Research on Fault Diagnosis Technology of Slewing Platform Bearing Based on Multi-Channel Fused CNN

The Bearing with low speed and heavy-duty is a core component of the large rotary machinery equipment. It is very important for improving the safe service life of equipment to detect and diagnose faults in a timely manner, avoid under- and over-maintenance, and reduce financial losses. We construct fault simulation test bench for low-speed and heavy-duty bearing with six types of bearing faults. The location of the installation of the sensors was determined using finite element analysis software, and fault signals were collected at seven measurement points. By comparing the data input of single measurement point and multiple measurement points, as well as the different input modes of vibration signal and spectrum signal, it is finally determined to use the multi-channel fused convolutional neural network model with multi-point spectrum signal input, and the accuracy of fault diagnosis reaches 98.64%.