多参数点焊质量的偏最小二乘综合评价方法

Partial least square approach for multi-parameter assessment of resistance spot welding quality

将一种新型的统计回归方法——偏最小二乘(PLS,partial least square)方法用于点焊质量的多参数监测与分析中。借助于提取主元的思想,利用PLS方法对于点焊参数和焊接质量之间的相关信息进行筛选和综合,讨论和分析了点焊参数的各种数据信息。在此基础上,建立了点焊质量多参数综合评价模型。结果表明,该方法可以有效解决点焊质量综合评价模型中点焊参数存在的高维及多重相关性问题,建立的综合评价模型具有良好的评估能力。

A new approach based on partial least square(PLS) was used in the multi-parameter monitoring and analysis of the resistance spot welding(RSW) quality.Based on the idea of extracting principal components,correlation information between the monitored RSW parameters and welding quality were screened and synthesized under the condition that the multi-parameters monitored exist multicollinearity.Moreover,various data message of the RSW quality process parameters were also shown and discussed.In addition,an integrated assessment model was built based on multi-parameter monitored.The experimental results indicate that the presented method can select and extract PLS components of RSW quality process parameters from sample data,and the problems of high dimension and multicollinearity are solved effectively in regression model.The integrated evaluation model has excellent estimation ability.