Investigation on Fault Diagnosis System Based on Time Spatial Information Fusion Theory

This paper presents a novel developed information fusion framework for machine system pattern recognition and fault diagnosis. It is named as time spatial information fusion fault diagnosis system. Because it makes the best use of information from multisensor and the advantage of neural networks, majority voting and Dempster-Shafer algorithm for pattern recognition, the accuracy of machine fault diagnosis can be improved. Experimental data of a diesel engine combustion system is used to evaluate the effectiveness of this method on machine fault diagnosis. It is can be concluded that this promising method contributes to development of machine preventative maintenance