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

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.