Research Of Condition Monitoring And Fault Diagnosis About Industrial Pump

Pump plays an important role in industrial production and daily life as a general machine ry, which has been widely used especially in the electric power and petrochemical industries. At the present stage, thermal power is still in the majority of the generating capacity and the c oal-fired power will still be the major power generation method in a period time in the future. Pump plays a large role as an important auxiliary equipment in power plant production. In the petrochemical industry, the status of the pump is also important. The materials, solvents and p roducts which production requires such as liquid are transmitted through the pump. Because o f the variety of craft materials, the very demanding material conveying conditions occasionall y and large amount of pump with large energy consumption in petrochemical plant, it’s very i mportant to ensure the normal operation of the pump in electric power and petrochemical indu stries. Industrial pump belongs to the high speed rotating machinery equipment with a high fai lure rate. This paper takes the pump equipment as the research object, summarize the types, st ructure, operation characteristics and typical failure of industrial pump. This paper mainly co mpleted research work as following:1) Statistic analysis on a variety of industrial pump failure cases, which can be references of the running state monitoring and fault diagnosis of pump.2) Vibration monitoring data processing of hundreds of pump in a petrochemical company, and get the signal features both in time and frequency domain, from which we can distinguish fault signals from normal ones.3) Make rules on classification of fault signals based on data mining method and decision tree algorithm. Then we can use these rules on other operation signals. It is more convenient and accurate when we use data mining method to analysis the signals. So it is a method of higher efficiency.