Published on August 31st, 2020 |
by Brad Rouse
August 31st, 2020 by Brad Rouse
I would love to have 100% clean electricity by 2035. Such a goal is admirable for its audacity and for its ability to get people motivated. For that I applaud it. But when we get down to it, will it be smart policy? Could it be that 90% would provide greater carbon savings than 100%?
The Biden campaign recently called for 100% clean electricity by 2035. But the issue we face is not clean electricity per se. The issue we do face is the imperative to reduce overall carbon emissions. To achieve 100% carbon emissions reduction, or 100% clean energy (as opposed to clean electricity), we need to do 3 things:
- Encourage energy efficiency and energy conservation wherever possible.
- Green the grid
- Electrify everything.
Currently, the grid produces 27% or so of US total carbon emissions. So when we say we are going to green the grid to 100% versus 90%, we are saying we are going to make an extra effort to convert the remaining 2.7% of US emissions from the electric sector by 2035, versus say 2050. Could it be better to hold this 2.7% on the table for a while, focusing on the remaining 73%, and get back to it later on?
The question comes down to what economists call “opportunity costs.” What is the cost of wringing the last 2.7% out of electricity emissions by 2035 versus doing it by 2050? Could we spend the same dollar cost on encouraging energy efficiency or on electrifying everything and achieve even greater carbon savings? Is there a better opportunity to spend the funds needed?
Carbon Pricing Will Help Us Find The Answer
I don’t know the answer to the above question, but I know how we can figure that out! You guessed it! Economy-wide carbon pricing is the answer. Put a fee on carbon that we project will drive us to 90% clean electricity by 2035, encourage conservation, and encourage electrifying everything. Then the extra benefit of going from a 90% clean grid to one that is 91% green can be weighed against the benefit of additional electrification or additional conservation. The cheapest form of carbon reduction will continue to win. By embedding the cost of carbon in our decisions throughout the economy, we don’t have to make decisions as central planners – the individual planners in enterprises throughout the economy can make those decisions for us. And if we as a democratic society are unhappy with the rate of greening, we can raise the carbon fee faster to speed things up!
The Pareto Principle, AKA The 80/20 Rule Applies
The folklore of many fields of endeavor contain references to “The Pareto Principle,” named for Vilfredo Pareto, an Italian economist and sociologist whose analysis I first studied in college. The principle is also known as the 80/20 rule, which means 20% of the work generally produces 80% of the results. Turning that around, it says the final 80% of the work only produces 20% of the result, after this first 80% of results have been achieved. That’s four times the effort to produce one quarter of the result.
Or, consider the aphorism commonly attributed to Voltaire: “don’t make the perfect the enemy of the good.”
Or, consider the mundane theory from investments and economics: the “law of diminishing returns,” which basically states that “after an optimal level is reached, additional effort produces declining result per unit of effort.”
It is common sense and conventional wisdom. That last 10% of going from 90% to 100% is going to be the hardest part.
How Does It Apply To The Energy Transition?
The Pareto Principle applies to all three main areas of the energy transition:
Energy Efficiency and Conservation: I know this from my own experience in the field working with low income families. You can choose multiple avenues to reduce energy consumption. Weatherstripping doors, adding LED lights, turning down the thermostat, using cold water to wash, etc., are examples of “low hanging fruit.” Other measures use more money to get a result, such as adding insulation or switching to new appliances.
Sometimes the benefit of one effort reduces the benefit of another. For example, consider adding insulation or switching to a more efficient appliance. If I add insulation first, then the incremental benefit of adding a more efficient heating system is reduced. If I add the more efficient heating system first, the incremental benefit of adding the insulation is reduced.
Electrifying Everything: As I discussed at length in an earlier article, some energy-consuming activities are easy to convert from fossil fuels to electricity (lighting), others are possible but depend on evolving economics (driving a car, heating a home), and others will be a lot harder (cement manufacturing, aviation). Market forces will lead us to choose the more cost-effective solutions first and leave the others for later. Carbon pricing will help those market forces make better decisions for us all. Obviously, the later ones will be more expensive (absent new discoveries). Those later ones are logical fields for subsidized research.
Greening the Grid: This is what I will focus on for the remainder of the article.
Solving Intermittency Is Hard. Solving Seasonal Intermittency Is Hardest.
In another earlier article I showed that with a modest fee on carbon it would be economical to replace all fossil fuels with renewables, even from highly efficient gas plants that have already been built, as long as the cost of storage or other solutions for intermittency are not considered. The problem is that as we continue to drive more and more fossil fuels off the grid, the cost of solving intermittency increases at an increasing rate.
One way to look at this issue is to consider battery storage economics as a function of the number of cycles of battery charge and discharge. The first battery a utility might buy will be used a lot – perhaps multiple times a day. But as you add more and more, the opportunities decline – so that at some point you add storage to replace that last kWh of fossil fuels left in the system. That need for the last kWh of fossil fuels in a year may mean that the last battery purchased is only cycled a few times in its life. This might be also called “seasonal storage” as opposed to daily or weekly storage. The economics of seasonal storage differ greatly from the economics of daily storage, and even with a high carbon price, seasonal storage might not be economical.
In my earlier article on intermittency I showed a chart (near the end) that related how the economics of storage would change under estimates of a declining cost of batteries from the National Renewable Energy Laboratory and a rising carbon fee from the Energy Innovation and Carbon Dividends Act. My conclusion was that under these circumstances, and assuming a 20-year life and 100 charge / discharge cycles per year, that such batteries would become insanely economical over the next 20-30 years.
Tightening just one of those assumptions leads to a different insight on the problem of seasonal intermittency. Shown below are the net economic benefits over time versus the number of cycles per year. The original assumption of 100 cycles is compared to results with 50, 25, 10, and 2 cycles per year. As you can see, the fewer times a battery goes through a charge / discharge cycle, the lower the net economic benefits. Even with the lower battery costs and high carbon fees, the cost of storage solutions that cycle as few as 2 or 10 times per year (dark blue and orange lines) does exceed the benefits.
You ask: “Why would I build a battery that only goes through a cycle twice a year? That’s preposterous!” And of course, it is preposterous with today’s battery prices. That’s the point. If anyone says batteries alone can solve the intermittency problem so that we can get to 100% renewable, they are saying that we will build batteries (or other forms of storage) that only get used a few times a year. And to account for variability in energy demand and in solar and wind output, they are saying that you are going to hold some batteries in reserve that you are only going to use once every 5 or 10 years to account for the most extreme conditions! That’s probably going to be a very high cost solution.
As we go from 90% to 100% clean electricity, we are making investments to solve intermittency that have higher and higher costs per unit of electricity converted from fossil to clean, and thus a declining benefit. We can show that declining benefit by focusing on 2035 to show declining economic benefits as a function of number of cycles, even with the assumed low battery cost ($25 – $112 per kWh) and high carbon price ($155 per ton) by that time.
The x-axis represents the number of cycles per year over 20 years used in the calculation. Decreasing the number of cycles means less and less use of the battery over a year, corresponding to solving deeper and deeper levels of intermittency. In all cases, as cycles decrease, the economic benefit of additional storage declines. Even at $25 per kWh for storage (1/3 of the cost projected by NREL for 2050), economic benefits with 10 or fewer cycles per year just are not there.
Recent Study Confirms The Cost Of The Last 10%
The University of California Berkeley Goldman School of Public Policy and the private think-tank GridLab undertook a detailed and comprehensive analysis: “2035 The Report – Plummeting Solar, Wind and Battery Costs Can Accelerate Our Clean Energy Future.” This report sheds some light on this discussion. Overall, the report is very good news indeed:
- 90% Clean energy can be achieved by 2035 at lower cost for electricity than current.
- The 90% clean grid can produce electricity reliably and dependably.
- Needs for solving the intermittency problem are manageable and low cost – modest improvements in transmission and new battery installations. Major new investments in a national grid are not needed.
- Many new jobs will be created.
- Public benefits include improved health and fewer deaths due to reduction in pollution.
- Sustained policy commitments are required to achieve this future.
- Going from 90% to 100% presents significant challenges, although a pathway was not considered in the study.
Reading through the details of the report are fascinating, even the Appendix! From the very beginning, the study attempts to find a fairly quick and cost-effective pathway to get most of the benefits of “greening the grid.” Early on, the report discusses why the 90% clean grid was chosen for 2035 versus the 100% goal, which the report suggested should be the plan for 2042 or 2050.
Here are some of my takeaways:
It’s easy to meet the final 10% of energy consumption in 2035 with existing high efficiency gas generators. No new fossil generation is needed, even with extensive electrification. Existing generators are needed, however, but they show gradually declining utilization over time.
By 2035, existing fossil generation is rarely used. Fossil generators that remain have an average capacity factor of only 10%. They are rarely used. The maximum gas generation needed at the peak hour is 360 GW out of a remaining capacity of 450 GW. 70 GW is used so rarely that it is only used 1% of the time. Think for a minute of the economics today or replacing that with some form of storage to replace that last 70 GW.
There are many technologies in the pipeline that might meet that final 70 GW. Remember, that 70 GW will require seasonal storage or massive overbuilding of wind and solar. But the reason the 90% scenario is so low cost, lower than today’s electric prices, is that we already have the capacity to meet the seasonal load in place today – it’s the fossil gas generating fleet.
Solving Intermittency Is Easy At 90%
The study did an exhaustive analysis of intermittency and resource adequacy for 2035 under the assumed conditions. They looked at 7 years of hourly weather conditions in over 100 separate US locations combined with the solar and wind conditions at each place and time for each hour. That is over 60,000 hourly calculations. The analysis was required to show that the electric system could meet the needs at each of those hours, and their final solution did.
In doing so, only two options for meeting the resource adequacy required were considered – battery storage and additional transmission. While earlier studies have indicated a need for long range transmission, the UCAL/Gridlab study indicated that no new long range transmission was absolutely required, but that substantial build-out of wind and solar at a regional level was required and that some new transmission was needed to move power from these new solar and wind installations all over the US to load centers. Again, no new long-distance transmission was necessary (but it could be helpful).
A substantial amount of storage was needed, however. Substantial until you look at it twice. 600 GWh / 150 GW was needed to ensure reliability. That is a lot more storage than is on the grid today. But that is peanuts when we look at just the storage which will require the electrification of the vehicle fleet. My Tesla has 75 kWh storage, so it would take the equivalent of 8 million of them to provide the grid storage we need by 2035. Given that there are over 250 million cars and trucks in the US today, 8 million is a drop in the bucket. If we are to electrify all of them, that is hugely more storage than what UCAL/Gridlab considered. Only a small portion of the fleet will need to be set up for vehicle-to-grid technology. Plus, by 2035 some of the EVs we build by then will already have ended their useful life. Remember, used EV batteries can still be used for grid-level storage. I would say that we can have vastly more than 150 GW by 2035, and even more by 2050. Relatively short-term storage should not be a problem.
I plan to write a companion article soon about options for seasonal storage. Seasonal storage is something we may prefer to go slow on until we are further down the road on electrification, the solar and wind build-out, and research advances in long duration storage and green hydrogen.
I’ve asserted that the cost of going from 90% to 100% zero carbon grid by 2035 versus doing it later may be very high. It may be better to focus our efforts on electrifying everything faster instead. Pareto pointed us in that direction, the data on storage cost per cycle suggest that conclusion, as does the UCAL/Gridlab study. Hopefully we can bring that all together in this last thought experiment.
What’s better for reducing carbon emissions – building batteries for the electric grid versus building batteries for electric cars? We assume the grid is 90% clean, so that the electric power for charging the cars uses electricity that is 90% zero carbon and 10% carbon from gas. Further, assume that we could build 1 GWh of batteries to further electrify the grid, or we could use the same batteries to build 13,333 new Tesla Model 3s, each with a 75 kWh battery pack. We assume no ability of “EV to grid” for these batteries. We can then compute the CO2 emissions per year of each option as a function of the grid charge/discharge cycles and the average miles driven per year of the cars. We assume 4 miles per kWh for electric and 30 miles per gallon for gas. Note: Grid charge/discharge cycles and miles driven both equate to how much the battery is used in a year. The following graph shows the results.
For 365 charge/discharge cycles, putting the battery to use on the grid is better, as the CO2 savings of the grid versus autos is higher. That’s a daily charge/discharge cycle. However, if we charge/discharge every other day (187) cycles, the results are mixed. Finally, if we discharge once a week (52 cycles) or once a month (12) or once a quarter (4), the results show ever increasing benefit for using our battery capacity to build more EVs versus investing in the grid.
This is exactly what you would expect. As we get deeper and deeper into grid decarbonization, we have fewer opportunities for additional battery use, and at some point it pays to place our decarbonization efforts elsewhere.
How will this work out in practice? There are many unknowns and far more variables at play. For sure though, no one can say today that rushing to 100% decarbonized grid is the best way to go. We shouldn’t make any sort of goal that constrains our options in the future. It would be better to put a price on carbon and then let the market work it out as all of the variables become better known. Pricing carbon allows us to decarbonize our economy and find the right mix along the way between electrifying everything and greening the grid.
Have a tip for CleanTechnica? Send us an email: [email protected]
Latest Cleantech Talk Episode