高级检索

基于变论域电阻点焊模糊神经网络控制方法

RSW fuzzy artificial neural network control method based on variable universe

  • 摘要: 为提高电阻点焊的控制精度和焊接质量,根据电阻点焊过程的特点和要求,通过集成变论域、人工神经网络和模糊控制技术,提出了基于变论域电阻点焊模糊人工神经网络控制方案,开发了四层模糊神经网络结构,分析了计算过程,推导了四层模糊神经网络各层的计算方法和计算公式,研究了输入输出变论域伸缩因子的确定方法,定义了输入变量的7个模糊子集和输出变量的13个模糊子集,确定了49条模糊控制规则,研究开发了一种电阻点焊变论域模糊人工神经网络控制器,结合实际产品的设计开发进行了试验研究与分析,证明了变论域电阻点焊模糊神经网络控制方法的优越性.

     

    Abstract: In order to enhance the control precision and welding quality of resistance spot welding(RSW),according to the RSW control characteristics,it put forward the RSW fuzzy artificial neural network(FANN) control method based on variable universe by integrating the variable universe technology and FANN.The network structure of FANN was designed and the process of neural computation dealing with the fuzzy,the fuzzy rules and the solution of fuzzy was analyzed.And the method of how to fix on the extension-contraction factors of input and output universe was researched.Then a kind of FANN controller of RSW with the variable universe was developed,being applied to a kind of spot welding experiments to study and analyze.The results prove the FANN control method is better than other fuzzy control methods.

     

/

返回文章
返回