Improved Production Performance Through Manufacturing System Learning

With the advancement of computing technologies in the manufacturing domain, more information is available on factory intranets. This paper introduces a framework to optimize the performance of a production line through multiple plant comparisons and learning among identical or similar production lines by leveraging the information stored on the factory intranet. In this work, production data from multiple identical production lines are collected and analyzed. A fishbone diagram is introduced to help find differences amongst plants. By taking advantage of the abundant information from multiple plants, the “best” feasible action can be learned on critical machines which offers a new way to optimize the product line in addition to root cause analysis. To predict improvements, machine learning is used including preprocessing, model selection and validation. Consequently, a cost-benefit evaluation is provided to help decision making. A case study is performed based on an automotive industry scenario where the method is demonstrated and an increase in throughput is predicted.