Adaptive Under Frequency Load Shedding Scheme Using Genetic Algorithm Based Artificial Neural Network

摘要

This paper presents two schemes of UFLS tokeep the system frequency within safe limits. GA-basedscheme is introduced as an offline method to get theproper amount of shed load achieving the minimumand steady state frequency within permissible limits.Due to the probability of generation variation andgenerating units outage during shedding process,ANN-based scheme is presented as an online methodto adjust the proper amount of load shedding at anyamount of power deficit. Multi scenarios ofcontingences are carried out on offline mode using GAoptimization technique to collect the training patternsfor ANN. The ANN-based scheme can consider thegeneration variations during the load sheddingprocess. Although using this scheme may shed moreloads, it maintains the frequency to be withinpermissible limits at various disturbance scenariosparticularly at the absence of secondary control. Ananalytic system frequency response (SFR) model withno secondary control incorporating UFLS scheme ispresented. The proposed method is compared with theclassical adaptive method to prove its effectiveness.Results are presented in the form of time domainsimulations via MATLAB/SIMULINK.