Wind adds small amounts to the overall cost of managing the variations in supply and demand in three ways.
First, more spinning reserve is required - "part loaded" generation capacity from any technology that can ramp up its output to meet increasing demand, or reduce production as demand falls. Spinning reserve adds to the overall cost because power stations operating below full output are less efficient and must spread their capital repayments over the sale of fewer kilowatt hours, increasing their unit cost of electricity.
Second, as the proportion of wind on a power system increases, it displaces generation from existing plants, which will consequently operate at a lower load factor. The reduced output means reduced revenues for meeting capital repayments, slightly increasing cost.
Third, at times of low demand and high winds, such as the rare occurrence of a windy summer period in northern Europe, wind power that is surplus to requirements becomes a cost rather than an asset, unless there are customers to pay for that generation.
To a greater or lesser extent, the principles that apply to wind apply to other generation too. To include in the cost of wind generation its specific contribution to the cost of maintaining security of supply without applying the same methodology to its competitors does not provide a fair basis for comparing the technologies' costs.
For wind, the modest cost of supplying reserves is hotly debated and often included in models of its generation cost. Yet, the specific cost of supplying reserves for nuclear are seldom, if ever, accounted for in its generation cost. For a fair comparison, such "system" costs, along with the economic benefits associated with all technologies, would need to be painstakingly allocated to each generation type - a complex exercise that would add costs for the consumer with no improvement of service or reliability.
A more straightforward and cheaper approach is to establish the overall cost of maintaining a low probability of supply failure and spread it across the entire generation base. And that is actually the model used by power system operators today. Theoretical models that apportion the specific cost of wind's share of reserve to the cost of its generation make wind look comparatively more expensive than it is.
For a 60GW power system with a mix of thermal generation, such as that of Britain, but with wind supplying 20% of electricity, the added cost to the consumer of maintaining a reliable supply amounts to about £2/MWh (EUR 2.3/MWh), or £0.002/kWh. That extra cost rises as the proportion of wind increases.
Cost versus savings
Britain has about 12% renewables supply today. If wind supplied 40% of electricity, the cost of maintaining reliability of supply across the entire generation base would be in the region of £5/MWh, or £0.005 per kilowatt hour. On a domestic bill that amounts to a 4% increase and 6% for a large industrial user. Savings associated with wind displacing other generation could more than compensate for that increase.
Intuitively, it might be presumed that provision of extra spinning reserve would always be more costly than displacement of thermal generation by wind, which requires more reduced output, or curtailment of wind production. That is not the case. Breaking down the £5/MWh extra cost for securing supply with 40% wind on a system similar to that of Britain reveals the increments on the final bill are around £2/MWh for provision of spinning reserve, compared with £2.4/MWh for running thermal plant in part load, and £0.6/MWh for taking wind off the system that is extraneous to requirements. For 20% wind penetration the corresponding extra cost to the consumer is £0.9/MWh for spinning reserve, £0.5/MWh for part loaded capacity, and £0.2/MWh for curtailment of wind generation, on average.
As the relatively low cost implies, required increases in spinning reserve when adding wind are small. They are much less than the actual capacity of the wind farm to deliver electricity. The size of the total reserve for the entire power system is governed by the need to make supply failure extremely unlikely, with or without a contribution from wind and taking account of the uncertainty inherent in predicting demand.
For a power system like Britain's, the hour-ahead uncertainty in predicting demand is about 400MW, or a bit more than 1% on average. With the addition of 30GW of wind, the level of uncertainty increases to about 900MW, or about 2.4%.
Given the size of the system, the quantity of extra reserve to cover for that rise in uncertainty is relatively small, demonstrated by the modest cost, as above.
WHAT IT COSTS THE CONSUMER SECURING A HIGH PROBABILITY OF SUPPLY WITH WIND ON THE SYSTEM
Cost of spinning reserve
A thermal power system similar to that of the UK, but with 57GW of thermal capacity and 30GW of wind, requires 4-6GW of short-term reserve - about double that of a 60GW all-thermal system, to maintain security of supply.
That extra reserve is likely to be a mixture of demand management (such as temporary reduction of electricity to appliances not in need of non-stop power, such as freezers) and part-loaded thermal plants. For 20% wind, the added cost of these measures, when allocated to each unit of wind produced, is £3-6/MWh (depending on variables such as the life of the plant and the mix of the plant margin).
As a system cost spread over the entire consumer base - which is how reserve power and plant margin is normally accounted for rather than being allocated to specific generation - the cost to the consumer lies in a range of £0.6 to £1.2/MWh, or between around £0.005 and £0.01/kWh.
Cost of operating a thermal plant at reduced output
The addition of wind power to a system can only displace some of the thermal capacity, or security of supply would be compromised. Thermal generation displaced by wind forces the plant to operate at reduced output, so it must spread its capital repayments cost over the sale of fewer kilowatt hours, pushing up its cost of generation.
For a combined-cycle gas unit with a capital cost of £700/kW, that additional cost of generation in a system with 20% wind energy effectively adds £2.5/MWh if added to the cost of generating the wind kilowatt hours, or £6/MWh with 40% wind. The addition of any sort of generation, however, has a similar effect and the cost of running any plant at reduced output is more usually treated as a system cost paid by all consumers.
The cost to the consumer of part-loaded generation attributable to 20% wind supply starts at about £0.5/MWh, rising to £2.4/MWh for 40% wind penetration.
Cost of excess wind
Once wind production is displacing 25% of thermal generation, it may on occasion exceed system demand. If no market is available to take the excess, generation has to be curtailed, increasing the level of debt repayment.
A cost is associated with wind curtailment: with wind providing 30% of electricity on a system similar to that of the UK, the cost spread across the wind generation is £0.6/MWh, and spread across all consumers it is £0.2/MWh. With 40% wind the corresponding cost is £1.5/MWh and £0.6/MWh to the consumer.
Interestingly, Denmark's energy plan to 2050, which is for a combined heat and electricity system with 80-90% variable supply from renewables - mostly wind - foresees an excess of supply for just 5% of the time over a year, thanks to a string of initiatives to mitigate the problem (few of which are "storage" in its generally understood meaning.)
With 50% wind by 2020, excess wind is not likely to be a problem, given initiatives already in the works.