Integrating Wind – a Theoretical Example

Wind is a variable resource, and its output varies depending on the winds. But consumers demand the services of electricity at their convenience, not just when the wind blows. So how can an electrical grid accept wind power, and what is the effect of wind power on emissions from fossil generating plants?

The supply and demand for electricity must always be in balance. So when demand increases, more power must be generated. Similarly, when demand falls, generation must be take off line.

Lets assume that we have an island, that is not connected to any other power system. The island has a minimum demand of 50 MW, and a maximum demand of 100 MW. The average demand is 70 MW. The island has 21 natural gas fired generating stations, each 5 MW in capacity, and which can be run at any level from 0 to 5 MW. It takes 1 hour to start up any generating station. How can we ensure that demand always equals supply?

One way would be to always have all 21 generators running, so we are ready to increase electricity supply at any time. We would be able to increase supply to 100 MW any time we wanted, and we even have an extra plant running, in case one failed. This arrangement would of course be very wasteful of fuel, but if our demand varied quickly, and without notice, it might be what we need to do.

To save fuel, the grid operator studies the nature of demand. How quickly does demand increase or decrease? Is there a weather related pattern to demand? Does outdoor temperature matter? Does sun affect air conditioning load, or decrease heating load? Does wind cause wind chill effect in buildings, increasing heating demand? Is demand different at night, when people aren’t working? Do sunrise and sunset times affect lighting requirements? Are there different demand patterns on holidays, or weekends? If we can estimate demand with 100% accuracy, and with a 1 hour notice, then we would only ever run exactly the number of generating stations we need, with a huge savings in fuel.

Grid operators spend millions forecasting demand for exactly this reason. It allows a more efficient dispatch of generating stations, and therefore saves money.

Our island has smart grid operators, so they made the investment in forecasting. And they achieved an accuracy level that means they are never more than 10% in error. So how much extra generation do they need to run? Only 10 MW. That’s way better than the 35 MW average they had before forecasting.

Now lets introduce 20 MW of wind. The wind will produce 6 MW on average, or 9% of the average demand of 70 MW. And production will range between 0 and 20 MW, depending on the wind. What changes will this allow in our system? How much natural gas will we save?

If we have no forecasting, and if wind can go from 0 to 20 MW and back to zero instantly, then we have to burn just as much fuel as before. We still need to keep our 10 MW of margin available. After all, the output from wind could be zero in the moment that we need the 10 MW due to our demand forecast error.

But in practice, wind does not go from 0 to 20 instantly. Winds pick up, and drop back, but they do so over time. So our smart system operator will come up with forecasts of wind. With simplistic forecasting, you would simply make the assumption that the output next hour will be the same as the last hour. And with this type of forecast you would be within about 10% of the installed capacity, or 2 MW.

So, lets assume that our wind output never varies by more than 2 MW. How much generation do we need to add to firm up the wind? We need to add no new capacity, as we already have 105 MW of capacity to meet 100 MW of demand. And we change from needing 10 MW of extra generation to 12 MW running at any one time. We need the 12 MW when we had a simultaneous unexpected increase in demand of 10 MW, at the same time the wind dropped by 2 MW. In this scenario, fossil fuel use would drop by an average of 6 MW of wind production minus the 2 MW of extra reserve. So wind, in this simplistic example, would reduce fossil fuel use by 2/3 of the output from the wind, or 4 MW.

But remember, we have smart system operators. What can they do to reduce the 12 MW reserve, thus saving even more fuel? It turns out there is lots they can do. They can improve the wind forecasting. If our forecast ends up better than just using the previous hour’s output, we can reduce the 2 MW added requirement. They can interconnect to another jurisdiction. If you had two equal islands as described above, you could have only 41 natural gas plants, not 42, as one redundant plant may be able to back up both systems. And you could probably reduce your reserves from 10 MW per island to 7, as the odds of simultanous forecasting error is reduce.  And the bigger your total system, the less variability wind has, as the wind doesn’t stop everywhere at once. You can install some “instant on” and “instant off” facilities. Waterpower often works in this way. If you had just 12 MW of “instant on/off” facilities, you could stop running extra generation altogether, and your windpower would be firm. You could implement load shedding, where hot water heaters, or air conditioners, or municipal water pumping facilities, or pool pumps, could be turned off when demand unexpectedly jumped. Ontario’s proposed smart meter program may allow some of this. If the system has pumped storage, you could run the pumps only when surplus electricity is available. The Cities of Guelph and Peterbourough used to have the ability to shut off all the hot water heaters on demand. This is neither new, nor far fetched, but it is an opportunity that Ontario has only recently begun to look at again. In the future, hydrogen production, or plug in hybrid vehicles may offer interesting ways to vary electricity demand at a moment’s notice.

How does this simple example apply to Ontario? Peak demand is 27,000 MW, minimum is 13,000, and average 17500. We have 31,000 MW of installed capacity. We have 14,000 MW of nuclear, and 1000 MW of natural gas co-generation, that cannot be either turned up or down. We have 6500 MW of coal that can be turned up, but requires a few hour lead time to get to full capacity. We have 7000 MW of waterpower, of which half can be turned up and down very quickly. We have about 4000 MW of gas that can be adjusted with about an hour’s notice.  And we have 400 MW of wind.

Studies done on Ontario’s electricity system demonstrate that we can add 5000 MW of wind with virtually no need to add back-up capacity, load shedding, or “instant on” capability.  And virtually all of the production from wind will reduce our use of fossil fuels.  This proportion of wind has been added in other jurisdictions, also with almost no need to make other significant changes.  Claims made by some that the cost of integrating wind is high are simply wrong.  But it does depend on the grid, the amount of load shedding, and instant on facilities that exist, and the accuracy of forecasting.  And there simply is no doubt that adding wind reduces the use of fossil fuels in any jurisdiction that makes electricity from fossil fuels.

Utility planners would be well advised to continue to refine their forecasting models (Ontario has this underway for demand), add wind forecasting to their models, and put a proper value on load shedding, instant on, or storage facilities.  Only then can an optimal mix from both a cost and environmental perspective be achieved.  And paying attention to these issues today allows maximum flexibility for future generation options.

Leave a Reply