摘要
针对目前铝电解的加料控制,提出了一种基于设定值优化的控制策略及智能加料控制方法。根据铝电解过程特性,采用了广义回归神经网络(GRNN)来辨识氧化铝浓度模型,并利用遗传算法优化寻找最佳光滑因子σ,当所辨识的实际氧化铝浓度模型的输出量误差最小化时,获得最佳的平滑因子。设计了模糊小脑模型神经网络(FCMAC)控制器,将氧化铝的浓度控制在理想区域内,提高了铝电解过程的控制性能。
As for the current feeding control of aluminum electrolysis,a control strategy based on set value optimization and intelligent feeding control method are proposed. According to the characteristics of the aluminum electrolysis process,the generalized regression neural network( GRNN) was used to identify the alumina concentration model,and the optimal smoothing factor σ was searched via the genetic algorithm optimization. When the output error of the identification model reached the minimum,then the optimal smoothing factor was obtained. A fuzzy cerebellar model neural network( FCMAC)controller was designed to control the alumina concentration within an ideal range and improve the control performance of the aluminum electrolysis process.
引文
[1]崔桂梅,杨海靳,刘丕亮,于凯.基于数据的铝电解槽氧化铝浓度预测[J].计算机仿真,2018,35(02):305-309.
[2]李界家,邓媛媛,郭宏伟,王梓翰.基于神经网络预测的铝电解模糊控制系统应用[J].沈阳建筑大学学报:自然科学版,2008(4):721-723.
[3]WU J,PENG D,LI Z,et al.Network Intrusion Detection Based on a General Regression Neural Network Optimized by an Improved Artificial Immune Algorithm[J].Plos One,2015,10(3):1-13.
[4]林九,喻小春,赵海兵,王玉杰.铝电解质分子比、Al2O3、CaF2的X射线荧光光谱法测定[J].广西轻工业,2011,27(07):17-19.
[5]朱跃华,李贺松.基于无线传输的铝电解槽极距分布监测系统的开发与应用[J].金属材料与冶金工程,2012,40(04):43-47.
[6]S.Y.WANG,C.L.TSENG,C.J.CHIU.Online Speed Controller Scheme Using Adaptive Supervisory TSK-fuzzy CMAC for Vector Controlled Induction Motor Drive[J].Asian Journal of Control,2015,17(2):569-581.