基于改进模糊神经网络的输电线缆制造质量预测方法

Prediction Method of Transmission Cable Manufacturing Quality Based on Improved Fuzzy Neural Network

针对输电线缆制造质量预测问题,采用制造工艺关键指标数据,基于改进模糊神经网络对制造质量合格率进行预测。首先,通过引入模糊化层和模糊规则计算层建立模糊神经网络的结构,采用高斯函数建立模糊化层的隶属度函数,构建了隶属度函数均值和宽度的训练过程;接着基于帝国竞争算法对模糊神经网络中的隶属度函数学习因子和后件网络权重学习因子进行优化,优化目标为最小化神经网络收敛误差;采用改进模糊神经网络设计输电线缆制造质量预测流程。最后,通过一个仿真算例验证了所建立的模型相比于传统神经网络以及基本模糊神经网络在输电线缆制造质量合格率的预测效果上更优,验证了所建立模型的有效性。

Aiming at the problem of manufacturing quality prediction of transmission cable,the manufacturing quality qualification rate of transmission cable was predicted based on improved fuzzy neural network based on key index data of manufacturing process.Firstly,the structure of fuzzy neural network was established by introducing fuzziness layer and fuzzy rule calculation layer,membership function of fuzzification layer was established by Gaussian function,and the training process of mean and width of membership function was established.Then,the membership function sub factor and the weight learning factor of the consequent network were optimized based on the imperial competitive algorithm.The optimization goal was to minimize the convergence error of the neural network.The improved fuzzy neural network was used to design the manufacturing quality prediction process of transmission cable.Finally,a simulation example shows that compared with the traditional neural network and the basic fuzzy neural network,the average absolute deviation between the prediction of the qualified rate of transmission cable manufacturing quality and the actual inspection qualification rate is lower,which verifies the effectiveness of the established model.