NREL Research on Renewables Backup

It's from 15 years ago - but it speaks to backup reserves

First off, I was on a panel discussion about Nuclear in Colorado. It was a really good discussion. You can watch it here.

One thing that has really bothered my from the start on my research into Wind is that I could find nothing from NREL about it. I didn’t think they would purposely not evaluate this issue. And now we have it - Wind and Solar on the Power Grid: Myths and Misperceptions.¹

First a sad comment - this was funded by USAID. This research can help us find the most effective solution to a carbon free future. Shutting down USAID doesn’t just harm people overseas, it harms us.

Myths and Misperceptions

This 2 pager is a summation of 5 studies (pertinent ones discussed below). There are several big issues about all this overall.

1. The studies were published in 2009 - 2014. So it’s based on the grid as it existed 10 - 15 years ago. A lot has changed since then, especially in how the BAs spin peaker plants up/down.

2. These studies view 25% of electricity provided by VREs² as large. Nowadays Germany provides 60% of its electricity from VREs. We in Colorado are aiming for 95%. Very different worlds.

3. It calls wind a money saver from the fuel savings as the backup generators now run only 70% of the time. That’s true. It’s also only part of the picture unless the wind turbines were manufactured and installed for free.³

First a major note - I did not read all of the below documents in detail. That’s 400 pages total for just the first 3 studies. But fear not, Perplexity (R1) to the rescue. I both read what I thought were the relevant parts and used the AI’s summation of the reports.⁴

Milligan and Kirby [1]

The first study is Market Characteristics for Efficient Integration of Variable Generation in the Western Interconnection.

They show three major ways to address wind backup. The first is to pull wind from turbines spread across a wide area. This is valid. However, it will require building a ton of new transmission lines across the country, at a cost of $1m - $7m/mile. And the wind is still, albeit less, variable.

Second is to consolidate balancing areas. This is interesting as a larger region will by definition even out some of the variability not just of the VREs but of demand also. On the flip side, it requires additional transmission lines for this to work. And problems with a blown transformer, fallen branch, etc. will then impact a larger area. This definitely should be looked at, balancing out the pros & cons, and adjust the balancing areas where they can.

And third, to use the existing flexible generation as needed. This is what we’re doing today. One interesting point on this, the study rejects the notion of 1:1 backup generation capability.⁵ We have numbers from eia showing zero wind at night at times in the PSCO balancing area. If we don’t have 1:1 we’ll have rolling blackouts at times.

I think the best example of how dated this report is, is the following from it:

Evidence from the United States is corroborated by evidence in Europe. Denmark, Germany, and Spain have all integrated large amounts of wind generation into their power systems. Their experience points to the benefits of operating in a region with a robust spot electricity market. (Holttinen, 2009)

The report was well before Germany pushed their wind up to 60% and shut down their nuclear plants. The above phrase does not reflect the situation today.⁶

Nothing in the report about the CO2 impact of the backup power sources. And nothing about the cost of the proposed improvements to the transmission network they’re suggesting.

Ela, Milligan, & Kirby

The second study is Operating Reserves and Variable Generation. This does get in to backup emissions - yah! But no explicit numbers - boo!

Their first point is no explicit backup is needed for wind as we can use the existing flexible resources. Well duh.⁷

Their second point is that reducing gas plant start/stops will lower emissions. This is a very good point and I assume that in the 15 years since this study, they’ve gotten good at this.

Their third point is to use low emission generators for backup/reserves. And I agree hydro is a better backup for no wind. Unfortunately there isn’t that much hydro.

Their final point is:

Increased gas use for reserves is offset by reduced coal baseload generation. The study cites the Eastern Wind Integration Study (EWITS), where 20% wind penetration reduced CO₂ by 25% despite added gas flexibility.

Well yes. Replacing coal with gas is a big win. Replacing coal with wind+gas is a big win. But no discussion of running a CCGT 24/7 vs wind + SCGT backup (or wind + CCGT backup).

Lew, Brinkman, Ibanez, …

The last relevant one is The Western Wind and Solar Integration Study Phase 2.

This is the best of all the studies I think. They dive into a lot and lay out their findings & solutions. With that said, a lot is a repeat of the above and so I’ll just hit the CO2 emission conclusions.

The following are the Perplexity summation of the CO2 output in the research results. I’m using this because the actual content in the study is long and complex which almost no one will want to work through. I think these summations are an accurate synopsis.

Gas plants as primary flexibility providers: Fast-ramping natural gas turbines are the most cost-effective solution for sub-hourly variability, but their runtime is minimized through optimized scheduling.

30% emissions reduction: Even with increased gas plant cycling to balance renewables, the study estimates a 30% reduction in CO₂ emissions compared to conventional fossil-dominated grids. This stems from displacing coal and reducing baseload fossil generation.

Gas plant efficiency mitigates emissions: Modern gas plants used for balancing operate at high efficiency (∼50% combined cycle), and their intermittent use results in lower cumulative emissions than coal baseload.

Renewables displace fossil generation: At 30% wind/solar penetration, coal generation drops by ∼40%, driving the bulk of emissions reductions.

I agree with everything stated there. Replacing coal with anything is a win. But I could not find, and Perplexity could not find, a comparison of wind + gas vs gas.

The Last Two

The fourth report is not a study more a summation of earlier work, that’s even older. And some calculations that less rigorous than what I have done in some of my blog posts.

The fifth report is good, but is about power controls so not germane to the discussion about backup resources.

Conclusion

The reports bring up a lot of good points. I think the increase in transmission lines proposed in places is way too expensive. But that’s open to discussion. And adjusting the balancing authority areas - well worth a look.

Of course replacing coal with wind + gas is a reduction in CO2, and a lot of other nasty stuff coal emits. And doing a better job spinning gas plants up/down - that’s a good win.

But nothing on the CO2 emissions of a CCGT running 24/7 vs. wind + gas. That’s the research we need. And also to measure wind + SCGT vs wind + CCGT. Ireland tried CCGT as backup and found it worse. But maybe the technology has improved.

So my conclusions are:

1. It was worth my time to dive in - this might have disproved my work.

2. Using Perplexity to summarize studies is amazing.⁸

3. I see nothing to disprove my work that says wind + gas emits more CO2 than just gas.

4. NREL needs to do a new study today.

And if anyone finds a study that does look at wind + gas vs gas, please tell me in the comments below. I would be thrilled to be proved wrong as I hate to think the hundreds of billions spent on wind was to no use.

And NREL, maybe you can get the PSCO balancing authority to hold coal constant for 2 months. The first month use SCGTs exclusively for wind/solar backup. The second month use CCGTs exclusively for wind/solar backup. Measure the generation and CO2 emissions from the gas turbines. And see if you can measure the use of gas backup in California and Germany. We need a detailed study of wind vs gas.

1. Thank you Bill Dugan

2. Variable Renewable Energy

3. Little games like this to favor wind annoy the snot out of me. When they do this, for me, it reduces their credibility. Own up to the fact wind + backup costs more.

4. I spot checked the AI summary - it was spot on each time. And it identified several point I missed in my skimming.

5. They do state “market access to flexibility” helps eliminate the 1:1 which I take as there will be companies providing power if you pay enough. Which I think means 1:1.

6. Dunkelflaute

7. By definition every electrical generator is flexible. Some just take longer to spin up/down.

8. Always check the AI results. But I have yet to find an issue when asked to summarize. (And if you’re a little nervous, use several AIs. I use Gemini, Perplexity, DeepSeek, & Qwen.)