In the process of distribution network load frequency setting,the PI parameters of the controller are affected by local control iterations,forming a time difference between the control layer and the feedback layer,resulting in frequency mileage de-viation.Therefore,a PI parameter setting method for distribution network load frequency control based on LSTM and transfer learning is proposed.The LSTM model is used to quantify the frequency modulation(FM)mileage parameters,and the related parameters of load frequency control are differentially updated with the transfer learning.The LSTM model of distribution net-work load frequency is constructed,and the transfer learning optimization of frequency control,PI parameter control con-straints and PI parameter setting output are carried out.The experimental results show that it has good performance both in control accuracy and control stability,and the setting results meet the expectations.
关键词
LSTM/迁移学习/负荷频率控制/PI参数整定
Key words
LSTM/transfer learning/load frequency control/PI parameter setting
维普
万方数据