Embedded Fault Diagnosis Expert System: Framework, Development Platform, Algorithms and Experiments

Abstract

The Internet of Things (IoT) connects all equipment on the industrial production site, realizing intelligent connection and freeing manpower. However, due to the large number of equipment and slow information transmission, it is difficult to transmit, store and diagnose the running status data of various equipment in real time. To address the above matters, a framework and a development platform for the embedded fault diagnosis expert system (FDES) are proposed to diagnose the running state of equipment in real time. Firstly, an inference engine framework is set up based on the C language integration product system (CLIPS) and probability neural network (PNN). Based on the framework, an object-oriented ontology system for fault inference is constructed, and an improved ontology traversing algorithm is proposed to increase the query efficiency of the inference process. Then, a fuzzy knowledge fusion algorithm is put forward to improve the clustering effectiveness of PNN and expand the industrial suitability of embedded FDES. Finally, the effectiveness of the proposed methods is verified by simulated cases and industrial instances.