Study on condition monitoring method of slurry circulating pump based on support vector machine

A slurry circulating pump is an important auxiliary equipment in a limestone gypsum desulfurization system. The normal operation of the slurry circulating pump is of great significance for the unit to meet the ultra-low discharge requirements. The reagent of the desulfurization system is limestone slurry, which makes the auxiliary equipment vulnerable to wear and blockage. As the main parameter of the slurry circulating pump, the bearing temperature can well reflect the operating state of the equipment. In this paper, the prediction model of temperature measurement points is established by using a support vector regression algorithm, and the temperature monitoring of a single temperature measurement point is realized. Secondly, considering that multiple temperature measurement point groups will fail when the temperature is abnormal, according to the actual performance characteristics of the measurement points, this paper uses the support vector machine classifier algorithm to realize the abnormal diagnosis of multiple temperature measurement point groups. The experiment with the field's actual measurement point data shows that the model proposed in this paper can realize the fault location when multiple measurement points rise abnormally.