摘要
交流电弧炉电极控制系统是一个多变量、非线性、参数时变、复杂强耦合系统,传统方法很难建立其数学模型。为此从电极控制的实际应用出发,提出了一种变结构遗传Elman网络预测建模方法,其中改进的混合遗传算法用来对网络结构和权值及自反馈增益的同步动态寻优。并将基于BP算法的改进Elman网络和本文提出的变结构遗传Elman网络都应用于交流电弧炉的电极模型建模中。通过基于安钢现场数据的计算机仿真实验表明:变结构遗传Elman网络克服了因复杂对象造成的网络结构复杂问题和采用BP算法带来的权值训练缺陷;具有更好的动态性能,逼近速度快,精度更高等优点。
Electrode control system of alternating current electric arc furnace is a nonlinear, parameter-time-varying, strong coupled system, and traditional method can hardly build up its mathematics model. From the application of electrode control, a variable structure Elman neural network prediction model based on a new hybrid generic algorithm is proposed in this paper for better efficiency. This learning algorithm which can simultaneously evolve the network structure, the weights and self-feedback gain coefficient based on improved hybrid generic algorithm. The improved Elman based on BP algorithm and the variable structure Elman neural network proposed in this paper are both applied in identification of electrode model. The simulation based on the spot real data of Anyang Steel indicates that the variable structure Elman neural network overcomes the problem of complex network structure, which is brought by the complexity of electrode control system and limitation of weights by BP algorithm. The proposed method based on a new hybrid generic algorithm has better dynamic characteristic, quicker approach speed, and better precision.
出处
《电工技术学报》
EI
CSCD
北大核心
2007年第12期175-179,共5页
Transactions of China Electrotechnical Society
基金
北京市教育委员会重点学科共建项目(XK100080537)
关键词
电极
混合遗传算法
ELMAN神经网络
变结构
Electrode, hybrid generic algorithm, Elman neural network, variable structure
