The coronavirus pandemic is having unprecedented impacts on society, and the power grid is no exception. Various reports are coming out almost every day about energy usage trends, whether from trade journals, grid operators or #energyTwitter. As most businesses remained closed and governors across the country have continued stay-at-home orders, an overarching trend is becoming clear: the pandemic is causing electricity demand to fall.

But grid-scale statistics don’t reveal the uneven distribution of impacts on different classes of customers. For example, commercial and industrial (C&I) load has dropped significantly, while residential load has increased. Fortunately, many Mission:data members have analyzed customer-specific trends, and we highlight their findings below.

Understanding the data is critical in several respects. By forecasting customers’ energy usage and bills, we can direct assistance to those who need it most amidst the crisis. We can also reconfigure energy efficiency (EE) programs to address high-usage customers who are most likely to be pinched by losses of income.

Let’s dig in.

The Overall Grid

Drops in demand were first seen in Italy, which initially led the world in the number of infections and deaths. Italy saw a loss of overall electricity demand around 18%-21% in mid-March. With a two-to-three week delay between European and American coronavirus cases, the U.S. has followed in Italy’s footsteps, but to a smaller extent. A University of Chicago analysis through April 7 showed that United States’ electricity usage dropped about 8%, with some areas dropping significantly more, such as New York City, the epicenter of the crisis, registering a 14% drop. Peak demand have also dropped in many wholesale markets.

The problem, of course, with grid-scale usage data is that it masks the individual impacts to residential, commercial and industrial users (C&I). Roughly speaking, C&I accounts for 50% of electricity use in America. If many businesses are shut, then we would expect a more precipitous drop in overall power demand. However, that overall drop appears to be offset by a significant increase in residential consumption. This is where analysis by Mission:data’s members come in.

Let’s Analyze The Data

OhmConnect was one of the first to release their analysis of residential customers. With thousands of users across California, OhmConnect found a 9.4% increase in electricity consumption for the week of March 19-25, the first week following California’s shelter-in-place (SIP) order. As a result of the increased consumption, OhmConnect predicted the coronavirus would increase Californians’ electricity bills by $1.2 billion.

Although this is generally bad news -- customers are facing higher utility bills at the same moment unemployment is skyrocketing -- there is some good news: by analyzing hourly power demand of its customers, OhmConnect found that the biggest increase is occuring during mid-day, when solar energy is abundant and prices are generally lower. With California generating approximately 13 gigawatts of renewable energy during spring days (composed mostly of solar), increased mid-day consumption is a great way to align demand with low wholesale prices, thereby “soaking up the sun.”

On the other hand, the bad news is that OhmConnect also found an increase in demand from 4pm-9pm as a result of California’s shelter-in-place (SIP) order. Not only is this bad for planners coping with the infamous Duck Curve, but it’s also bad for customers because of time-of-use rates, which in 2018 were gradually phased in, causing increased prices in the late afternoon and evening.

Time-of-Use Will Sting In Summer Months

Let’s consider what this means for a typical PG&E residential customer. The average electric bill -- before widespread TOU or the pandemic -- was $113.64/month. Assume the typical customer uses a modest 500 kWh/month, and let’s say the customer’s “baseline” is also 500 kWh/month. The increase in usage beyond 500 kWh is priced at PG&E’s TOU rates, given below. Note these have a “winter” and “summer” schedule; the changeover occurs June 1st.

Unfortunately, it seems likely that some travel and business restrictions will continue into June, increasing power usage (and air conditioning) at home.

OhmConnect calculated a 9.4% increase in power demand overall, but the portion that is on-peak vs. off-peak was not specified. Looking at the graph above, we took a few guesses at the percentage of on-peak increase due to the pandemic and calculated bill impacts:

Had the pandemic occurred three years ago, before millions of Californians were on TOU rates, its impact on utility bills would have been much less severe. Unfortunately, however, we are seeing a “perfect storm”: residents are experiencing a once-a-generation paradigm shift in electric rates at at the same time as a global pandemic. This is especially concerning as unemployment is exploding, putting massive strain on household budgets. And the worst may be yet to come: April and May are months with relatively mild weather; consumers benefit from lower “winter” prices in those months. As we enter June, however, on-peak air conditioning loads, particularly in inland California, are unavoidable for those families stuck at home.

Gridium Offers Commercial Buildings an “Energy Covid Toolkit”

In commercial buildings, energy demand is undoubtedly falling. But, despite office buildings being empty, their load isn’t falling to zero. As Gridium points out, the coronavirus creates some opportunities for EE. First, a building management system’s (BMS) normal schedules for HVAC equipment should be eliminated. And as the daily load curve turns flat, facility managers can finally investigate their “base load” -- the 24x7 power draw -- without contaminating their load profile data with occupant-driven variations. What is that long-forgotten electrical equipment? Now is a great time for professional building managers to investigate and make changes without those pesky occupants getting in the way. Retire the lava lamps in the lobby! Ditch the old copy machine that doesn’t have a “sleep” setting!

Detailed Grid-Scale Models: Amperon

Many energy efficiency (EE) professionals know that merely comparing energy use with a prior period leads to misleading interpretations of the causes of change. Large drops in power demand, such as Italy’s, are made plain with simple comparisons to a prior period. But what about a small reduction, such as 3%-4%, like ERCOT is experiencing? How do we know the reduction wasn’t due to weather? In order to calculate an accurate counterfactual, one must control for variables such as air temperature and other exogenous changes to the system.

Amperon, a leading provider of advanced electricity demand forecasting, has controlled for weather in its analysis of over 20 regions’ power usage across the world. Amperon calculated the percent reduction in power demand as a function of number of days after a lockdown. It appears that while some region’s energy use has stabilized 10-15 days after lockdown, many others are seeing increasing rates of decline 20+ days after lockdown, such as Spain and France:

“Weather normalization” sounds like a deceptively simple concept. Surely there’s an Excel tutorial for that, right? Nope -- as all EE professionals know, weather normalization is complicated. In order to compare energy usage to a prior period, there must be a single date that marks the change from “open” to “closed down.” Predictive models need a time period to delineate a “training” period from a “prediction” period. Unfortunately, shutdown dates within countries or regions are unclear: Many governments had conflicting lockdown orders that were not uniformly enforced on day one, and some cities issued their own stay-at-home orders prior to state action. So how did Amperon pick the correct date for each region? Demonstrating their creativity and analytical prowess, the company used OpenTable’s database of restaurant reservations to determine the start date of lockdowns. As Amperon explained, the collapse of restaurant reservations represents the de facto lockdown date, whereas government statements could differ from the “real” lockdown date by up to 10 days.

Adapting to Coronavirus: Three Reasons Energy Data Is Essential

If anyone is capable of both quickly analyzing trends and rapidly implementing changes to accommodate health and safety requirements, it’s the entrepreneurs and innovators of the advanced energy industry.

But conditions are changing fast. In-home retrofits are banned in state powerhouses of energy efficiency like California and New York due to physical distancing requirements. This has taken a huge toll on “boots-on-the-ground” implementers, for whom business has fallen off a cliff. Even when some of these activities are able to resume, supplies and customer pipelines will be interrupted, and laid-off employees need to be re-hired. All of this costs time and money.

That’s why the coronavirus is making access to energy data essential in three different respects:

Remote energy analysis is critical: Everyone knows that customer acquisition costs in the efficiency business are high -- particularly for residential. But getting easy access to energy information in digital form can both help reduce those costs and speed a deal’s time to close. When coupled with savvy electronic customer communications, a lot of the in-person functions associated with “business as usual” sales can be postponed, while keeping a customer engaged and maintaining a pipeline for when business resumes. New York startup Sealed is doing this. For its part, New York agency DPS recently met with EE implementers such as heat pump installers to discuss reliance on remote customer qualification tools (such as Sealed’s) to help keep the industry afloat during quarantine.

Measurement and verification (M&V) will become increasingly measurement-based: Calculating energy savings is complex, particularly when “non-routine events” (NRE) occur. An NRE is a variable that must be controlled for, such as a drop in a building’s occupancy rate or a change to an HVAC system. Is coronavirus a NRE? “It’s the mother of all non-routine events!” one company told us. While it’s possible to do custom calculations for every residential or commercial site, the reality is that one-off analysis is untenable at a large scale. And those “deemed” savings values for devices and appliances that were estimated based on “normal” operating conditions? Those went out the window with Covid-19. Instead, estimating energy savings in a pandemic means taking advantage of energy data. By comparing energy use at one location against a population’s energy use -- all of which was affected by the virus -- the effects of the pandemic can be mathematically eliminated. Of course, you’ll need access to the entire population’s individual energy usage data in order to use this cost-effective approach.

Energy management is a macroeconomic imperative: While it’s nice that many states have instituted moratoriums on utility shut-offs throughout the duration of the pandemic, moratoriums are not a waiver of utility bills. Ultimately, whenever this ends, families will face an eye-popping balance owed. Yes, that persistent ticking is the sound of a $26 billion time bomb of utility bill debt that will accrue in just four (4) months’ time, according to calculations by VoteSolar. With an estimated 20% of American families unable to pay their utility bills, relief is best accomplished through a combination of approaches: Direct bill assistance from government; write-downs of some unpaid balances, as in a “utility bill jubilee”; and, of course, energy efficiency, particularly for low-income households. An investment in efficiency will pay dividends -- particularly in the long term by insulating households against high bills during the next health or economic crisis.

Energy Usage During a Pandemic: Analyzing the Data

Find us: