新型模糊神经网络控制器在GTAW焊中的应用

Application of New Fuzzy Neural Controller in GTAW

钨极气体保护电弧焊(GTAW)过程是一个复杂的、多参数耦合的高度的非线性系统,难以实现实时、有效的在线控制 模糊控制吸收人的经验思维的特点;神经网络则对信息的处理具有自组织、自学习的特点;遗传算法是一种全局优化搜索方法,具有简单通用、适合并行处理的特点 笔者将三者有机地结合起来,在模糊神经网络控制器的基础上利用改进的遗传算法,并分析其网络结构和离线学习的方法,协调利用三者的优势设计一种新型的模糊控制器,并使之用于脉冲GTAW焊仿真中。

Gas Tungsten Arc Welding(GTAW) is a strong nonlinear system of complexity and multiparameters. It is difficult to realize real time and effective control. Fuzzy control has characteristics of human experiential thought. Neural Network has characteristics of selfcontrol and selfstudy for information disposal. Genetic algorithm is a new method of global optimal searching with characters of simpleness, parallel disposal and mass application. By combining the above three together, and based on fuzzy neural network, transformed genetic algorithm is applied to the design of a new fuzzy controller. The result of simulation in impulse GTAW indicates that the new fuzzy neural controller is better than traditional fuzzy controller.