A rolling optimization dispatching method is proposed including the optimal selection strategy of units to be adjusted. According to ultra-short-term predictive information of load and wind power

the units to be adjusted are optimized and selected first by considering the matching between the prediction deviation of equivalent load(load and wind)and units' regulation capacity

the matching between the units' climbing amount and the fluctuation quantity of equivalent load

as well as the principle of reducing risks of the interface power off-limits. On this basis

with minimum unit adjustment cost and abandoned wind power as the objective

an optimizing mathematical model is developed and solved by the improved particle swarm optimization(PSO)algorithm. A practical example shows that the rolling optimization dispatching method considering the optimal selection strategy of units can realize automatic tracking and regulation according to the prediction deviation of load and wind power on an ultra-short-term time scale

while improving the accuracy of generation schedule and reducing the comprehensive generating costs and the risks of the interface power getting off-limits to enhance the economy and security of power grid.