Application of Soft Sensor in Welding Seam Tracking Prediction Based on LSSVM and PSO with Compression Factor
作者单位：College of Information Science and Engineering Northeastern University
会议名称：《第25届中国控制与决策会议》
主办单位：IEEE;NE Univ;IEEE Ind Elect Chapter;IEEE Harbin Sect Control Syst Soc Chapter;Guizhou Univ;IEEE Control Syst Soc;Syst Engn Soc China;Chinese Assoc Artificial Intelligence;Chinese Assoc Automat;Tech Comm Control Theory;Chinese Assoc Aeronaut;Automat Control Soc;Chinese Assoc Syst Simulat;Simulat Methods & Modeling Soc;Intelligent Control & Management Soc
会议日期：2013年
学科分类：12[管理学] 080503[工学-材料加工工程] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0805[工学-材料科学与工程（可授工学、理学学位）] 0835[工学-软件工程] 0802[工学-机械工程] 0811[工学-控制科学与工程] 080201[工学-机械制造及其自动化] 0812[工学-计算机科学与技术（可授工学、理学学位）]
基 金：supported by Liaoning Key Industry Programme (2010220014)
关 键 词：Particle Swarm Optimization with Compression Factor (PSO-CF) Least Squares Support Vector Machines (LSSVM) Soft Sensor Welding Seam Tracking Prediction
摘 要：There are some problems with the strip steel heated in the annealing furnace such as the multiple correlations between temperature in the annealing furnace, steel strip tension, tension roll speed and other variables, noise in field data, which lead to the difficulty to predict the time of welding seam achieved air-knife. The least square support vector machine （LSSVM） inductance model optimized by the particle swarm optimization with compression factor （PSO-CF） algorithm is presented for the difficulty of the time prediction in this paper. The improved algorithm can improve PSO convergence accuracy, and effectively avoid falling into local optimum. It can be both the global fitting ability and local fitting ability of least squares support vector machine. The parameters of LSSVM model are optimized by improved PSO-CF algorithm to escape from the blindness of man-made choice. Using the algorithm in prediction of the arrival time and the position of welding seam, the numerical simulation results illustrate good generalization performance and prediction ability of the proposed method