Updated: Jul 10
The first two parts of this series on the integration of renewable energy with existing power grids covered typical supply-demand mismatches of the two “volume” providers of renewable power, namely wind and solar. These mismatches can occur both on a seasonal and a daily basis. The second part of the series then introduced the added complication of variability and uncertainty in the output of wind and solar plants, which make life much harder for the electricity grid operator, particularly as the proportion of renewable power on the grid increases, notionally above about 15% of the available capacity.
This third part of the series will look at commercial models of power purchase and how renewables do or don’t fit into existing power bidding models. The discussion here is based upon a hybrid of typical “market models” in open power regimes and is used to illustrate the typical development of thinking. It therefore does not reflect the detail of one particular market.
1st generation: traditional “dispatched” power markets (that is where a generator is called upon to produce or “dispatched” by the grid operator as and when required) are based upon the stacking of the various generation sources under an annual load duration curve of electricity demand. A hypothetical example is shown in the figure below and shows along the x axis, the 365 days in the year sorted by electricity demand with the highest demand days on the left (winter months for example in temperate climates) and the lowest demand days to the right (summer in the temperate zones). The resulting curve, shown in red in the figure, is called the Load Duration Curve (LDC). The grid operator spends a lot of time estimating the LDC for the coming year to obtain the best possible estimate of which generators will be called at which points in the year.
For a given day or hour, at the base of the curve will be the generator which has bid the lowest unit cost to the grid operator for a particular time period, followed by the next most expensive and so on until we reach the marginal generator, who just meets the last part of the demand for that period. The unit price bid by the marginal generator is then treated as the system clearing price for that period and is paid to all the generators equally.
The system operation takes into account many other factors, of course, such as ramp up and down times of each of the plants and other services provided such as frequency balancing.
The problem with this first generation approach is that the marginal generator, the one who sets the market price for that period, does not get any contribution to his capital cost as he is, by definition, bidding at close to his variable or operating cost. There is thus little incentive in the market for new entrants to deliver peaking capacity, as they will not make a return on their investment under this system. In an expanding market, there is therefore likely to be an erosion of available peaking capacity (as a percentage of the required capacity) over time, as electricity demand grows but little or no new peaking plant gets built.
2nd generation: to get around this problem of peak generators not recovering their capital costs via the market spot price, grid operators moved to a system of mixed capacity and commodity payments. Under this arrangement generators are paid a capacity payment for having generating capacity available at a defined notice period and then paid a further commodity payment for each kWh they actually produce. Provided the payments are set to allow the peak provider to receive a reasonable return on capital, this system has worked with conventional power sources.
3rd generation: high volume renewable generation plant generally has very low variable cost of operation. It can therefore bid into a 2nd generation system very competitively. The grid operator should treat such plants as "must run" sources and put them towards the bottom of the load curve. This causes two immediate problems for the grid operator:
the issues of intermittency discussed in briefs one and two of this series, mean that the grid operator now has an “unreliable” part of its base load power supply, which it will need supplementing frequently with other sources;
the introduction of highly flexible renewable power as base load (when it is available) pushes less flexible plant into the marginal generation position, which is the reverse of what is required to respond quickly and effectively to changes in demand. Natural gas generation plant will require around 30 minutes for a "warm start"; coal plants up to a couple of hours. This problem is particularly troublesome during low demand months, for example in the summer.
These two problems are exacerbated as the proportion of renewable generation on the system increases.
The final part of this series will look at new models which system operators are developing or trying out to get around these problems as renewable concentrations on a grid increase.