柿竹园公司生产实现开门红.pdfVIP

64 第 32 卷 肖本贤，王小伟，朱志国，等 基于改进 算法的过热气温神 ， ， （）： [8] . PSO Applications 2008 25 3 569-573. 经网络预测控制 控制理论与应用， ， （）： [9] 卓金武，魏永生，秦 健，等 在数学建模中的应用 [J]. 2008 25 3 569-573. .MATLAB [M]. XIAO Benxian ，WANG Xiaowei ，ZHU Zhiguo ，et al.Neural network 北京：北京航空航天大学出版社，2013 ：64-65. predictive control for superheated steam temperature based on [10] 史 峰，王小川，郁 磊，等.MATLAB 神经网络 30 个案例分析 modified partical swarm optimization [J].Control Theory [M].北京：北京航空航天大学出版社，2011:241-242. Leａkａｇe Ｔｉme Preｄｉcｔｉｏｎ ｏf Ｔｕｎｇｓｔeｎ Ｉｏｎ Excｈａｎｇe Bａｓeｄ ｏｎ PSO-ｏpｔｉmｉzeｄ LS-SＶM LIU Feifei, ZHU Pengke, WU Yijun (School of Electrical Engineering and Automation, Jiangxi University of Science and Technology, Ganzhou 341000, Jiangxi, China) Abｓｔrａcｔ:To solve the problems existed in tungsten ion exchanges leakage time, such as frequent detection, detection lag, large relative error and high labor intensity, this paper established a leakage time LS-SVM prediction model using PH value, Baume degree, conductivity and flow rate of the pre-delivery solution as inputs and leakage time as output. Kernel parameter and error penalty parameter are selected by particle swarm optimization (PSO). The simulation results show that the model can effectively predict the ion exchange leakage time with high precision. The maximum absolute error is 8 min, and the maximum relative error is 1.5385 % . It can meet the requirements of tungsten ion exchange process by replacing the manual detection. Keｙ wｏrｄｓ:tungsten ion exchange; leakage time; LS-SVM; PSO !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!