Beware the “Duck Curve”
For several years now, there have been complaints from utilities about the resulting electrical load shapes from owners with solar systems. This is the so called “duck curve” which has gotten a lot of attention from utility alarmists. In some sense, it has been a way of claiming that rooftop solar is a profound problem for utilities. We will show here that this issue can be (and will be) solved.
Figure 1 shows a classic illustration of the duck curve. Data here are those from the California Independent System Operator (CAISO) as logged on October 22, 2016. This was a day when the wind power output was low and steady throughout the day, but solar output varied strongly with its characteristic bell-shaped noon-centered production curve. Examining the resulting daily net loads, one sees the duck with low loads during the day when the sun is up. Note the duck curve’s steep rise from 17:00 to 18:00 as the sun sets, requiring some 5 gigawatts of generating capacity to come online within the hour.
More recent analysis by Scott Madden has shown that the duck curve is not only a product of increasing saturation of rooftop solar in California, but even more strongly a reflection of the saturation of utility-scale solar. This also explains how it has become much more pronounced in recent years.
Figure 2 shows the lowest load days in California in March from 2011 to 2016 so that it is possible to see that the belly of the duck is getting lower.
In examining Figure 2, note not only that the belly of the duck has been getting lower with increased utility-scale and rooftop solar, but also that since 2011, the nighttime loads have been falling in magnitude. That’s likely reflecting the increased saturation of efficient lighting in the California residential and commercial market.
The cure for the rapid ramp up seen in the duck curve from 4:00–7:00 pm is electrical storage in one form or another. With the very low cost of generating solar energy in the daytime hours seen in recent years, what is a problem now, becomes an opportunity for innovative electrical storage in the future. The promise is plentiful low-carbon electricity. This will be done by storing excess solar during the daytime hours and then applying stored energy to early evening loads.
One helpful weapon against the “duck curve” is obvious, if seldom recognized: more efficient lighting will help flatten the Duck curve given that inefficient lighting is partly responsible for the rapid rise in early evening electrical demand. Further, by substituting LED lighting sources, the amount of released interior heat to the home is dramatically reduced (often by several thousand Btu/hour), reducing air conditioning in Florida between 7:00–11:00 pm as well. Indeed, if one carefully examines the demand profiles between 7:00 pm and 5:00 am in the California data in the preceding Figure 2, it can be seen that the nighttime loads are dropping from 2012 to 2016 as the saturation of more efficient lighting grows in that state. Between electrical storage, LED lighting, and more efficient air conditioning in Florida, the infamous “duck curve” can be effectively flattened. How would efficiency and electrical storage work together in Florida homes to de-fowl the duck completely?
In Figure 3, we see how the electrical demand on a peak summer day in a typical Florida home are reduced by various measures that are sequentially installed as done in our Phase Deep Retrofit project (PDR, described at the end of this segment). These were simulated using an hour-by-hour simulation tuned to the characteristics of a typical existing Florida home. On the plot, we also superimpose the output of a typical simulated rooftop solar system. Thus, it is possible to see how the various efficiency measures cut demand on the peak summer day. The summary in the later table shows annual results and costs of the various options as experienced in the PDR project.
We see that the residential loads are made more flat and stable when a greater penetration of efficiency measures is achieved in a household. Indeed, simply through efficiency, the household electric demand is cut in half during the summer peak between 4:00 pm and 8:00 pm.
Beyond the profile-sculpting help of efficiency, the final coup de grâce to finish off the “duck curve” is with distributed electrical storage such as is currently being installed with solar by Sonnen in a new housing development in Arizona. The potential here is not only alleviation of the “duck curve,” but smoother electrical loads than without solar. However, efficiency is meaningful part of the package since electrical storage is expensive (often $700–1000/kWh or more in residential systems). Thus, with efficiency, it becomes possible to modestly size home electrical storage systems to make a big difference in serving the evening loads or even make the home 24-hour power-independent and virtually autonomous.
Figure 4 is an expansion on the presentation shown in Figure 3. Here we plot the “net demand” for the same Florida home with the 6.5 kW PV system before and after retrofit measures are installed. The red line shows the characteristic “duck curve” with a standard efficiency home with PV sending demand negative during the day with a total daily grid energy use of 27.3 kWh. The average demand during the peak period rises quickly between 4:00 and 8:00 pm (the “duck”) and is 2.62 kW. A second line shows the demand with all the efficiency measures and the PV system. Here, more power is sent to grid during the daytime hours and the total daily net demand is 0.3 kWh per day. The “duck curve” is much lower, averaging 0.72 kW during the peak period, but still quite noticeable.
A third plot line (bright green) shows the result when the excess daily electricity is stored in a battery system similar to the Tesla Powerwall 2 system (13.5 kWh of electrical storage). We assume that 10% of the excess PV energy going to the battery rather than to grid is lost due to inefficiency. With the efficient home with all the improvements, 12.6 kWh is stored during the middle of the day in the battery that is then used during the early evening hours to flatten out demand. Getting the electricity out of storage and to the loads also is associated with a further 10% efficiency loss1. Nevertheless, as shown in the bright green trace, the efficiency measures along with the battery are able to zero-out all building electrical loads between 9:00 am and 10:00 pm on the peak summer day and effectively eliminate the duck curve. The only remaining loads are 5.3 kWh of grid energy supplied between 10:00 pm and 8:00 am. It can also be seen in the plot that without the efficiency in place (red trace), there would be little electricity available for storage at midday to meet early evening loads. Thus, to eliminate the summer duck curve with PV and batteries, efficiency is needed unless a large PV array and a large system of batteries are available. Typically, roofs have finite space facing south that is not marred by obstructions and shading — the reason that we assumed a typical 6.5 kW system.
As this evaluation has focused on a single peak day in summer, what would be the impact to total energy as these improvements were installed? The table below summarizes the cumulative impact as simulated based on the PDR project. We include the cumulative measures shown above plus the solar electric PV system along with key energy and economic parameters.
The table also shows the impact of three other measures which are often desired: sealing building air leakage (air infiltration control), new advanced solar control windows, and exterior insulation over the outside of concrete block construction2. These measures tend to be more expensive, although potentially increasing interior comfort in a big way. For instance, double-glazed, solar control windows not only save energy, but result in a more quiet and thermally balanced interior.
In conclusion, we find that through two mechanisms — efficiency and electrical storage — the problem of the duct curve can be effectively smoothed away.
Finally, we summarize details regarding the Phased Deep Retrofit project indicating what could be accomplished in Florida with improvements to efficiency. The findings from that project were the basis for the values in the above table.
In our next segment, the last before we conclude this series, we will see how the efficiency improvement outlined here and electrical storage can also provide significant improvements to post-hurricane resiliency in Florida homes. We’ll also showcase the actual experience of nearly 2,000 Central Florida solar installations and how they came through Hurricane Irma and helped provide emergency power for homeowners.
The Phased Deep Retrofit Project
The Phased Deep Retrofit research project was accomplished in conjunction with FPL and U.S. Department of Energy monitored 60 Florida homes in great detail relative to end-uses (even down the televisions) as well as to collect information during the very simple pass through audit. This included dimensions, amount of insulation, type of windows and specific appliances and their characteristics in every category. We also monitored most of the circuit breakers in the electrical panel in order to obtain a detailed view of household energy use in Florida.
Total house electric power (and all the items below sub-metered)
- Heating and inside unit blower
- Cooling: compressor power
- Electric water heater
- Clothes dryer
- Range and ovens
- Pool pumps
- Central television and entertainment center
We also measured a number of unusual energy systems as part of the monitoring protocol; such as second refrigerators, home offices, wine cooler, and even home fountains.
Beyond these unusual power measurements, we took a detailed temperature measurement by the heating and cooling system thermostat. And since these measurements were so critical, we also always installed a redundant portable temperature and humidity logger so we could track how well the efficiency measures provided comfort.
For the shallow retrofit portion of the project we wanted to find easy-to-install measures that might save electricity and in a very cost-effective and easy to install fashion.
To save even more electricity later in the deeper retrofits, where more equipment and goodies would be installed, we collected detailed audit data on each home that might allow a customized more extensive approach to saving energy. The installed shallow measures:
- High efficiency LED and CFL lighting substituting for incandescent.
- Hot water tanks wraps and offering to change shower fixtures with high flows.3
- Cutting back on pool pump time if greater than six hours per day.
- Cleaning fouled refrigerator coils.
- Smart power strips that attempt to reduce phantom loads from devices plugged in but not doing anything.
The shallow retrofits were installed in 55 houses. To evaluate savings we used an analysis method where we examined each involved end use, looking for changes looked in consumption for a 30 day period before and after install. The unaltered segment of home was used to examine how weather-related loads changed without the intervention. All of the various measures were available at each site, but we found very different customer acceptance of some measures. For instance, although smart power strips offered some savings, we found many homeowners objecting to doing anything that would alter the mass of wiring that was typically behind the main home entertainment center or the similar tangles found in the home offices.
Based on evaluations, the shallow retrofits, really worked – saving 9% of measured average household electricity. That’s a big deal as the cost to do these measures was very low. In fact they are an easy project for do-it-yourselfers.
We evaluated each sub-segment of the retrofits. The measures which saved most consisted of changing out of all the lighting to LED and CFL. Also effective were hot water tank wraps and showerhead substitution – these measures seemed to really impact electrical loads and not just lighting. The lighting retrofit and tank wrap and showerhead were exceedingly effective at cutting lighting and water heating loads. Details here.
For instance we found that the very simple hot water measures saved 10% of hot water electricity. But changing out lighting not only directly reduced lighting electricity but also altered the waste heat profile in the home (and the need for cooling to cover it), but seemed to have also altered the thermostat temperature preferences of homeowners. We hypothesize that this may be due to altering the interior global mean radiant temperature. With cooler light sources, the interior environment feels cooler as well.
Those receiving the lighting retrofit seemed to prefer a 0.8 higher interior temperature weather winter or summer. They also saved an average measured 1.7 kWh per day in reduced lighting energy. Considering the interaction with thermostats observed, just changing out the lighting in a Florida home will typically reduce annual electricity use by 5%.
A series of deeper retrofits installed in ten of the PDR project homes that had already received the shallow retrofit. Those doing the deeper retrofits were already committed to replacing an aging or poorly functioning central heat pump, but would also have insulation levels brought up to codes, heat pump water heaters, nest thermostats and repair duct systems.
Using detailed data over a year-long period, pre and post we measured at 38% averaged measured reduction in household energy from more extensive retrofits, saving an average of 6,760 kWh. Mean cost of the measured for the deep segment was $14,700 with a median annual saving of $810 a year. While payback times were much longer with the extensive retrofit, it would be easily possible to run such an efficiency program where appliances and equipment are changed and improved at end of useful life where the incremental costs over standard efficiency air conditioners and appliances would be less than half those described here.
And not only did homes receiving the deeper retrofits save a lot, but some of the homeowners were inspired. For instance, Site #19 in the project, went from one of the most energy-guzzling households (an average of 67 kWh/Day) to one that achieved 38% efficiency savings (32 kWh/day) or roughly an annual household savings of $1400.
So enthused for the big monthly savings was the head of the household that he decided to install a 10 kW west-facing PV system to bring it close to net annual zero. The west-facing PV system in this case produced an average 31 kWh/day. The reduction in grid purchased kWh at this site from before the PDR project until afterwards was 82% or 60 kWh/day. This energy savings level has an annual value in Florida of about $2630, including local taxes. Both efficiency and PV elements showed reasonable economics.
From the PDR project, Site 19 is also a good example of how each home in the U.S. will, in some way, be unique. At this home, clothes dryer energy use was enormous, entering the project the household was measured to use 8.03 kWh/day (2,930 annual kWh) for the dryer. However by participating in a follow-on project to test and evaluate a new Whirlpool heat pump clothes dryer. This was able to cut their consumption for clothes drying by 2 kWh/day or 25% — saving $80 a year.
To be forthright, we also tried some things that didn’t work. Cleaning of refrigerator coils saved a bit, but not enough to be cost effective. But we did learn that if the refrigerator is using more than 3.5 kWh per day, it would be prudent to consider replacing it. It’s easy to measure them in an audit with a portable Kill-a-Watt meter and find out. We also found that cutting pool pump hours is a measure with no persistence. Within a year, pool maintenance people had brought the operation hours back up– a failing that fortunately we found was completely addressed by the fantastic results we saw from substituting variable speed pool pumps (typically an 80% savings).
In the final two years of the project we evaluated a plethora of advanced technologies:
- Very high-efficiency air conditioners and heat pumps
- Smart or connected thermostats such as Nest or Ecobee
- Heat pump water heaters (HPWH)
- Heat pump clothes dryers (HPCD)
- Fully variable speed pool pumps
- High efficiency refrigerators for those found to be using more than 3 kWh/day
- Supplemental ductless mini-split heat pumps to drop central AC load
- Adding more ceiling insulation or floor insulation
- Duct sealing after test
- High efficiency dishwashers
The technology evaluations were revealing, precisely because we knew so little about expected performance beforehand. Behavioral savings are notoriously difficult to evaluate.
Indeed within a second phase of the project we found that smart thermostats like the Nest and Ecobee are good bets for Florida homes. For instance a sub-sample of 22 homes only had the shallow retrofit + the connected thermostat (Nest). However, evaluation use a year of pre and post sub-metered data on the heating and cooling systems showed a savings of about 9% on cooling and more for heating — approximately 500 kWh per year. We also found significant reductions in kW for both summer (0.18 kW per site) and the winter (0.25 kW) peak condition.
A separate, even larger monitoring study done for FPL by Itron verified cooling energy savings of about 450 kWh per year and even much greater levels in attached homes which are more likely to be unoccupied for longer periods4. This technology is developing rapidly. Indeed, FPL is promoting smart home technology using smart thermostats (Ecobee4Smart Thermostat now has Amazon Alexa control built in for easy voice-activated convenience in heating and cooling control), likely because demand reductions from these devices are on a par with what can be obtained by their more expensive load control programs — a facet unmentioned in their promotions.
There are even greater potentials, however. My family and I enjoy the Alexa personal assistant at home and find particularly convenient the ability to control high-efficiency wireless LED lighting while providing easy dimming and even color control.
Heat pump water heaters may be the most important new appliance for Florida’s energy future. We measured them to be extremely effective in reducing measured hot water energy. They reduced electricity consumption for water heating by an average 68% or 5.3 kWh/day. With interior installations in utility rooms or innovative ducted designs we have found that these may reduce cooling by about 5% or about 280 kWh5.
Economics suggest that anytime in Florida that your water heater goes out, to replace it with a HPWH. Depending on household size and water heater location, savings will vary from $100—300 per year and average about $180.
The heat pump clothes dryers (HPCD) in our study saved an average 34% of clothes drying electricity, but with some interactions with the house cooling climate unvented models should be avoided unless located in a garage and likely the new vented HPCD (from LG) would be most appropriate for Florida conditions.
Supplemental Mini-Split Heat Pumps (centrally located and operated to supplement the main central system) saved 33% of space cooling and 59% of heating energy. If the installed models are 110 volt capable, they can easily be operated off a generator as they typically draw less than 500 watts in operation. 120 volt units up to SEER 17.2 are available from the Home Depot for as little as $7306. Thus, such a system could help cut one’s cooling load by a third if used over the year, at the same time providing a redundant cooling system that would provide air conditioning in the event of a prolonged power failure. With home battery storage and a rooftop PV system, such an arrangement could provide indefinite air conditioning even without grid power.
And if we did the deep retrofits over, we would include the supplemental mini-splits along with the very efficient conditioning system as one of the improvements. They are clearly a unique solution for Florida homes.
We learned that monitoring process itself can be fruitful. For instance, it appears likely from the project that if a refrigerator is found using more than 3.5 kWh/day, then it would be ripe for replacement with the most efficient models for the type and size needed.
And relative to the shallow retrofits, the connected thermostats save so profitably, that one should be installed in the process of doing the other shallow measures — such as a LED lighting makeover, a hot water tank wrap and potentially showerheads. With the better thermostatic control and lower internal heat release from the lighting change out, household cooling can be reduced by 15% or more as well as the direct savings to lighting and water heating electricity. The expenses for these modifications are modest while being highly cost effective.
Since the DOE research was government funded all of the data and findings were in the public domain. Thus, the fact that we found that we easily can cut the energy use of typical Florida houses by 9% or 4 kWh/day by spending $375 is (including all hardware, labor and costs) is a big deal. For a typical Florida home, that means a reduction in energy use of nearly 1,500 kWh per year with an annual value to those saving to the homeowner of about $180 a year, just over a two-year payback and a rate of return superior to conventional economic investments. Where can one find a stock or annuity or any other investment instrument that pays a 50% after tax rate of return?
Why would the Florida Public Service Commission not consider potentials like this and not decide that utilities should undertake a massive program to reduce Florida’s residential energy use? Would it not pass RIM test screening? After all a nearly 50% rate of return would seem intensively attractive if utilities were really interested in saving energy. In fact, the utilities could likely have it flunk by citing the fact that it could pay for itself in about two years. Households should be doing it without any incentive. But here’s the rub: most of Florida’s homeowners have no idea how cost effective it is to change out all their lighting to LED lamps and to install a smart thermostat. Moreover, we demonstrated in a conclusive fashion that deeper retrofits in Florida homes can save up to 38% of annual electricity with similar impacts to utility-coincident peak loads. Still, the utilities in Florida have seemed resistant to further efforts towards energy efficiency for their customers.
While, many in the efficiency community in the U.S were enthused by the measured statistical shown in the PDR project in Florida, FPL was unmoved. And when it came time to continue the research to evaluate other important questions, FPL was unresponsive to further proposals. And so key research questions remain:
- How well would vented heat pump clothes dryers work in Florida? (there are issues we discovered with fully condensing non-vented models).
- Would rooftop solar help control attic heat gain and help reduce cooling?
- Might combining rooftop PV with electric vehicles, electrical storage and time-of-day rates help utilities realize load shape smoothing goals while controlling revenue losses?
- Could detailed sub-metered data like that in PDR be used to verify end-use dis aggregation schemes using utility smart meter data?
We had hoped our long-time utility partner would be inspired by the great data emerging from PDR. However, instead they elected to remove their long-time Conservation R&D program manager, downsized of their DSM department and petitioned the PSC to gut their efficiency programs. Why? One possible explanation is that Florida IOU utilities saw energy conservation savings as something that was good green PR for the utilities when demand was growing for a time. But now, with falling household electric demand, these savings are seen as lost revenue — that may adversely affect the bottom line for profits and hence investors.
Let’s hope that someone reading this, be it at the Florida PSC or the utilities will have a change of heart and find out how to do what’s right for Floridians and still make that profitable for utilities and homeowners alike.
After all, given the likelihood of future challenges in Florida, we are going to need each other.
2 The improved windows and added wall insulation were both studied as case studies in a second segment of the PDR project. These both showed savings potential, although with large variability depending on interior temperature preference. Both were expensive measures with the window retrofit averaging $9,000 per home and the exterior wall insulation at near $20,000 as it included exterior insulation over masonry, then stucco and paint for finish. Although building air tightness is a known significant energy saving feature, generally tightening the structure should be associated with mechanical ventilation which was not done in the PDR project due to potential liability concerns.
3 Note that the substitute shower heads were not of the annoying flow restriction type, but a variety of fixtures chosen based on high consumer acceptance in Consumer Reports from which homeowners could choose. Still shower quality is most important to people and only about 40% of high flow fixtures here changed.
4 D. Parker, K. Sutherland, and D. Chasar, “Evaluation of the Space Heating and Cooling Savings of Smart Thermostats in a Hot Humid Climate using Long Term Data,” Proceedings of the 2016 ACEEE Summer Study on Energy Efficiency in Buildings, American Council for an Energy Efficient Economy, FSEC-RR-647-16, August 2016.
From the details of the FPL study of the same technology: “Smart Home Technologies to Help You Save,” Teslarati, Florida Power and Light, October 18, 2017.
5 C. Colon, E. Martin, D. Parker and K. Sutherland, “Performance of Ducted and Space-Coupled Heat Pump Water Heaters in a Cooling Dominated Climate,” Proceedings of the 2016 ACEEE Summer Study on Energy Efficiency in Buildings, American Council for an Energy Efficient Economy, FSEC-RR-644-16, August 2016.
Appreciate CleanTechnica’s originality? Consider becoming a CleanTechnica Member, Supporter, Technician, or Ambassador — or a patron on Patreon.