Comparison of three change detection algorithms for an electronics manufacturing process

Abstract

In a sequential manufacturing process, a product proceeds through different manufacturing stages. At these stages, sensors monitor the features of the product. In this paper, the information produced by the sensors is employed to detect abrupt changes in process variables. The developed algorithms contribute to an on-line application to a manufacturing system. A literature survey revealed the most common methods utilized in change detection. On-line applicability and transferability to new manufacturing lines are the most important features for real applications. During both on-line and off-line tests, some of the presented methods showed satisfactory results. Real-time, on-line manufacturing environment sets also its requirements for the applications. In the future, the possibility of combining expert knowledge with the aforementioned methods is the crucial point to study. The information thus received has usage in the preventive maintenance and quality control.