Mid-Atlantic grid operator PJM has modeled a future with more wind and solar power, cheap natural gas, and retiring nuclear and coal power plants — integrated under projected peak conditions and in hypothetical heat waves and cold snaps — and found that this increasingly diverse generation mix doesn’t threaten grid reliability. At least for the near term.
The model sets an upper bound of about 20 percent on how much renewable energy — wind, namely — that the 13-state transmission grid operator can take, before reliability issues start to arise.
PJM’s analysis of reliability and fuel diversity (PDF), released Thursday, is among many across the country tackling the challenge of integrating intermittent renewable resources into the grid. In PJM’s case, however, it was triggered mainly by the growth of natural gas as a primary generation fuel, raising the question of whether that was reliable enough to replace retiring coal-fired and nuclear power plants.
PJM’s resource mix has changed rapidly over the past decade. In 2005, coal and nuclear resources generated 91 percent of its electricity. But over the past six years those figures have dropped to 33 percent for coal and 18 percent for nuclear, while natural gas has grown to represent 33 percent of the power mix.
Meanwhile, renewables including hydro made up 6 percent of PJM’s 2016 generation mix. Demand response, which had accounted for roughly the same amount of capacity as renewables as recently as 2015, has shrunk a bit in the past two years after PJM implemented new capacity performance rules.
Given these changes, PJM needs to find out “whether the resource attributes necessary to maintain system reliability will continue to be available in sufficient quantities,” given a variety of possible future scenarios, it wrote.
To get there, the grid operator modeled 11 different types of resources — coal, natural gas steam, natural gas combustion turbine, oil steam, oil combustion turbine, nuclear, solar, wind, hydro, battery storage, and demand response — against a set of reliability needs, ranging from frequency response and voltage control to contingency reserves and dispatchable load following.
It then ran them through a variety of forecasts, and grouped those into four broad portfolios. The first is PJM’s existing near-term portfolio out to 2021. The second accounts for “moderate” levels of coal and nuclear retirement largely being replaced with natural gas. And two categories project high coal and nuclear retirements, one with natural gas taking up most of the slack, and the other filling the gap with renewables.
For the next four years at least, PJM’s analysis shows little to no reliability risk. “The expected near-term resource portfolio is among the highest-performing portfolios and is well equipped to provide the generator reliability attributes,” the report noted.
To be sure, with less coal and nuclear caseload resources, some grid reliability metrics go down, such as “generator reliability attributes of frequency response, reactive capability and fuel assurance.” But others, such as flexibility and ramping attributes, increase, it noted — largely due to natural gas power plants’ relative flexibility compared to nuclear and coal.
PJM’s study did find reliability worsening in high wind and solar penetration scenarios, however. And growing beyond 20 percent renewables — a scenario most likely to occur through rapid growth of wind power — were deemed “infeasible,” because it could lead to “loss of load expectation” going beyond acceptable boundaries at night, when wind power is both a large share of load and subject to intermittent dips and surges in production.
“Nevertheless, PJM could maintain reliability with unprecedented levels of wind and solar resources, assuming a portfolio of other resources that provides a sufficient amount of reliability services,” the study noted.
But this study doesn’t get into how these new services could be provided: “This paper does not address any specific areas involved with the evolution of how prices are formed in the capacity, energy or ancillary service markets.”
PJM has a lot going on to manage its changing grid. The list of related efforts includes: “Capacity Performance, enhanced standards for inverter-based resources, centralized forecasts of wind, solar and distributed energy resources, business rules to support dispatchability of variable energy resources, 'Pay for Performance' regulation service, 15-minute interchange intervals, and the PJM Renewable Integration Study.”
It’s worth comparing PJM’s latest analysis to the aforementioned renewable energy integration study, conducted from 2011 to 2014. It found that, “with adequate transmission expansion (up to $13.7 billion) and additional regulation reserves (up to an additional 1,500 MW),” the grid “would not have any significant reliability issues operating with up to 30 percent of its energy (as distinct from capacity) provided by wind and solar generation.”
The difference here is between energy, or how much electricity is produced over time, compared to reliability, which measures a host of requirements like 24/7 availability to perform critical tasks. In those terms, “we did see a renewable limit at 20 percent of operational capacity. That's still significantly more than what we have now,” Mike Bryson, PJM’s vice president of operations, said in a Thursday conference call.
On the original question of natural gas, the study found it could meet the model’s needs up to a hypothetical 86 percent of the entire resource mix, giving it “no upper bound” in terms of reliability. It did warn, however, that “additional risks, such as gas deliverability during polar vortex-type conditions and uncertainties associated with economics and public policy, were not fully captured in this analysis” — a risk that PJM has already faced in the 2014 polar vortex.
Importantly, having a diversity of resources is as important as which resources they are, Bryson said in a prepared statement. “We found that the risk to the system wasn’t that resources couldn’t necessarily provide reliability attributes but that the potential concentration of a single fuel source or low-probability, high-impact events could cause significant impacts to the system.”
GTM Research analyst Elta Kolo noted PJM's “seemingly robust approach towards achieving long-term resource adequacy and operational reliability by continuously striving for fuel mix diversity.”
Its Generator Reliability Attribute Matrix, the formula by which it translates its 11 different resources and multiple reliability metrics into a measurement that can be used to compare them all against the other, “is one that can be adapted by other ISOs/RTOs to provide a high level comparative assessment across systems.”
As for distributed energy resources, they’re largely beyond PJM’s control, since they lie behind the meter at the distribution utility level. At the same time, the growth of rooftop solar, which shows up as a loss of load on sunny days to PJM’s perspective, could be a problem in years to come: “Depending on the level of penetration, DERs could require increased de-commitment/re-dispatch of centrally dispatched resources or off-system schedules to meet the balancing obligations,” the report noted.
PJM is working on DER integration rules and solar forecasting “to lessen the impact on existing resources and improve market efficiency.” These are early-stage DER integration efforts, compared to California grid operator CAISO, with its distributed energy resource provider product set to allow virtual aggregations of DERs to bid into its energy markets as early as this year.
At the same time, PJM hasn’t yet come to face the real-world challenge of integrating double-digit percentages of renewables into its grid, as California has. Last month CAISO predicted that it could face the need to curtail up to 8 gigawatts of generation this spring, if must-take hydropower and midday solar power exceeds demand for energy.