Modeling Alternatives to Buying More Natural Gas Generation for Platte River Power Authority

00:11 - Platte River Power Authority, which covers the power generation for Fort Collins, Loveland, Estes Park, and Longmont, is leading the state in its efforts to transition to renewable energy. They’ve created a very detailed plan on how they want to transition to mostly renewables by 2030, and they think they can do it at business as usual costs, maintaining their reputation for having high reliability and some of the lowest rates in Colorado. In their plan they laid out four different options on how they might want to accomplish this transition, and they chose their second option which involves completely retiring all coal generation that is in use today and promises a 90% reduction in CO2 emissions by 2030.

00:50 - This plan was approved by PRPA’s board of directors in October of 2020. In this talk, I want to answer these questions: what is PRPA’s plan for 2030, and what does it entail? And in this plan they are proposing new natural gas generation; do we actually need it? PRPA has made it clear that this is a default solution for them and that they are planning on exploring non-fossil fuel alternatives. So what are the non-fossil fuel alternatives that we could be spending that money on instead? To answer these questions I’m going to be looking at this model created by Energy Should Be in excel.

Models don’t predict the future, but they can help to better understand and explain certain problems. This model is based on PRPA data from 2018, and in it you can input the amount of storage, renewable generation, coal generation, etc, and it gives you these graphs and other outputs which model what PRPA might look like in the future based on how it looked in 2018.

01:46 - So we can input PRPA’s numbers for their 2030 plan into our model to get a picture of what it might look like. This table shows where PRPA is now, versus where they want us to be in 2030. As you can see, they want to completely retire the 429 megawatts of coal generation that is in use today. They want to build a lot more solar and wind generation, and a lot more storage; they want to build 300 megawatts of storage in four hour batteries for a total of 1200 megawatt hours of storage.

And, in addition to these new renewable resources, they want to build a new RICE plant. RICE stands for reciprocal internal combustion engine, and that’s a new natural gas plant that would add 104 megawatts of natural gas to the 388 megawatts that PRPA already owns. This is a picture of PRPA’s fossil generation and what it would look like in 2030. Again, they want to completely retire the two coal plants that are in use today. But, crucially, the five gas turbines that they own right now which make up 388 megawatts will still be operational in 2030.

Our question is: while we are trying to transition to renewable energy–we’re retiring coal, we’re building a lot more wind and solar–why are we building new fossil fuel generation on top of the fossil generation that is already in use? Additionally, in order to get an understanding of what PRPA’s power generation and use will look like in 2030, we need to understand how the load, or the total power used by all of PRPA’s consumers, will rise by 2030.

This graph from PRPA’s plan shows several different predictions of where they think load might be in the future. And I wanted to look at PRPA’s highest predicted load in order to get a better understanding of a worst-case scenario, the hardest scenario to cover with our existing power generation. PRPA’s highest load forecast predicted a peak load of a little under 900 megawatts in 2030 so I set the peak load of our model to be 880 megawatts in order to model this scenario.

Now we can get into the actual model. This is the graph that shows the entire year of our modeled 2030 scenario, and it has data plotted at every single hour of the year. In this graph this red line is our load, and again, that’s the total amount of power used by all of PRPA at any given hour. Green is renewable generation, and yellow is unmet load; that’s anywhere where our renewable generation is not meeting our load and so that is power that will have to be made up by natural gas in our scenario.

Because we want to understand how much natural gas we will need in 2030, we wanted to look at dark-calms. Dark-calms are periods of low sunlight, hence dark, and low wind, hence calm, which lead to low renewable generation so we’re going to have to rely on natural gas to keep the lights on in those scenarios. To find these dark-calms, I had the model look for the biggest periods of uninterrupted unmet load; essentially the biggest areas of yellow on the graph.

And these bars at the top of the graph signify the 10 worst dark columns that the model found. We can actually meet all but two of these dark-calms using just our existing 388 megawatts of natural gas and by charging the battery storage that PRPA wants to buy with unused natural gas capacity. And let me zoom in on our worst dark-calm here in February so I can explain more. This is the graph zoomed directly in on our February dark-calm, and again as you can see between these two pink bars which mark the period of the dark-calm, there’s not enough green renewable generation to meet this red load line and so all of this yellow area is unmet load that has to be met by natural gas.

And again because we’re trying to understand how much natural gas we actually need, it’s important to understand how much unmet load there is. It’s a little hard to tell how much unmet load there is on this graph because there’s a lot going on, so let’s take a look at a graph of the unmet load by itself. This is the same area of yellow here as it is right here only now it’s graphed by itself. And now that we can see how much unmet load there is, we can add in our existing natural gas to this graph.

So here we have our existing natural gas added in. This line is our existing gas capacity line, and that’s at 334 megawatts which is our 388 megawatt capacity minus a 14% that has to be kept in reserve per FERC regulations. Then this gray area is unmet load that is now met by that existing gas capacity. The yellow are peaks above the existing gas capacity line. That’s where we can’t cover our unmet load using our existing gas capacity. The green area is unused gas capacity.

That’s anywhere where we could be using natural gas, but we’re not using it to meet the load, so it’s unused gas capacity. It’s available to be used. Now we have these peaks above our existing gas capacity line, but we could charge our 1200 megawatt hours of storage that PRPA is planning on buying with this unused, available natural gas capacity. However, the reason why this is our worst dark-calm that the model found is that would actually still not be enough.

The biggest peak above the existing gas capacity line is 1970 megawatt hours, which is above the size of the battery that PRPA is planning on, so there’s 770 megawatt hours that we can’t cover in this dark-calm. We have a similar story in our second worst dark-calm here in July. These two peaks add up to about 1300 megawatt hours, but we could actually cover them because of this bit of unused gas capacity which is a little over 100 megawatt hours. So if we charge the battery using this bit, then we can completely meet those two peaks.

The same could not be said of this peak here, which is 1220 megawatt hours, and that is just above the size of the battery that PRPA is planning on. Here’s our problem: the model found two occasions where we can’t quite meet load with our planned battery storage and our existing natural gas capacity. This bar graph here shows the size of the peaks above the existing gas capacity line, and as you can see we can’t quite cover them. PRPA’s default solution to this problem is to buy enough natural gas so that we could meet load during these two dark-calms.

That would cost us 140 million dollars up front. And our question is: what else can we buy for 140 million dollars for what is a twice-a-year problem? One alternative we can look at is buying enough battery storage so that we could cover the peaks above the existing gas capacity line. If we bought 800 additional megawatt hours of storage, for a total of 2000 megawatt hours of storage, we could completely cover our worst February dark-calm and easily cover any other dark-calm that the model found.

As you can see it can cover the July dark-calm with a lot of unused storage capacity to spare.

08:39 - Our model is based on only one year of data from 2018, and so you can imagine that there are years of data out there with worse and longer dark-calms than the one that we have access to. To simulate this scenario we thought, well, what if we had several of our worst dark-calms all in a row? So here we have our February dark-calm, followed directly by our July dark-calm, followed by our February dark-calm again. And if we had 800 additional megawatt hours of battery storage, for a total of 2000 megawatt hours, we could completely cover even this worst-case scenario.

There’s enough unused gas capacity in each of these green regions here to completely charge the battery each time and completely cover each of these peaks above the existing gas capacity line. Additionally, battery costs are falling, and they’re likely falling faster than PRPA anticipates. In their plan, PRPA supplied this graph which is a prediction of what they think 4-hour batteries will cost in the future and in 2030 they believe that four hour batteries will cost eight dollars a kilowatt month, over the battery’s 15-year lifetime, which is equal to 360 dollars a kilowatt hour.

If we compare this to another reputable source, NREL’s annual technology baseline for 2020, NREL’s mid-level prediction for what four hour batteries will cost in the future predicts that batteries will cost only two hundred dollars a kilowatt hour in 2030, which is significantly lower than the cost that PRPA is predicting. And it’s worth noting that even NREL’s high-end prediction for what batteries might cost in the future is still lower than the price that PRPA is predicting.

10:12 - Now we can compare the battery costs to the price of the natural gas generation that PRPA is planning on. In their plan they supplied this table which lists different prices of different natural gas generators, and they’ve selected this RICE plant which I’ve highlighted in yellow. That RICE plant has an installation cost of one thousand two hundred and fifty two dollars a kilowatt. And that’s where I got my price of 140 million dollars of a total upfront cost for the RICE plant that PRPA is planning on.

If we use PRPA’s numbers for what batteries will cost in 2030, they believe that 800 megawatt hours of additional batteries will be significantly more expensive than that new RICE plant that they selected. However if we use NREL’s price prediction, NREL’s price prediction says that 800 megawatt hours of batteries will be a similar cost to the new RICE plant that PRPA is planning on. Now they might be slightly more expensive, but there are a couple reasons why batteries might still be a worthwhile investment.

For one, by definition, every single time that we use a RICE plant it will be adding more carbon emissions into the atmosphere. So if we take into account the environmental and social costs of those additional carbon emissions, a new RICE plant would be even more expensive than the 140 million dollar price that is listed here. Additionally, the RICE plant can only be used on the couple of occasions where we don’t have quite enough power generation to meet our load during dark-calms.

Whereas batteries can be used year round and they also cut down on renewable surplus so we can more effectively utilize the renewable resources that we have, and it gets us closer to our end goal of 100% renewable energy.

11:53 - Another alternative that we can look at is demand management. Demand management means essentially lowering our load or shifting our load so that we can more easily meet it using our existing power generation. One way of doing this is that PRPA actually has the ability to reduce the voltage that it sends out to consumers. For example, electrical sockets in your house operate at around 120 volts. But they actually operate at a range around 120 volts; they can operate at voltages higher than 120, and they can operate at voltages as low as 110 volts.

So during a period of a dark-calm, PRPA can reduce the voltage that it sends out to consumers, and consumers would only experience very minimal effects. For example, maybe your incandescent bulbs might be the slightest bit dimmer, but it would be almost imperceptible. Another tool of demand management is that PRPA can have people operate at off-peak times. So PRPA can communicate with industry and individuals to have them operate at off-peak times. For example, factories could operate in the early morning or the late evening instead of midday when most people are using electricity, and that will reduce the peak load because we’re not using as much energy all at once.

However, this is actually more of the historical way of looking at this tool demand management, and when we’re relying on high amounts of renewable energy, the story changes slightly.

13:12 - Here we have the entire graph of our worst dark-calm again and as you can see, right here we actually have more renewable energy than we need to meet the load. Whereas right here, we have very low renewable energy. So instead of having people simply operate at off-peak times, PRPA can actually ask people to operate when there is more renewable generation available. Say, shifting our peak from here to here. This more effectively alleviates dark-calms, and it maximizes our use of renewables.

We call this “time of renewables. ” Here again is a graph of the worst dark-calm with our load unchanged, no demand management, and it has a maximum load of 580 megawatts. If we reduce our load only five percent, taking our maximum load from 580 megawatts to 550 megawatts, we take our dark-calm from looking like this with 1970 megawatt hours that we can’t cover, to this where the peak above the existing gas capacity line is only 1200 megawatt hours, which is, as you’ll recall, the size of the battery that PRPA is planning on buying in 2030.

So with only a five percent reduction in load, we can now cover this worst dark-calm using only our existing 388 megawatts of natural gas capacity and our planned 1200 megawatt hours of storage. Another option we can look into is changing the way we dispatch our hydro power. Again this is our graph of our worst dark-calm, only now it’s just showing our hydro power generation. And as you can see, right now hydro power is dispatched to follow our load. It’s dispatched at higher levels when we are using more energy, and at lower levels when we’re using less energy.

Now this does make sense, but like the last tool of demand management that I talked about, this is more of the historical way of looking at things. And again if we look at our entire graph of our worst dark-calm, right here we have a lot of wind and solar generation–more than we need to meet the load–whereas right here we have essentially no wind and solar generation and we’re entirely relying on hydro power in terms of our renewable generation. So instead of having hydro power follow our load, it might make sense to dispatch hydro power at lower levels when there’s high amounts of wind and solar generation, and at higher levels when there’s low amounts of wind and solar generation.

So here is our hydro power zoomed directly in on that hydro power generation, and here is where it is following the load–this is how it is dispatched normally–and here is where I have redistributed hydro power so that instead of following our load, it’s dispatched to follow our dark-calms. So again, here is our regular normal dark-calm, our worst dark-calm, and here it is where I’ve dispatched hydro power differently. And these two peaks add up to about 1260 megawatt hours, and this bit of unused gas is about 60 megawatt hours, so again now we can cover our worst dark-calm using just our existing natural gas capacity and our planned storage capacity.

16:01 - Another alternative is the energy imbalance market. PRPA will join the Western Energy Imbalance Market, or WEIM for short, in 2022. In WEIM, different power authorities can join to exchange energy. Some areas may experience dark-calms while others are experiencing surplus, and the energy imbalance market allows power authorities who have surplus energy to sell that surplus to other authorities who need it. So in 2030, when we will have already joined this market, PRPA will be able to buy energy from neighboring power authorities during dark-calms, so that we can more easily meet our load.

Now this may not be able to entirely solve our problem; weather patterns are generally larger than the area of one power authority. Usually multiple power authorities are experiencing dark-calms or experiencing surplus at the same time. However, it could only help us, and we will have already joined this market in 2030 when this plan goes into effect.

16:53 - One last option that I want to look into is electric vehicles, or specifically what’s called electric vehicle to grid charging. For some background, Colorado Energy Office put out a plan in 2020 which stated a goal for Colorado to have 940,000 electric vehicles within its borders by 2030. According to U. S. census data from 2019, PRPA’s cities make up a little over six percent of Colorado’s population. So we can estimate that there will be around 60,000 electric vehicles within PRPA’s borders by 2030.

Within a few years, and certainly by 2030, most electric vehicles will have battery capacities of around 100 kilowatt hours. And with 60,000 electric vehicles, that’s 6000 megawatt hours of storage. To take advantage of this massive amount of battery storage, PRPA can actually consider incentivizing customers to discharge some of their battery capacity back to the grid when there are dark-calms ahead. If there’s six thousand megawatt hours of battery capacity in electric vehicles throughout PRPA, we would only need 13% of that to cover our worst dark-calm.

If you recall we needed about 800 additional megawatt hours in order to cover that worst February dark-calm. Right now, there are only prototypes of vehicle to grid charging, but PRPA should definitely investigate this possibility, especially as electric vehicles are going to become more and more prevalent. To restate our problem, we have these two dark-calms that we can’t quite meet with our existing natural gas capacity and planned battery storage capacity.

PRPA’s default solution to this problem is to buy enough natural gas [generation] so that we can meet load during these dark-calms. However, they have made clear that they are planning on exploring non-fossil fuel solutions. And these non-fossil fuel solutions can include things like: buying enough battery storage so that we can cover those peaks above our existing gas capacity line, we can use the tools of demand management to lower and shift our load so that we can more easily meet it using our existing power generation, we can and will join the Western Energy Imbalance Market so that we can buy power from neighboring authorities in our times of need, we can look into dispatching hydro power differently so that instead of following our load it follows our dark-calms, and we can investigate electric vehicle to grid charging so that we can take advantage of the massive amount of battery storage that will be in electric vehicles by 2030.

19:09 - There are multiple next steps that we can take to further our analysis, but one big development that came up recently is that in mid-February a significant cold snap hit large portions of the country, and it especially affected Texas. In Colorado, although we were slightly more prepared for those low temperatures, we still experienced some problems. For example, like Texas, PRPA had trouble getting access to natural gas to fuel its natural gas generation.

This is probably a worse dark-calm than what is found in the 2018 data that our model is based on. So this really emphasizes to us that we need to be looking at more years of data. Research into dark-calms is still ongoing, but it will be really helpful to analyze data surrounding this new February cold snap to compare to other years of data and to our own model to see what we can learn.

19:55 - To conclude, PRPA wants to buy 104 additional megawatts of natural gas generation, but our model found only two occasions where that additional natural gas would actually be needed. We really don’t have to resort to buying more fossil fuel. There are multiple non-fossil fuel alternatives like buying more battery storage, demand management, hydro dispatch, electric vehicle to grid charging, and the Western Energy Imbalance Market. We could spend 140 million dollars on new fossil fuel generation that would only be used a few times a year, or we could spend that money on other solutions.

If you have any questions, please feel free to contact me at Kirya@EnergyShouldBe. org.