On December 17, 2019, the New York Independent System Operator (NYISO) Installed Capacity Working Group meeting included a presentation on the NYISO Climate Change on Resilience Study – Phase 1 by Eric Fox from Itron. The study included estimates of future load expected as a result of the Climate Leadership and Community Protection Act. This post primarily addresses that aspect of the study as it pertains to energy storage requirements.
The purpose of the analysis was to develop long-term energy, peak, and 8,760 hourly load forecasts that reflect the potential impact of climate change. Itron evaluated temperature trends using state climate impact studies and developed scenarios that reflect state policy goals with climate change impacts. The summary of results states:
- Analysis of weather trends across the state show statistically significant increase in average temperatures of 0.5 to 1.1 degree per decade
- State average 0.7 degrees per decade
- Temperatures on the coldest days are increasing faster than temperatures on the hottest days
- Trend likely to continue through the future and could be worse depending on long-term greenhouse gas path
- Warming trend will contribute to increase in summer peak demand and lower winter peak demands. Increase in cooling energy requirements will partially be offset by declining heating related requirements.
- State policy to address greenhouse gas emissions will have more impact on loads than the impact due to temperature trends. The end-use modeling approach provides a framework for translating energy policy into impacts on energy, hourly loads, seasonal peak demands, and changes in emissions of greenhouse gases.
The primary topic of this post is the impact of state policies but I have to address the weather trends in this report. As far as I can tell the primary source for the temperature trends presented is a study by two economists. While on the face of it measuring temperature trends should be relatively simple in fact it is not and I have no faith that a couple of economists are aware of any nuances to these numbers. To cut to the chase I do not find a claim that over my lifetime (approaching 70) that the average temperature of New York State has gone up nearly 5 degrees. I am sure that if I dug out representative climatic data that I would find a different story but I don’t have the time. Instead I refer readers to a recent post at a different site that shows that the temperature sensing network that was designed to be representative does not show such alarming trends.
More important is the report’s conclusion that temperatures on the coldest days are increasing faster than temperatures on the hottest days. This is a typical result from analyses of the data but is often ignored. The only reason I can imagine that it is not publicized more is that it can be argued that warmer winter days are more beneficial than not. That is to say it does not fit the alarmist narrative.
Future Load Analysis
Itron’s analysis of the load impacts of the CLCPA focused on the residential sector. They estimated the reduction in residential emissions needed to meet the CLCPA targets. Then they calculated the cumulative increase in electricity needed overall and on the average per residence. They found that in their policy case with accelerated energy efficiency gains and behind-the-meter PV adoption that the savings outweigh gains from electric vehicles and electrification. However, the CLCPA scenario has much higher electrification targets and state-wide annual energy use goes up about 50%. They predict that the summer peak will increase from 43,317 MW in their reference case to 57,109 MW in the CLCPA scenario. Importantly they expect that the electric load peak will shift from summer to winter. They predict that the winter peak will increase from 31,131 MW in their reference case to 71,859 MW in the CLCPA scenario.
My primary concern in this post is to estimate the amount of energy storage that will be needed to meet the CLCPA targets. My initial thought when I saw the projections available in this analysis was that I could repeat the evaluation of energy storage necessary that I did using the Citizen’s Budget Commission report with another projection. While doing this analysis I came to the conclusion that the emphasis of this post should not be on those results but rather the future load projection calculations.
My methodology for estimating energy storage requirements is based on evaluation of an example worst-case meteorological period. In the future CLCPA electric energy sector of New York, electricity will only be generated from hydro, nuclear, on-shore wind, off-shore wind, utility-scale solar, behind-the-meter (residential) solar, and other renewable. I estimate capacity for each renewable source category and apply those estimates to the energy requirements during the worst-cast period to calculate deficits and then estimate how much power will be needed to cover those deficits. All my analyses show that there are periods when the winds are light at night so significant energy storage will be required.
The NYISO Climate Change on Resilience Study – Phase 1 Forecast Comparison of Winter Peak figure lists four different projections for the winter peak electric capacity for 2020, 2030, 2040, and 2050. The most notable thing is the massive increase in capcity needed for electrification of the residential, commercial, industrial and transportation sectors in the CLCPA scenario. I calculated energy storage required for their 2040 and 2050 capacity projections for a light wind night time worst-case period on Jan 3-4 2018.
The Comparison of Energy Storage Required and Potential Price Necessary to Prevent Deficit on January 3-4 2018 table lists the results for the two projections. For the 2040 capacity projection of 56,071 MW $52.5 billion would be needed for the Li Ion storage batteries necessary to provide electricity during the renewable deficit periods. The increased capacity projected for 2050, 71,859 MW, reduces the battery storage cost to $43.1 billion. As long as batteries are this expensive over-building renewable capacity is a cheaper alternative so developing a way to optimize the costs of energy storage relative to the cost of additional renewable development would be a worthwhile investment.
My work on the CLCPA impacts on the electric system has emphasized energy storage requirements. My approach has three components: the potential renewable resources available, the generating resources capacity from each generation type, and how much load will be needed by the system. As shown below, the load projection for worst-case energy storage may differ from a worst-case analysis that uses the traditional peak energy load.
I believe an analysis of state-wide solar and wind resources is absolutely necessary in order to ultimately determine what could be available. NYISO had the Analysis Group do a forward-looking assessment of the fuel and energy security of the New York electric grid during winter operations that targeted potential reliability risks and impacts under severe winter conditions and adverse circumstances regarding system resources, physical disruptions, and fuel availability. The November 2019 final report was titled: Fuel and Energy Security In New York State: An Assessment of Winter Operational Risks for a Power System in Transition. I have previously analyzed the effect of winter peaks and used a period, 12/29/17 to 1/12/2018, chosen based on a cursory check of extreme conditions. The Analysis Group did an analysis over the last 25 years and this period was called out. They defined extreme weather events including the largest increase above average daily load over a long period as 14 days from 12/25/2017 to 1/8/2018 and more extreme shorter periods where they found in the last 25 years the fourth lowest 3-day cold snap was 1/4/2018 to 1/7/2018. Note that my attempt to estimate potential solar and wind resources came up with an overlapping time frame because energy storage requirements are driven by the renewable resources available. The question that needs to be answered is whether there are periods when the loads are unremarkable but the availability of renewables is so low that even greater energy storage resources are needed.
The second component needed is the generating resources capacity from each generation type. For my analysis of the Citizens Budget Commission results I calculated the MW capacity based on their power (GWh) estimates. Itron listed capacity in MW for 2040 and 2050. I arbitrarily split the relative amount for each renewable type and matched their totals. Eventually I hope that the State does a least-cost optimization study that balances how much of each renewable generating source and energy storage is needed.
The final component is how much load will be needed by the system. In this evaluation I arbitrarily chose to scale the observed load during the Jan. 3-4 2018 study period by the Citizen’s Budget Commission projected annual load in 2040 over the existing annual load. As far as I can tell, both the CBC report and the Itron analysis estimated future loads from increased electrification based on annual numbers. I believe that the increased load from heating electrification will increase the magnitude of the peak load more than the annual load will increase. In particular, the preferred option for retrofitting home heating is an air source heat pump. During extreme cold (less then zero degrees F) air source heat pumps don’t work efficiently so in order to keep warm people will have to rely on much less efficient radiant heating. I have shown that this will exacerbate the winter peak and further complicates the winter peak load projections. In order to incorporate this effect into the future winter peak projections detailed temperature and wind data should be used.
The NYISO Climate Change on Resilience Study – Phase 1 is a useful report and makes a couple of important points. I am not a big fan of studies that try to estimate the impact of climate change on future operations simply because weather variability is larger than climate variability. I believe trying to tease out a small climate effect is mostly bogus. The report does note that “state policy to address greenhouse gas emissions will have more impact on loads than the impact due to temperature trends”. They also come to the same conclusion that I have that the winter peak will be more important in the future.
There also are important limitations. I think that future capacity estimates based on annual energy usage have to be replaced by estimates based on shorter time periods and that the air source heat pump inefficiency issue has to be considered. The analysis underscores the importance of a comprehensive study of renewable resource availability relative to expected load when the CLCPA electrification requirements kick in because the peak energy storage requirement may not necessarily occur during the peak load period.
Two points need to be emphasized for the energy storage calculations using my methodology and the projected capacity from this report. Energy storage costs are extraordinariy high. In 2040 energy storage could cost $52.5 billion for 56,071 MW of capacity and in 2050 the cost would be $43.1 billion for , 71,859 MW of capacity. Incredibly according to the NREL report Life Prediction Model for Grid-Connected Li-ion Battery Energy Storage System on the life expectancy of lithium-ion energy storage systems: “Without active thermal management, 7 years lifetime is possible provided the battery is cycled within a restricted 47% DOD operating range. With active thermal management, 10 years lifetime is possible provided the battery is cycled within a restricted 54% operating range.” In other words the energy storage costs for the 2040 to 2050 time frame needs to sum the costs for a total of $95.6 billion! These numbers certainly support the notion that over-building renewable capacity will reduce energy storage costs and point to the need to optimize the tradeoff between renewable capacity and energy storage.