Knowledge-Driven Analysis Framework of Anomaly Propagation in Manufacturing Workshop

With the improvement of customization demand and the increase of market fluctuation, manufacturing enterprises gradually turn to intelligent manufacturing mode. However, there are various abnormal events in the manufacturing process, which hinder the normal production of the workshop. The analysis of anomaly propagation is considered to be an effective technique to solve this problem. In this paper, we propose a knowledge driven analysis framework of anomaly propagation in workshop. Firstly, a temporal knowledge graph(TKG) is built for anomaly propagation analysis based on the open platform communications unified architecture (OPC UA) information model and the ontology web language (OWL) mapping model. The time-series data in the manufacturing process is contained in the TKG, such as workpiece process data, machine maintenance data, etc. These data are the basis for the analysis of anomaly propagation. Then, a time-series analysis method for anomaly is proposed. This method gives the breadth and depth of abnormal events propagation in a period of time, so that decision makers can make decisions in advance to precisely control the production process. The analysis framework provides a route for manufacturing workshop to resolve abnormal events.