September 22nd, 2020 by Brad Rouse
For those deeply interested in our future energy system and how it will evolve, I highly recommend two in-depth papers published in the last year.
The 2035 Report, a product of a team led by UCAL Berkeley and GridLab, explores a 2035 US energy system that is 90% carbon free, results in lower electric rates than today, and is highly reliable. Steve Hanley reviewed the report in more detail in CleanTechnica, which is shorter reading than the full report, available here.
The second report, “Impacts of Green New Deal Energy Plans on Grid Stability, Costs, Jobs, Health, and Climate in 143 Countries” published in One Earth, provides a conservative roadmap for supply of all the energy needs of the US (and 142 other countries) that is 99.7% renewable energy by 2050. The 2050 study was written by a team at Stanford under the leadership of Mark Jacobson. Michael Barnard reviewed this study in two great CleanTechnica articles, here and here.
Instead of just reporting on these very exhaustive, well documented, and complex reports, I will show a similar vision step-by-step using some of the same assumptions and publicly available data using “back of the envelope calculations” (spreadsheets). My purpose is to: (1) demonstrate to myself the reasonableness of their more sophisticated approaches and (2) show the basic assumptions and logical processes in a way that a lay reader can follow.
In other words, to see if less of a “black box modeling” approach can show much the same result. In other words, a thought experiment. Quick summary – it shows much the same result and yields a few insights (at least to me).
This thought experiment begins with my most recent CleanTechnica article “100% by 2035? 90% Might be Better,” continues with this, and will be completed in a future article on what it will take to reach 100% carbon free energy by 2050.
Getting To A 90% Grid By 2035
Let’s start with assumptions for converting the grid to 90% renewable.
Assumption 1. The grid in 2035 and 2050 (before considering electrification of other energy uses) will use the same amount of electricity as it uses today. There has been little or no growth between 2005 when electric production was 4,055 terawatt-hours (TWh, or a thousand gigawatt-hours) and 2019 when it was 4,153 TWh. 2020 will undoubtedly be well under the figures of 2019 or 2005. Ongoing efficiency improvements are nullifying economic growth and electrification. Electricity demand before additional electrification would probably decline with carbon pricing and slowing population growth. This assumption is roughly consistent with the 2035 Report and Jacobson’s, although neither of them provide a “no electrification” estimate.
Assumption 2. Modest investments in storage, modest build-out of grid infrastructure, and limited use of fossil fuel assets can pretty much take care of intermittency issues that are not long duration or seasonal with a 90% grid. This was a conclusion of the 2035 Report, which showed only 600 gigawatt-hours (GWh) of storage (equal to 8 million Tesla Model 3 batteries) would be needed for 90% renewables in 2035. In my calculations I assume any intermittency that within a calendar month can be overcome by these limited storage investments.
Assumption 3. Replacement of existing fossils with 50% wind and 50% solar to get us to a 90% clean grid. This mix is consistent with Jacobson, whereas the 2035 Report evaluates multiple scenarios ranging from a roughly even mix of wind and solar to a more wind-dominated system.
Assumption 4. Existing hydro plants, other existing renewables, and existing nuclear plants stay the same in all scenarios. This is consistent with the 2035 Report, whereas Jacobson eliminates nuclear by 2050. In the next article I’ll look at the no nuclear option for 2050.
Assumption 5. Ignore regional imbalances and look at the US as a whole. I realize that this last assumption is not realistic, so just take that as a given. Both the 2035 Report and Jacobson’s 2050 study look at regional differences in detail.
With those assumptions, electricity production today can be compared to what the 90% grid in 2035 would look like, using monthly and annual data for electricity production.
We Will Need Lots Of Wind & Solar
Electricity production for the latest year available, 2019, was generated by fossil fuels at 62.1%, nuclear 19.5%, wind and solar 9.8%, and other renewables 8.6%. What if we build out enough wind and renewables to meet 90% of energy needs by nuclear and renewables, as in the 2035 Report?
Fossil would be 10%, nuclear 19.5%, and wind and solar 60.50%. Here’s what annual energy production would look like for the existing grid before and after the conversion to 90% carbon free:
Note: Units are in terawatt-hours (TWh) which is 1,000 gigawatt hours (GWh). Just for reference, this is about 500 million times the amount of electricity (8,000 kWh) typically used over a year in my home.
Here is the result for the 90% grid(2035), broken out by month. Notice the system has greatest demand in summer months with a lesser winter season peak.
Study this graph for a minute, because understanding it is a key to what comes next. Fossil fuel is zero in April and May. In these months, we have curtailments, or solar and wind energy. We have solar and wind that we can’t use. This is what is meant by “overbuilding wind and solar.” Remember, under our assumptions, energy is not stored in one month and used in another. If we had enough storage to do that, energy produced in March and April could be used in later months and we wouldn’t need to build quite as much wind and solar.
Notice that with the 90% clean grid, the greatest need for fossil fuels is in July and August when wind energy is lowest. This confirms results from the 2035 study.
Remember also, solar and wind that is not used for electric production and which exceeds battery storage capacity, is still AVAILABLE FOR SOMETHING. And it’s available at zero marginal cost.
Can We Build This Much Wind & Solar?
With this as a starting point, our thought experiment can proceed quickly to understanding how much wind and solar we need to build. In the first graph, wind and solar in 2019 is 407 TWh, whereas wind and solar in 2035 (assuming we reach 90% that year) is 2,640 TWh, or about 6.5 times greater. Combined wind and solar capacity would grow 960 GW. By way of comparison, the 2035 Report shows capacity growth of 1300 GW, but that report included a significant amount of electrification.
Either way, this is a huge transformation! But how does it compare to the progress of recent years?
Growing the installed production of wind and solar from 407 TWh to 2,640 TWh requires a compound annual growth rate of 12.2% a year. From 2014 to 2019 the rate of growth in production was 14% a year and it shows no sign of slowing. Long term growth of 12.2% seems in-line from a growth rate perspective.
But relying on compound rates of growth can be tricky. This level of growth also means that we need to add over 130 TWh of new production of wind and solar on average, so every year we need to add the equivalent of about 30% of 2019 production, and every year on average we will need to add double our best year so far, which I am estimating the be 67 TWh in 2020. Still, this seems doable to me, and it likely doubles the jobs in wind and solar. Probably we will need to increase wind and solar at an accelerating rate for a few years and then level off at a steady rate for the ongoing build-out. And if the cost keeps coming down, doubling the output of wind and solar is likely to be less (possibly much less) than doubling what we are paying now.
But on top of this we must begin the transition to an all (or mostly) electric economy!
Electrifying Many Things By 2035
We can’t just green the grid by 2035 and only then start seriously to electrify everything. We’ve got to start now to move end uses like electric cars and heating systems to electric so that when the grid is green, it will all be green. The problem is that the electric grid is only 27% of carbon emissions and only about 18% percent of total energy consumption.
We’ll continue our thought experiment by exploring non-electric energy consumption, which is basically fossil fuels like natural gas, oil, and coal consumed directly in homes, businesses, factories, and vehicles.
Non-electric energy consumption is measured in BTUs (British Thermal Units). The EIA tracks non-electric energy consumption by major sector – residential, commercial, transportation, and industrial.
What if we were to convert all of these non-electric energy uses to electric and not assume any efficiency gains are possible when going from fossil fuels to electric? Now this is totally unrealistic, because there are MAJOR efficiency gains when going from fossil fuel to electric (unless that electric is mostly produced by fossil fuels). But this is a thought experiment, after all.
Energy demands on such an electric system would far outstrip the current electric system, with a total of 22,666 TWh for 2019 versus the 2019 electric system demand of 4,153 TWh. Without assuming any efficiency gains in electrification, we are really going to need a lot of wind and solar to be even at a 90% fossil free grid – not 6.5 times the 2019 amount, but 40 or 50 times!
This scenario before efficiency gains are factored in is the first column in the graph below.
I show this scenario because sometimes people do these conversions, do not make realistic conversions for efficiency, and then try to convince people that this entire enterprise is hopeless.
Realistic Assumptions For Electrification
Here are our assumptions for conversion of non-electric energy use to electric.
Assumption 1. Non-electric energy demand in 2035 and 2050 (before we consider conversion to electric) stays the same as 2019, as discussed earlier for electricity only. From 2000 to 2019, non-electric primary energy demand increased just .2% a year.
Assumption 2. An exception to Assumption 1 is we assume 10% decline in primary energy for commercial and residential due to a presumed focus on making buildings more energy efficient, and the use of carbon pricing.
Assumption 3. Eliminating fossil fuel energy production saves energy by eliminating production, refining, and transport in the industrial sector (estimated at 15% of industry devoted to refining oil) and 5% for transportation. This is far more conservative than Jacobson’s assumption of 12.1% reduction in overall energy use due to ending the production of fossil fuel.
Assumption 4. We assume the following reductions in energy use for each sector to account for the immense efficiency gains of converting end uses to electric:
- Residential and Commercial – 75%. This is consistent with advanced air source heat pump house heating and water heating, and on the low side for geothermal. Cooking is somewhat less.
- Industrial – 50%. Much of industrial use is process heat, which can also be created more efficiently by directly using electricity. Some of this can come from heat pumps, direct solar heating, nuclear power, or burning hydrogen. Some processes can be changed to become dramatically more efficient.
- Transportation – 75%. Battery electric transportation is far more efficient than petroleum transport. Hydrogen fuel cells are also more efficient, but offer less of an efficiency advantage than batteries.
I explored some case examples of reduction in primary energy use in an earlier article. And my assumptions are informed by similar ones made by Jacobson, et al.. Another good resource for those interested in a deeper dive can be found in NREL’s “Electrification Future’s Study.”
The following compares annual electricity use in these three scenarios:
The first column above represents a fully electrified energy system with no efficiency gains for conversion of fossil fuels to electric. The second column is the system as it would be with all assumptions and 100% conversion – what we might hope to happen by 2050. The third column is the electric system we might expect to happen with a 50% conversion of non-electric to electric, and that is what we will assume for 2035.
The 2035 system represent a huge increase over current demand, going from 4153 TWh in 2019 to 7121 TWh in 2035. This will represent growth of 3.4% per year in annual electric demand versus basically no growth from 2005 to 2019. This may seem like a lot, but when I first got into the electric utility planning business, utility planners were expecting 7% growth. As it turns out, from 1960 to 2005 electricity production grew by 3.7% a year. And that included lots of spending on nuclear power during that period! Growing the electric system at 3.4% is doable, we just need to get back to the way we used to do it, but this time with wind and solar and batteries!
The 2035 system will also yield a huge reduction in CO2 emissions for the US. US emissions in 2019 were 4.8 gigatons (GT). The CO2 emissions of the 2035 system after 50% electrification and a 90% clean grid would be 2.2 GT, a 53% reduction. To get more CO2 reductions than 53% by 2035, we will need more electrification and/or more energy efficiency and conservation among the end uses that have not yet been electrified. I see lots of opportunity for that. Going from 90% to 100% on the grid could be an option, but the max benefit of that would only be another 3.5%.
The seasonal distribution of the 2035 system will be quite different from the 2019 system. Instead of a strong summer peak, the winter and summer peak months will be almost equal. This will change the nature of how we solve the intermittency problem. This is contrasted with the 2035 Report, which continues to show a July peak, but the 2035 Report has about half the electrification that I am assuming for this analysis for 2035.
How Much Wind & Solar Will We Need To Electrify 50% By 2035?
Earlier I noted that to green the grid to 90% by 2035 we would need to grow solar and wind 6.5 times from 2019 levels to 2640 TWh. If we also electrify 50% by then, we will need to grow wind and solar 13 times, adding an additional 2737 TWh for a total of 5377 TWh. Here are the monthly results:
Notice that with the winter peak in effect, our biggest need for fossil power in 2035 will no longer be in July and August, it will be in December and January when converted residential and commercial heating needs are combined with low solar output. Curtailments still occur in April and May and there are neither curtailments or fossil fuel use in June.
Adding 5377 TWh from a base of 407 TWh over 16 years is a lot of solar and wind AND a lot of transformation of direct fossil fuel end uses to electric. We will need to average production increases of 311 TWh a year, which basically means we need to add 75% of the solar and wind built to-date each year between now and 2035. Solar and wind need to grow at a 17.5% compound rate.
I ran some quick cost calculations and came up with about $200 billion a year for the US. Right now we are spending around $1.2 trillion a year on energy, so an investment of 1/6 of that amount actually doesn’t sound so bad, when you consider that the return on that investment is an energy infrastructure that is pollution-free and which no longer requires a continuing stream of fossil fuel mining and importing to support it, and which should lead over time to much, much lower spending on energy. That’s a future worth working for.
Conclusions
I plan on examining the implications for 2050 and 100% renewable in a second article, which is the apt place to compare to the Jacobson study.
My analysis gets into the same ballpark as the 2035 Report results using a simplified spreadsheet analysis. That I can mostly confirm their results this way makes me feel more confident in their very hopeful conclusions.
Technologies and markets are continually evolving. It is impossible today to anticipate what will end up being the best plan to achieve the energy transition we need in 2035 and 2050. Nevertheless, you can tease the data to yield insights about the future and to counter those who say it can’t be done.
Our best strategy as a nation is to aggressively price carbon, return the proceeds of the carbon fee to the people, and make selective investments in research and in clearing obstacles to the future we want. With carbon pricing in place, the market can choose that combination of resources that will be the best option. The two reports I have been studying and my own analysis give me confidence that if we do that we will build a much better future – one that is cleaner, healthier, wealthier, fairer, and doesn’t sacrifice in terms of the reliability of electricity, which is required for all those other attributes.
Appreciate CleanTechnica’s originality? Consider becoming a CleanTechnica member, supporter, or ambassador — or a patron on Patreon.
Sign up for our free daily newsletter or weekly newsletter to never miss a story.
Have a tip for CleanTechnica, want to advertise, or want to suggest a guest for our CleanTech Talk podcast? Contact us here.
Latest Cleantech Talk Episode
About the Author
Brad Rouse lives in Asheville, NC and is deeply involved in local efforts around the energy transition. He lobbies Congress for carbon fee and dividend as a volunteer for Citizens Climate Lobby. In 2016 Brad started a non-profit – Energy Savers Network – that mobilizes volunteers to help low income people save energy. He has a rooftop solar installation and his family cars are a Tesla Model 3 and a Prius Plug-in hybrid with 150,000 miles and still about 9 miles of EV only range. He has been studying energy economics for over forty years and holds a BA in economics from Yale University, where he learned about pricing pollution through a fee in freshman economics class. He also holds an MBA from the University of North Carolina at Chapel Hill.