The “Happy Hour Effect” – PV, Electric Vehicles and Learning Thermostats Walk into a Bar…!
Utilities are facing the confluence of three residential trends over which they have little or no control, but are tasked with meeting its consequences.
The effects of all three of these trends coincide at about the same time of day, potentially wreaking havoc on both supply and distribution hourly requirements in the late afternoon period from 5 to 8 pm, the traditional “Happy Hours”. The technologies involved are:
- Solar PV
- Electric Vehicles
- Smart thermostats
Each of these technologies is considered beneficial, as they include or combine energy efficiency, renewable energy, and reduction of fossil fuel use. However, as the adoption of these technologies increases, the unintended consequences of their combined effect on the grid may end up creating problems for those responsible for serving that load.
1. Solar PV – Residential homes are increasingly adopting solar PV, with solar panels on rooftops becoming more common as prices come down, more financing options are made available, and more initiatives are introduced, such as the new requirement in California for all new homes to include solar starting in 2020[i]. However, as the sun goes down, so does solar output, resulting in a shift from solar to electric grid demand between 5 – 8 pm, with variation by season and latitude.
The result of the increased adoption of solar PV is already being felt in areas of higher adoption, such as in Hawaii and California, leading to the term “duck curve”. As noted below for the California grid (CAISO) for actual data for the years leading up to the Spring of 2016 and projected through 2020, the effect of increasing adoption of Solar PV has resulted in reduction in grid load during the day when solar resources are highest, followed by the load ramping up as the sun goes down and solar PV’s contribution is reduced.[ii]
Figure 1 – CAISO Historical Spring Load Curve[iii]
The ramp up in load in 2017 of almost 7 GW between 5 – 8 pm is projected to grow even more extreme as solar roofs continue to increase. While this may not be entirely attributable to PV (note EV and Thermostat effects cited below), but given the relative adoption rates of Solar PV, it is the primary driver. Considering just the Residential sector, PV capacity has been growing at a rapid pace over the past few years and is expected to continue to increase in the next few years.
Figure 2 – CAISO Residential Solar PV Capacity by Year
In terms of timing, reduction in loads associated with this capacity has already reached over 3,500 MW[iv], or 15% of the peak evening load in the Spring, and almost 8% of the 2016 summer system peak of 46,232 MW[v], which occurred at 5 pm.
Given that sunset occurs about 7-8 pm in California in Spring, declining solar insolation would start about 5 pm and recede by about 8 pm, when the Spring peak now occurs. In the Summer, the peak occurs about 5 pm, when the solar insolation starts declining. So, the additional grid load caused by the declining solar PV resource would coincide with Spring peak and be worsened in the Summer
2. Electric Vehicles – The adoption of electric vehicles is already ramping up, with DNV’s Energy Transition Outlook projecting cost parity with standard cars by 2022[vi]. While much of the focus has been on charging stations, a significant amount of new electric load has already been added by consumers charging at home after work, with California being the leader in adoption, with 300,000 electric vehicles as of June 2017. Policy initiatives in California alone, including an executive order for 1.5 million zero emission vehicles by 2025[vii], and the retention of tax credits under the newly-passed tax bill in December 2017[viii].will mean a continued increase in the next few years.
The figures below show a typical July EV charging profile (left) and a TOU participant profile (right) for July[ix], demonstrating how incentives can shift the load profile for EV to avoid Happy Hour:
Figure 3 – Electric Vehicle Home Charging Profiles for Non-TOU (left) vs. TOU (right) Participants
3. Learning thermostats – The breakthrough in thermostat performance that Nest has pioneered with its innovative “learning” features means that Nest and its competitors[x] have made occupancy-based temperature setback automated and have helped overcome the complexity of programming that has traditionally limited consumer use and the associated success of thermostats in affecting electric grid demand. However, this means that while utilities can expect lower daytime demand with more use of automatic summer daytime setbacks, it also means ramping up of cooling demand as consumers arrive home and those setbacks expire. A pilot study performed for FPL Energy showed that for customers using a temperature setback of at least 4 degrees, the early evening peak was increased by 25%[xi].
As shown below, customers who programmed their thermostats with daytime setback (P/Blue) lowered their daytime loads but spiked their peaks starting hour 17 (5 pm) compared to both non-setback programmers (N/Red) and non-programmers (H/Green).
Figure 4 – FPL Thermostat Pilot Peak Day Duty Cycle Profile by Programming Segment
The confluence of these three trends will have a major impact on the future shape of the load that utilities must serve, principally for summer peaking utilities who must meet the post 5 pm peak. This will only be exacerbated by the declining solar production, additional load from EV’s being charged after the work day, and the effect of additional thermostat-driven cooling restoral as people come home from work when their learning thermostats react to the restored occupancy. The coincidence of these three will mean a spike in loads after 5 pm – the “Happy Hour” effect, normally a term considered a cause for celebration, but not for utilities in this case.
What can be done about this “Happy Hour Effect”? These three effects are basically independent phenomena, so a single initiative will not resolve it. But there are solutions already being applied, for a variety of reasons, that could be re-targeted to reducing the Happy Hour Effect.
For Solar, the impact on the grid of solar roofs is already being touted in articles and projections about the “duck curve” increasing the need for more flexible generation. Traditionally, base load power plants (coal, oil, nuclear) are designed to operate most efficiently at high capacity factors, meaning that they are not intended to cycle on and off. Shutting down and starting up a large thermal power plant is not efficient or practical! Add to that the need to ramp back up after 5 pm – often rapidly – compounds this problem. Natural gas plants are a bit more flexible, and that has become the dominant fuel for new plants anyway. Coal plant use has declined due to “lower-cost natural gas, renewable energy and regulations designed to protect public health”[xii] and oil plants have been replaced by gas as a long-term trend to decrease dependence on foreign oil, as well as environmental concerns. There are still about 100 nuclear power plants, which are particularly inflexible for cycling, most are over 40 years old and new plants are only now being considered. Wind plants do provide more hours of use during evenings and off-peak, providing options during post-Happy Hour, but are even more intermittent than solar, which only operate in daylight.
Storage is touted as one solution, both for local homes and in the distribution system, as well as a supplement to generation. Combining solar with storage is the best option, since customers with solar and storage can shift solar charging and discharging to optimize the load shape to offset duck curve effects if enough storage is available and placed strategically on the grid, if there is an economic reason for them to do so.
For EVs, some utilities (e.g. Dominion Power) are offering time-of-use rates and/or demand response to ensure that charging is shifted to later in the evening. EV’s typically include a deferred charging option so consumers can automatically start charging later (e.g. after midnight) rather than immediately after arriving home. These TOU rates can provide a strong incentive but special metering (such as automated metering infrastructure, or AMI) must be in place. Customers may not want their whole home to be on a TOU rate and separate metering adds cost, which may constrain enrollment. Other technology options could be used, such as a Wifi switch that would lock out charging during specific hours or a logger that could detect peak use and offer credits/penalties to ensure/reward compliance with off-peak charging.
For thermostats, increasing enrollment of Wifi thermostats in demand response can provide a reliable means of smoothing the “happy hour” spikes, especially given the two-way communications enabling real-time response options for utilities for enrolled customers. For example, depending on the grid load, setbacks during the low end of the duck curve could be canceled or even pre-cooling could be used to smooth the Happy Hour effect. Also, storage could be used (per home or on the feeders) to offset the increased Happy hour loads even for homes without solar.
Finally, AMI adoption means that utilities can monitor loads for both distribution and individual homes in real time, enabling true load management with any load (A/C, EV) or supply source (storage) to optimize the grid and smooth out the “Happy Hour” effect.
How Can DNV Help You Better Assess Solar, EV and Thermostat Impacts?
Based on DNV’s extensive experience in solar integration, storage, and distribution planning, as well as load analytics and program evaluation for electric vehicle and thermostat programs, we can help you assess the load effects of these technologies.
Once your information needs are identified, our teams of statisticians, engineers, market researchers, and load data modeling specialists have the experience and expertise to help you develop program potential, strategy and design for pilot programs and full-scale program evaluations for these and related technologies.
DNV is regular contributor to industry conferences, including recent papers presented at ACEEE, AEIC, BECC and IEPEC. Please contact us to discuss this blog or how we can help you with technology assessments for planning and evaluation efforts. Visit www.dnv.com to learn more.
[i] https://www.nytimes.com/2018/05/09/business/energy-environment/california-solar-power.html
[ii] https://www.caiso.com/Documents/FlexibleResourcesHelpRenewables_FastFacts.pdf
[iii] https://www.greentechmedia.com/articles/read/eia-charts-californias-real-and-growing-duck-curve#gs.wAAGNJk
[iv] https://www.californiadgstats.ca.gov/charts/nem
[v] https://www.caiso.com/Documents/CaliforniaISOPeakLoadHistory.pdf
[vi] https://eto.dnvgl.com/2017/renewables-power
[vii] https://electrek.co/2017/06/28/california-electric-vehicle-adoption-bill/
[viii] https://www.usatoday.com/story/money/cars/2017/12/19/electric-car-buyers-still-get-incentives-under-tax-bill/964543001/
[ix] Source: DNV Proprietary Client Project.
[x] https://thewirecutter.com/reviews/the-best-thermostat/
[xi] https://aceee.org/files/proceedings/2010/data/papers/1953.pdf
[xii] https://climatenexus.org/climate-issues/energy/whats-driving-the-decline-of-coal-in-the-united-states/
6/4/2018 9:00:00 AM