution as conventional power plants, but the same logic would apply to frequency regulation.
In fact, wind plants will likely be better than conventional power plants at providing primary frequency response and regulation. Wind plants typically respond to frequency regulation signals an order of magnitude faster than conventional power plants can, while the North American Electric Reliability Corporation has found that around 90 percent of conventional power plants fail to provide sustained primary frequency response. PJM’s report does note that nuclear plants do not provide frequency response.
Regardless, PJM’s report found that reactive power capability in the high renewable future was essentially unchanged from today’s levels, and frequency response capability only declined by a few percent. Given that the frequency response capability of the current power system greatly exceeds the need, such a small decline is not cause for concern.
PJM’s report did not analyze resources’ ability to ride through voltage and frequency disturbances on the grid. This is a critical reliability service, as the failure of conventional generators to ride through disturbances has been implicated in many recent grid reliability events. Had PJM examined this, it would have found that wind plants, again thanks to their power electronics, far exceed the capability of conventional power plants to remain online following a disturbance.
Availability of on-site fuel was the other category in which PJM projected a power system with large amounts of renewable and natural gas generation may differ from the current power system. However, it should be noted that again PJM assumed current operating practices, even though other grid operators have already required gas generators to have dual-fuel capability and store liquid fuel onsite. PJM’s report also does not account for the potential growth of innovative resources that can provide these services, like energy storage and demand response.
It should also be noted that, while PJM runs a sensitivity to examine resilience to a polar vortex event, for the reliability analysis that underlies the report’s main conclusions it does not quantify the risk that a common mode failure will take out a significant share of conventional generation. Rather, PJM uses the industry standard but invalid assumption that conventional power plant failures are random events, with no correlation between the failure of one conventional power plant and another. As real-world reliability events like the polar vortex have shown, that assumption overstates the reliability of conventional resources by ignoring the risk that many of them will be forced offline simultaneously by correlated, common mode failures. Wind and solar resources are held to a higher standard, as the impact of weather and other correlated events on their output profile is taken into account in this and other analysis.
Regardless, electricity markets will ensure that the power system continues to have enough capacity to operate reliably. The fact that prices in PJM’s capacity market are a fraction of the cost of building a new power plant indicates that there is no need for new capacity on its system. Existing power plants are almost always the cheapest form of capacity, and there is minimal impact on emissions from maintaining existing power plants that only operate during a small number of hours to maintain system reliability. If it becomes economic to replace those existing generators with new resources like demand response, energy storage, or gas generators, the market will provide the price signal to do so.
PJM’s study did examine the geographic diversity of wind and solar resources across PJM, which helps to increase the contribution of those resources to meeting capacity needs because if renewable energy is unavailable in one area it is typically available somewhere else on the power system. PJM’s study also captured the synergy between wind and solar output profiles, which makes their combined capacity contribution greater than the sum of its parts.
PJM’s report makes important contributions to understanding power system reliability and resilience, and confirms that very high levels of renewable energy can be reliably integrated. However, at times the report falls into the common misconception that conventional power plants provide essential reliability services while renewable resources do not. In reality, wind plants exceed the reliability contributions of conventional power plants in many cases.