Wind and solar energy sources vary rather unpredictably in output during the day, from 0% to 100%. Yet they are on a path to eventually constitute a large portion of the of the world’ energy portfolio.
The power grid manager needs to match at all times its customers’ demand with its energy generation. To do so, she must match her grid’s variable power sources with a power source that she can easily turn on or off (when her variable power source goes down). Right now, this power source is typically a gas-powered plant (it’s really a bit more complicated).
How does that impact the true system cost of solar or wind?
In 2014, in most countries of the world (except Germany, which generated 25.8% of its electricity from renewable sources in 2012, and had a peak of 59% instantaneous renewable energy use in late 2013) these on/off gas-powered plants are already a part of the grid. So, when solar and wind remain a small part of the grid, there is no hidden cost.
But, as soon as wind and solar become a significant portion of the grid, using traditional grid management practices the construction of new wind and solar plants would need to be accompanied by that of gas-powered plants, so that, when solar or wind generation goes down while demand remains high, gas-powered plants can match energy supply with demand.
How does this impact the cost of wind and solar?
Let’s take the example of a 350MW solar plant. According to the EIA, the capital cost is about $4/W, i.e. $1.4B. But, to add a true capacity of 350MW to the grid, you also need to add a gas-powered plant of 350MW as well, that you will turn on when solar is down. Building such a plant with carbon capture and sequestration, to minimize emissions, costs about $2/W, so another $0.7B. The soalr plant costs $25/kW.yr to operate, so the operation costs are doubled. Finally, if we assume that the solar plant will produce power 35% of the time, (very optimistic), the gas plant will need to burn gas 65% of the time, with a fuel/ ops/ maintenance cost of about $35/MWh (source EIA2).
Let’s assume that both plants last 30 years. Over 30 years, the cost of the solar plant alone would have been (very simplified, no actualization):
Csolar = $1.4B + 30*25*350,000 = $1.7B in 2012$
However, the additional cost of the gas plant over the same duration is (also very simplified):
Cgas = $0.7B + 0.65*30*35*350*24*365 = $0.7B (capital) + $2.1B ( ops and fuel) = $2.8B in 2012$
So, in the end, when we want to add 350MW to the grid with solar, our real cost (since we also need to add gas generation for the times when there is no solar) would be $1.7B + $2.8B = $4.5B, which is over 235% of our life cycle cost for solar plant alone.
What this means is that, with optimistic estimates on the availability of solar power generation, the variability of the resource has cost us 135% more than the total life cycle cost of the solar plant!
In that case, why not replace the adjunct power plant by an energy storage system?
Note: it is important to understand that the present model is simplified. This is how:
- cost analysis neglects second order costs and actualization
- the actual practice of generation pairing is much more complex, and can use, in some cases, other assets than gas-powered generation
- it is possible to reduce, to some degree, variability in wind and power by spreading generation capacity over a large area.
Yet the analysis, as a whole, is true to reality.