When the scenario analysis method is used to describe the uncertainty of renewable energy

it can not only characterize the probability distribution of random variables(i. e. predicted wind/solar power)in time-power space

but also can further reflect the temporal correlation of these random fluctuations. The scenario dimension disaster is a common problem in scenario analysis. As it comes to reserve optimization

the combination explosion of control measures is also encountered

which exacerbates the difficulty of the solution. Scenario reduction is beneficial to the practical application of scenario analysis. However

traditional scenario reduction methods cannot ensure that low-probability and high-risk scenarios would not be cut off

and thus lead to risk leakage. A multi-scenario reserve optimization method of gradually filtering scenarios is proposed considering the residual risk of scene sets

which integrates the scenario reduction and optimization process. Compared with the traditional “scenario reduction first

and optimization second” approach

the proposed scheme can adaptively select the scenarios to be optimized

and can effectively identify low-probability and high-risk scenarios. Simulation results indicate that the proposed multi-scenario optimization method has significant advantages in balancing optimization effect and computational efficiency compared with the mixed-integer linear programming optimization method based on the traditional scenario reduction.