Data Fusion of Noise and Temperature Field of Magnetic Bearings

Data fusion method of noise and temperature field is presented for the fault diagnosis of magnetic bearings. Noise distribution of magnetic bearings is measured with B & K instrument. From the experimental results, it can be known that two radial bearings produce the most large noise. Temperature distribution can be tested with an infrared imaging system. The results show that the component with the highest temperature is the motor, and the back end and the back radial bearing come second. The temperature is higher with the increasing of the rotation speed. It is proposed that noise and temperature distribution of every parts of a magnetic bearing be discretized to be three states respectively for data fusion. Data fusion results reveal that noise and temperature are relatively independent. A component producing large noise is uncertain to lead to high temperature rise, and vice versa. Noise is also related to temperature to some extent