Product design change propagation analysis in a manufacturing environment with machine learning

Product design changes are implemented to remain competitive in the market. Product change propagates in design and manufacturing system. Completing product change prediction analysis well in advance is the key in successful engineering change management. Manufacturing companies are facing challenges to control changes in a process during implementation phase. Process parameters are required to be reset against product changes with multiple production trials. This delay in change implementation phase increases obsolescence cost along with high chances of product quality failures. Our research objective is to reduce overall time required to implement product design change in manufacturing system. It is proposed to extend the concept of design structure matrix from design phase to production environment through a machine learning tool. A case study is conducted on the riveting process parameters in an automotive company. Riveting process is decomposed with process parameters by use of design structure matrix. Machine learning tool is used to extract relationship between process parameter and part features with scientific method. Prediction results of the case study from machine learning tool are evaluated with actual design of experiments trials. Prediction results are found very close to results from DOE trials. The results from case study supports to research objective by early prediction of impact on process parameter.