A survey of knowledge-based intelligent fault diagnosis techniques

Abstract

With the development of information technologies, more and more real-time data can be obtained from production and operation process. Thus, how to extract effective information from these massive data, so as to carry out in-depth statistics and mining of faults, and gradually explore the faults laws and causes are crucial for intelligent factories. In recent years, a variety of statistical learning and data analysis methods have been used in fault diagnosis. Due to the complex structure, multi-source failure and suddenness of the industrial production system, the combination of empirical knowledge and mechanism principles can solve various fault problems. This paper summarizes several commonly used fault diagnosis methods, and focuses on knowledge-based intelligent fault diagnosis, including first-order logic knowledge representation method, production knowledge representation method, framework knowledge representation method, object-oriented knowledge representation method and Semantic-based knowledge representation methods.