Research on Composite Impact Fault Identification Algorithm Based on Improved MLP and Single Sensor Data

In the actual complex industrial environment, as an important part of mechanical equipment gear box in variable speed, variable load operation environment after a long time of work may appear compound failure -- gear, bearing failure. However, due to the influence of mechanical equipment structure design and other factors, it is usually not possible to install sensors near every important component to perceive its health status, and only a few vibration sensors can be installed in a few locations. Therefore, it is very important to identify and diagnose faults of multiple components through sensor data of a few locations. In this paper, several improved MLP neural network models are proposed to diagnose and classify gear and bearing faults using single sensor data, and the model is verified and analyzed by independent experimental data sets.