摘要
针对电炉的执行机构的调节总是滞后于电弧炉实际状态变化的情况,在专家系统的基础上增加弧炉神经网络预估模型,通过它预估出电弧炉下一时刻的状态,并经过特定的优化程序(ANN预估补偿程序)对专家系统的输出做出优化补偿。系统的实际运行证明:神经网络预估补偿的电弧炉自适应控制是一种可行的电弧炉电极控制方法;在电弧炉的点弧过程中,电弧的稳定性有较大改进,提高了功率因数,降低了电耗。
The regulation of executing agency always lags behind when the actual state of electric arc furnace changes. The electric arc furnace neural network predicting model, which can predict the electric arc furnace＇s state for the next time,is added based on expert system. The output of expert system is optimized compensated by special optimired program（ ANN Predicting Balance program）. The operating result of system shows that adaptive control of electric arc furnace based on ANN is a feasible way of electrode controlling. In arc ignition process ,it has apparent effect and improves the stability of arc, so that power factor is advanced and current drain is reduced.
出处
《化工自动化及仪表》
EI
CAS
2006年第1期70-72,共3页
Control and Instruments in Chemical Industry
关键词
神经网络
预估补偿
电极
权值
控制嚣
neural network
predicting compensation
electrode
factors
common controller
