Up to a small percentage of the total load on a grid, or on a network of grids tied together by adequate interconnectors, wind power can be integrated with the use of existing reserve capacity, i.e., with the redundant availability of extra generators kept running – burning fuel but not generating electricity – so that they can switch to production in case of the sudden failure of other generators. Thus wind power, despite being highly variable, intermittent, and nondispatchable – responding only to the wind, not to actual customer demand – can be taken into the grid, especially during periods of peak load, when there is a greater variety of sources providing power and the grid is therefore more flexible.
As the percentage of wind-generated power on the grid (called its “penetration”) increases, however, the ability of the grid to cope with its unpredictable fluctuations diminishes.
Unlike conventional generators, which are run continuously to provide base load or called on as needed to provide intermediate and peak loads, power from wind turbines cannot be dispatched on demand, and it often surges when there is no demand for it. To preserve stability of the electricity supply, grid operators often have to resort to shutting off the wind turbines.
Furthermore, if peak wind production by chance does coincide with peak demand, the transmission capacity of the grid may not be great enough to shift it and again, the wind turbines must be shut off.
Because of laws requiring the grid to accept all wind power, the wind companies may be paid for this curtailment as if they had nonetheless contributed power to the load.
Notably, curtailment occurs frequently in Great Britain (with payment), Ontario (also with payment) and Texas, both with substantial penetration of wind on an isolated grid. Curtailment has been documented also in northern Vermont.