基于主成分分析的光纤网络激光切口粗糙度预测研究

Research on laser cutting roughness prediction based on principal component analysis

当前光纤网络激光切口粗糙度预测方法存在预测精确度低,且预测过程复杂的问题,提出一种基于主成分分析的光纤网络激光切口粗糙度预测方法,通过主分量分析的方法,根据光纤网络激光切口粗糙度数据的信噪比,实现光纤网络激光切口粗糙度数据的处理分析,利用处理后粗糙度数据不相关变量之间的正交关系,确定粗糙度数据的平均值,构建协方差矩阵,确定主成分分析的主分数值。并根据处理后数据之间的相干关系,对预测模型的参数进行计算,构建光纤网络激光切口粗糙度预测模型,实现激光切口粗糙度预测。实验结果表明,本文方法进行光纤网络激光切口粗糙度预测,预测过程较为简单,且预测结果与实际结果较为接近。

In the current fiber network laser notch roughness prediction method,the prediction accuracy is low and the prediction process is complex.A method based on principal component analysis for fiber laser roughness prediction is proposed.The signal-to-noise ratios of the network laser slit roughness data is used to process and analyze the laser slit roughness data of the fiber network.The orthogonal relationship between the uncorrelated variables of the processed roughness data is used to determine the average value of the roughness data,construct the covariance matrix,and determine the principal score of the main components analysis.According to the coherent relationship between the processed data,the parameters of the prediction model are calculated,and the laser cutting roughness prediction model of the fiber network is constructed to predict the roughness of the result.The experimental results show that use the proposed method to predict the laser cutting roughness of the fiber network,the prediction process is relatively simple,and the prediction results are close to the actual results.