Research on Information Fusion Model Used in Complex Mechanical System State Monitoring

According to the complex characters of mechanical system such as structure complexity, un-stability failure, multiple and coupling characteristic parameters, a new state monitoring method based on information theory and data fusion principle is proposed here. The main work is to present an information measuring guideline called state information entropy to describe the uncertainty of the system state, then to design a information fusion model for the synthetical analysis of the multiple characteristic parameters to get the decision-making of the state entropy, and to realize the state monitoring of the system. The fusion model is based on a integrated neural network so that it has the ability to learn the running mechanism and experience data of the system and also to realize the associate analysis and layered fusion of the multi-source signals. Efficiency of the method is shown by the theoretical analysis and simulation experiment.