Published on August 1st, 2020 |
by Brad Rouse
August 1st, 2020 by Brad Rouse
Please bear with me. I know this title is a turn-off to any but the most abject energy nerd, but this is a really important issue if we are to solve climate change. As noted in my first two articles (here and here) we’ve got to green the grid.
To green the grid, we must adopt a strategy of meeting our energy needs with low-cost renewable wind and solar resources. The obvious question is how to resolve the intermittency issue (what happens when the wind doesn’t blow or the sun doesn’t shine). Solving this problem is a necessary ingredient for “saving the planet.”
Wind and solar, now the lowest-cost sources of energy, are subject to “intermittency,” otherwise known as, “what to do if the wind isn’t blowing or the sun isn’t shining.”
Intermittency of wind and solar is a subset of the problem of resource adequacy – do grid operators have the resources to meet the demand for electricity at every hour?
The Real Problem: Resource Adequacy
The real problem is resource adequacy. Are the resources that I can deploy at this instant sufficient to meet the demand for electricity? Intermittency MIGHT be a problem if other resources are not available to step in when needed. The variation in solar or wind output is just a subset of the problem of resource adequacy, which includes such factors as:
- Variability of demand for electricity
- Unavailability of resources due to mechanical breakdown
- Inability of resources to respond quickly enough (or at all) to changes in demand
- Variations in supply due to the availability of sun or wind or water flow (intermittency)
Renewables add new wrinkles to the question of resource adequacy, helping in some ways and making it more challenging in others. The amount of electricity you can get from them is not fully controlled by electric grid operators, but is subject to variations in time of day, time of year, and weather. Renewables output is only controlled by grid operators to the extent that their output can be reduced.
Renewables help, however, because they come in small increments (1 or 2 megawatts (MW) versus 500 or 1000 MW for a fossil fueled plant). Any mechanical reliability problems that wind and solar might have do not require the same high level of backup capacity available at an instant that large fossil or nuclear plants require. It’s possible that a renewable dominated grid might be able to get by with lower reserves at peak demand than the current grid.
Finally, the state of emerging battery technologies offers cheaper, denser, lighter, and more powerful storage assets. Using different chemical and material properties, we can store energy output from renewable resources and discharge the power when needed for grid reliability.
A Personal Digression
The combination of my having studied economics and speaking “Southern” helped me get my first job as an energy consultant. My new employer had agreed to build a computer model to forecast the demand for electricity for the “Southern Company” which was then and still is one of the largest utilities in the US. Part of that was to build a “load shape forecasting model.” I had no clue what that was, but the client was paying good money, so I had to learn!
It turned out the load shape is just the hour by hour demand for electricity over the course of the day, week, month, and year. It’s critical in utility planning, because you have to meet that demand with power resources every minute of every day. You can imagine how blown away this newly minted MBA was to learn that they wanted us to forecast 8760 data points a year! For 30 years! We never did make an accurate load shape forecast, but we did have some useful insights along the way.
Let me show you what they meant when they talked about a load shape. The Energy Information Administration (EIA) publishes hourly loads for every “balancing area” of the US grid. In my case this is the Western North Carolina (WNC) balancing area hourly load shape for my area for two days this year:
In my next job, I went from working on forecasting electricity demand to planning the electric supply system. My colleagues specialized in modeling to simulate the power system. Their secret ingredient was a breakthrough in modeling the unavailability of resources due to mechanical breakdown. I was a lonely economist surrounded by engineers! But I learned a lot more about intermittency and developed software for long-range planning that is still used by power companies today. This problem of resource adequacy is something that grid operators have long been dealing with.
Solar & Wind Turn Utility Planning On Its Head
Don’t just think that because solar and wind are cheaper than gas or coal that they will immediately take over. They need a little help. The problem is that the amount of output hour by hour is not under the control of grid operators and is often a severe mismatch with the hourly electric demand.
Let’s do a thought experiment using the 7.5 kilowatt (KW) solar array on my home in Asheville as an example. The hourly output on the February and July days are below (note, my panels are west-facing so their output peaks later in the day than south-facing panels and is at a particular disadvantage in the winter). So here’s a thought experiment. Let’s scale up my solar panels to try to meet the entire regional demand for electricity for the day (but not hour by hour) for those specific January and July days. Here’s the profile for that very day in February:
Wow. A HUGE amount of solar is required to do this. I know this is true because at my home I run surpluses in the summer and then I don’t have enough solar output in the winter. I use Duke Energy as my giant battery! But as you might expect, if we took the incumbent utility out of the picture, we would need a giant battery charging during the day that could meet the load at night. (That is made worse by assuming that the roundtrip efficiency of charging and discharging a utility scale battery is 85%.) Meeting the energy needs for that day would require 8 gigawatts (GW) of solar, or about 20 KW per person in the region. At say $1 per watt, a good price these days, it would cost a cool $8 billion. And on top of that we would need storage for 8.5 gigawatt-hours (GWH) of electricity to meet the demand at night, which at a cost of $100 per kilowatt-hour (kwh) (the holy grail battery price for EV dominance) would cost another $850 million.
But look at the graph below representing that July day. To meet the energy needs that day we need only about 2 GW of solar versus 8 GW on the winter day. And due to the longer hours of summer sunlight, we need 6.7 GWH of storage versus 8.4 GWH on the winter day.
I provide this illustration because in a fully renewable electric system, this is the kind of mismatch that utility planners will have to deal with. Fortunately, a lot of factors will work to make the actual solution much more workable than this admittedly far-fetched example, which is nevertheless what an innocent bystander might understand when they hear the words – “but, but, but, but, but the sun isn’t always shining and the wind isn’t always blowing”!
Adding Supply Diversity To The Mix
There are numerous solutions to the problem of resource adequacy, and many of them are probably more economical than just adding batteries. The most obvious is to increase supply diversity! There are lots of ways to do this – adding wind power (offshore and land-based), increasing transmission ties, adding solar facing in different directions, etc.. Let’s expand the thought experiment with wind power. In this case we add wind in equal proportions to solar on the July day. How much storage would we need in this case?
For this thought experiment, I used the hourly profile of wind power in Texas on similar days. (I know, you can’t get Texas wind to WNC at the moment, but maybe in the future.) I assume that half of the daily energy need is met with wind and half from solar. Good news! Wind blows at night AND it blows more in the winter than in the summer. To simplify, I assumed that wind and solar would each meet half of the daily energy for the July day. Then, given how much wind and solar that amounted to be, I would see if I had enough to meet the day in February.
With the same amount of energy coming from wind and solar on the July day, there is much less storage needed to meet the load than just with solar. The problems of resource inadequacy and intermittency have been reduced. Diversity helps! The load and renewables are well matched in the early morning hours while batteries are needed to supply power till around noon. From noon to about 9:00 PM solar is charging the batteries, and battery power is needed again during TV “prime time”’.
When that same MW of wind and solar are applied to the February day, we find that there is very little solar needed that day (thank goodness) since there is much more wind output on that day, and in fact the whole system is surplus to the point that the amount of energy available to be stored exceeds the energy needed for that day:
Battery discharge is needed in the morning and a tiny bit in the early evening, but otherwise the system is producing more energy than needed. Depending on the battery capacity, the system may have a “curtailment” event (so much solar and wind that the batteries can’t hold it all).
Bottom line: I have added just one potential solution to the mix and had a dramatic reduction in the amount of battery storage needed. If this were a real utility planning exercise I would have much more powerful analytical tools at my disposal and would be able to draw from many other options to ensuring resource adequacy. My conclusion is that resource adequacy is a very solvable problem. From a policy perspective, of course, we need to continue to improve technology through research and there may be options for targeted government investments. Overall, solving this problem is well within the experience of utility planners, but it takes a new mindset that starts from the idea of meeting the energy needs and then having a set of tools like energy storage to allow exact matching of supply and demand.
I’ll dig into these issues and examine some comprehensive studies of this subject in a later article.
Carbon Pricing Can Play A Huge Role In This Part Of The Energy Transition
I’m a big fan of putting a price on carbon because it sends a signal to all players in the economy that they have a role in the energy transition. And the signal is, you will be paid according to your contribution to reducing your carbon footprint. Electric companies today have a huge carbon footprint (27% of total carbon emissions) and they will have a huge incentive to reduce emissions.
My favorite carbon fee proposal is the Carbon Fee and Dividend proposal filed as a bill in Congress called the Energy Innovation and Dividend Act (EICDA). This bill calls for a rising goal-based fee on carbon with all revenues returned to Americans in the form of a dividend.
The EICDA will increase the rewards to finding solutions to resource adequacy problems to the extent that they reduce the carbon footprint of the grid. A simple way to look at this is based on the economics of bringing on battery storage to allow substitution of renewable energy for fossil energy. At the current time, most utilities can simply add renewables and reduce fossil fuel use and reap the benefits. Battery storage will come into its own when there is too much zero carbon energy for the grid to handle without moving that energy to a different time period. This can come either when the renewable resource is likely to be curtailed or when the economics favor increasing fossil fuel use at one time and decreasing it at another, more carbon intensive time.
A simple approach to understanding how this will play out is to compare the cost of adding battery storage to the resulting decline in fossil fuel cost from the storage being utilized. The factors that go into this evaluation include the cost per megawatt-hour (MWH) to produce electricity with fossil fuels, the cost of the battery, the lifetime in cycles of the battery (number of times you can expect this battery to actually reduce the fossil fuel cost), and the charge/discharge efficiency of the battery. Carbon pricing affects the cost to produce with fossil fuels.
For a new thought experiment, let’s look at the economics of batteries compared to the natural gas peaking plant, which is normally used as an option to provide resource adequacy. We assume the battery has a 20-year life and goes through 100 cycles per year, for 2000 cycles over its life, with a round trip efficiency of 85%. The cost of battery storage is assumed to decline in accordance with “Cost Projections for Utility-Scale Battery Storage,” a June 2019 study from the National Renewable Energy Laboratory (NREL). We use the mid case scenario (from $287 per kwh today to $76 by 2050). Carbon fees rise in accordance with the low boundary carbon fee increases under the EICDA. Both the carbon fees and gas costs are consistent with my earlier analysis in this series.
The economics move positive (green bars) with the projected decline in storage cost, and they become extremely high with a fee on carbon. And even though the economics are not positive today based on gas costs alone, storage is being added to the grid anyway due to the other benefits of storage beyond the simple cost per MWH comparison – particularly the small size and quick construction time of storage versus the much larger size of a peaking plant, which means that a utility can bring on storage and more evenly match it to the need as it evolves.
But what of the case of an existing combustion turbine (peaker) plant? For plants already in service, it does not pay to bring on a battery to offset its use unless carbon is priced, even with the super cheap batteries expected by 2050. Clearly carbon pricing, or some sort of mandate, will be required. The following graph shows the situation:
Grid intermittency from cheap renewable energy brings new problems to grid operators and planners as we add more and more renewable energy to the grid. Fortunately, there are many tools at their disposal, chief among them (1) seeking a diversity of zero carbon supply resources and (2) storage batteries, which are declining in cost. Incorporating a price on carbon into grid planning and operations decisions will be one effective mechanism that will result in solutions becoming more and more economically.
Have a tip for CleanTechnica? Send us an email: [email protected]
Latest Cleantech Talk Episodes
Latest CleanTechnica.TV Episode