A study of cause-effect structure acquisition for anomaly diagnosis in discrete manufacturing processes

In a factory setting, the zero-breakdown concept of manufacturing processes has been eagerly anticipated. In this article, the notion of machine health management technology is introduced in response to this requirement. The technology employs a cyclical and evolutional problem-solving approach using cause-effect structures. This paper will utilize an automatic acquisition method by using sensing data of machine behaviors. The problem to be solved here is to diagnose an anomaly using tact time delays of manufacturing lines. The cause-effect structure will illustrate the relationship between tact time delay anomalies and machine behaviors that lead to the anomalous conditions. The cause of the tact time delay can be identified by applying the cause-effect structure acquired through adopting a novel statistical processing approach using I/O signals of programmable logic controllers (PLCs). Experiments were conducted using an experimental manufacturing line to validate the effectiveness of the proposed method.