A novel sensor was designed to realize real-time monitoring of sliding bearing wear in magnetic drive pumps. Based on the structure of magnetic drive pumps, a special sensor structure in which the curved surface measurement was converted into a plane measurement with a detecting wheel was put forward. The detecting wheel made of an anti-corrosion treated magnetic material was fixed on the inner magnetic rotor, and the detecting sensor was installed in the ring-shaped recessed groove in the containment shell. The sensor was a container with a flat detecting plane, which was opposite to the detecting wheel in the axial direction. The sensor was composed of magnet, magnetic flux concentrator and its forehead as well as the Hall IC. The detecting sensor and the detecting wheel form a closed magnetic circuit. Through magnetic circuit analysis, the magnetic circuit was simplified. After the sensor size was decided, all the parameters of the magnetic circuit such as the magnetic permeability of the magnet, the magnetomotive force of the magnet, and the cross-sectional area of the magnetic circuit were constant, thus the output voltage of the Hall IC was related to the change in the area of the magnetic flux concentrator forehead, So, it was associated with radial displacement δ of the detecting wheel. The resolution was 0.122 mm, the linearity was 0.1-2.3, the sensitivity of the sensor was 10mV/mm and the linearity of the displacement characteristic curve of the sensor was improved. The sensor designed in this paper has certain application significance for real-time monitoring of sliding bearing wear in magnetic drive pumps.
YUAN Tiejun, ZENG Pei,KONG Fanyu. The applica-tion of multiple neural network model to the monitoring system of magnetic drive pump sliding bearing[J]. Machine tool and hydraulics,2007,35(2):159-161.(in Chinese)