摘要
The welding process essentially is a complicated nonlinear system with time-varying, uncertain, strong-coupling characteristics, so it is difficult to get high welding quality by traditional control approaches such as the standard proportionalintegral （ PI） algorithm. A new algorithm based on artificial neural network （ANN） is presented to achieve optimal P1 parameters and improve its adaptability. First, main parameters of artificial neural network are researched to improve the convergence rate and system stability. Then, six expert rules are proposed to constitute the expert adaptive ANN-PI algorithm. Experimental results show that the welding current control system＇has high dynamic response rate, and the welding process is stable.
The welding process essentially is a complicated nonlinear system with time-varying, uncertain, strong-coupling characteristics, so it is difficult to get high welding quality by traditional control approaches such as the standard proportionalintegral （ PI） algorithm. A new algorithm based on artificial neural network （ANN） is presented to achieve optimal P1 parameters and improve its adaptability. First, main parameters of artificial neural network are researched to improve the convergence rate and system stability. Then, six expert rules are proposed to constitute the expert adaptive ANN-PI algorithm. Experimental results show that the welding current control system＇has high dynamic response rate, and the welding process is stable.
基金
This work is supported by the National Natural Science Foundation of China （No. 51207083）.
