Research on fault diagnosis health monitoring algorithm in electromechanical control

A new method of fault diagnosis is proposed, which combines deep learning with timing analysis, to improve the precision and real time performance of mechanical and electrical control system. BP neural network is able to extract the latent characteristic from the fault data efficiently by combining the timing analysis method. A typical electrical and mechanical equipment fault data set was used in the experiment. Simulation results indicate that this algorithm has better performance than the conventional model and signal processing methods in the aspects of fault diagnosis precision, false alarm rate and missed alarm rate. In particular, the accuracy of the fault diagnosis is 95.6%, the error rate is 3.2%, and the failure rate is 2.1%. Moreover, the proposed algorithm is robust in processing time series data and has a good future. The results show that the proposed algorithm not only has good performance, but also has good real time performance, so it can be used for on-line monitoring of large electrical and mechanical systems.