基于灰色神经网络的绕丝参数回归分析设计

Regression Analysis Design of Wire Winding Parameters Based on Grey Neural Network

现阶段工业生产中绕丝加工存在一定的报废率,对原材料造成一定的浪费,尤其在贵重金属加工行业。以铂丝绕丝加工为例,首先对绕丝工艺中的5个关键参数,即拉力强度、延展值、断裂负载、灌注孔径、阻力进行回归分析,证明合格率与各参数之间不存在线性关系;其次利用灰色神经网络对来自工厂的加工数据进行预测分析,找出绕丝生产过程中的最大影响参数,最后使用MATLAB软件进行编程并进行计算和仿真之后,对产品进行纳米压痕测试对比,经过优化后所生产出来的产品,硬度高且分布均匀。

At present,there is a certain scrap rate in wire winding processing of industrial production,which causes a certain waste of raw materials,especially in the precious metal processing industry.This study takes platinum wire winding processing as an example.Firstly,five key parameters including tensile value,elongation value,breaking load,perfusion diameter and resistance in the wire winding process are analyzed by regression method.The results show that there is no linearity between the pass rate and each parameter.Secondly,the grey neural network is used to predict and analyze the processing data from the factory to find out the maximum influencing parameters in the wire winding process.Using MATLAB for calculation and simulation,the products after optimization have the higher hardness and equal distribution as the results from nano-indentation experiment.