会议专题

An improved feed-forward neural network based on unscented kalman filter and strong tracking filter to establish dynamic evolutionary modeling for aluminum electrolysis process

  This paper presents a modeling method for dynamic process of aluminum electrolysis based on a new neural network.The proposed Neural Network(NN)is based on the theories of Unscented Kalman Filter(UKF)and Strong Tracking Filter(STF),which is shortened as ST-UKFNN in this study.Moreover,the new training algorithm and robustness analysis of the ST-UKFNN are presented.The final section of the paper shows an illustrative example regarding the application of the proposed method to estimate the technical energy consumption of the aluminum electrolysis process,compared with the modeling methods of Back-Propagation Neural Network(BPNN),Extended Kalman Filter Neural Network(EKFNN)and Unscented Kalman Filter Neural Network(UKFNN).The analysis and results show that the method improves the real time tracking ability of dynamic interference in aluminum electrolysis process,and the accuracy of ST-UKFNN is better than the other three modeling methods.

Aluminium electrolysis Feed-forward Neural Network (FNN) Strong Tracking Filter (STF) Unscented Kalman Filter (UKF) Energy consumption model Dynamic interference

Lizhong Yao Taifu Li Hengjian Zhang Jun Yi Yingying Su Xinghua Liu

Department of Electrical and Information Engineering,Chongqing University of Science and Technology, Department of Electrical and Information Engineering,Chongqing University of Science and Technology, Department of Electrical and Information Engineering,Chongqing University of Science and Technology,

国内会议

2016第七届中国石油化工重大工程仪表控制技术高峰论坛

银川

英文

1-13

2016-05-31(万方平台首次上网日期,不代表论文的发表时间)