论文部分内容阅读
在铝包钢丝生产过程中,采用在线中频感应加热方式。基于中频感应加热条件,针对钢丝直径、钢丝运行速度、加热温度、中频频率、中频电压、中频电流、中频功率、加热炉长度之间的函数关系,建立了连续包覆过程的多变量神经网络控制方法。经过多次实验和筛选,选出有效数据,多个模式反复学习,直至网络全局误差函数E小于预先设定的一个极小值。通过网络训练和仿真分析得到较佳的解决方法,该方法应用在铝包钢丝连续包覆生产线上,较好地解决了钢丝加热控制难的问题。
In the aluminum clad steel wire production process, the use of online IF induction heating. Based on the medium frequency induction heating conditions, a multivariable neural network control of continuous coating process was established for the function of the diameter of steel wire, running speed of wire, heating temperature, intermediate frequency, intermediate frequency voltage, intermediate frequency current, intermediate frequency power and furnace length method. After many experiments and screening, valid data are selected, and multiple modes are repeatedly studied until the global error function E of the network is smaller than a preset minimum value. Through the network training and simulation analysis, a better solution is obtained. The method is applied to the aluminum clad steel wire continuous coating production line, which solves the problem of difficult wire heating control.