弧焊电源电压电流的自适应神经网络控制  

Voltage and Current Self-adptive Networks Control of Arc Welding Power

作　　者：高忠林[1] 胡绳荪[1] 殷凤良[1] 

机构地区：[1]天津大学,天津300072

出　　处：《电力电子技术》2007年第11期103-105,共3页Power Electronics

基　　金：天津市自然科学基金资助项目(05YFJMJC08900)~~

摘　　要：焊接过程是一个复杂、多参数耦合的高度非线性系统,在实际焊接过程中难以实现实时、有效的在线控制。根据焊接工艺要求,设计了弧焊电源输出电压电流波形。在常规PID控制的基础上,运用神经网络控制理论,建立了自适应神经元PID控制器,确定了自适应神经网络PID学习控制器的学习算法。建立了二氧化碳气体保护焊自适应神经元网络控制系统,并通过数字信号处理器TMS320F2407和单片机MSP430F149加以实现。通过常规PID控制与自适应神经元网络控制输出波形的对比,证明了其控制效果优于常规PID控制。Welding process is a strong nonlinear system of complexity and multi-parameter.h is difficult to realize real time and effective control.The voltage and current waveforms according to the demand of the welding techniques were designed using the neural networks control theory,the self-adaptive networks eontroller based on the PID control theory was set up.A CO： welding self-adaptive networks control system was established and DSP TMS320F2407 and MCU MSP430F149 were adopted as the central eontroller.The result shows the self-adaptive networks control has good control performance.

关 键 词：直流弧焊电源 逆变 自适应控制/神经网络 鲁棒性 

分 类 号：TG434[金属学及工艺—焊接]