Dynamic responses of area frequency and tie-line power under disturbance are supposed to be suppressed by automatic generation control. Because of the large scale integration of wind power and the inaccuracy of describing wind power randomness in day-ahead scheduling, an intraday automatic generation control parameters optimization considering the wind power forecast error is required. Therefore, a look-ahead parameters update framework both considering AGC controller parameters and participation factors of each type unit is proposed. In this framework, hundreds of wind power forecast error scenarios are first clustered and sampled. Subsequently, the optimized parameters are obtained under the initial scenarios selection set and adjusted according to the check results under the rest scenarios. In order to match the computation requirement of the proposed framework, a novel group-based particle swarm optimization-pattern search algorithm with high calculation speed and high-quality solution is developed. The quality of solution is guaranteed by the combination of the global searching ability of particle swarm optimization and local search ability of pattern search. Meanwhile, the calculation speed is guaranteed by cluster method and reasonable iteration stop condition. The proposed framework and algorithm are verified in a modified multi-area and multi-source power system. The simulations results show that this on-line optimization method provides better AGC response with less deviations of system frequency and tie-line flow. Moreover, the novel group-based particle swarm optimization - pattern search algorithm proposed shows promising calculation speed in look-ahead time scale.
A Novel Group-based Particle Swarm Optimization-Pattern Search Algorithm for AGC Parameters Optimization with Wind Power Integration
20.10.2021
2129397 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
An Adaptive Particle Swarm Optimization Algorithm with Local Search
British Library Online Contents | 2015
|Particle Swarm Optimization Algorithm
British Library Online Contents | 2005
|