PJM presented updated estimates and analysis of coronavirus-related impacts to electricity demand at the April 14 Planning Committee meeting.
During weekdays, peak load – the highest point of electricity demand during a given day – registered approximately 8 to 10 percent (or about 7,500 MW) under what would be expected minus COVID-19 social distancing restrictions, said PJM’s Andrew Gledhill in a presentation to the Planning Committee.
The impact on total daily energy use has been slightly less, with an average weekday decrease of about 7 percent, or about 140 GWh.
The most dramatic reductions in weekday energy usage were observed on March 26 and 27, at over 10 percent (around 9,500 MW per day), in the wake of acute, state-mandated shutdowns of business activity, travel and more.
During weekends, the negative impact caused by virus-related shutdowns is “noticeably less” at about 2 to 4 percent Gledhill said, reflecting lighter load corresponding to traditional reductions in weekend business activity.
In general terms, the COVID-19 virus has smoothed out peaks compared to traditional hourly load patterns of late winter/early spring. Recent weekend peaks illustrate softer impacts, reflective of traditionally less intense weekend peak patterns.
“The load shape has been somewhat flatter than typical over the last weeks,” he said.
Gledhill said these figures will be variable as more data becomes available and hence should be viewed “as a rough guide.”
Fine-Tuning Models to Isolate COVID-19 Behavior
To isolate the impact of the COVID-19 virus apart from weather, PJM analyzed recent load patterns adjusted for actual temperature conditions dating back to early March, when notable impacts began, through March 23, when states and localities began implementing strict stay-at-home mandates, through April 11. By inserting actual temperatures retroactively, the models will therefore predict loads without significant error for weather.
PJM also attempted to account for the impact of behind-the-meter solar – such as solar-powered homes – that reduce peak demand on sunny days but must draw from the grid at other times. PJM assumed an average March/April behind-the-meter solar reduction of around 1,300 MW daily, but the historical range is actually between 0 MW and 2,400 MW, so the conditions on any given day may provide a margin of error of about 1 percent, Gledhill said.
For revisions to its load analysis and long-term forecast activity, PJM is closely monitoring inputs and updates from economic firms, namely Moody Analytics, as conditions evolve.
Evolving Load and Long-Term Forecasts
PJM will discuss long-term forecasts and evolving load patterns at the May 5 Load Analysis Subcommittee.
For that meeting, PJM invited transmission owners and members to share data and analyses from their experiences reflecting changing peak and energy usage patterns at the local and regional level. As the economy shifts, new patterns have emerged including higher residential load, Gledhill said, in addition to expected decreases in commercial and industrial load.
PJM does not have direct visibility into these changes in electricity sector consumption but depends on its member utilities and other load serving entities to describe those changes.
“That kind of information would be useful to PJM and other stakeholders to try to assess how best to move forward,” Gledhill said.