Design Of Intelligent Monitoring System For Canned Motor Pump Based On Health Assessment

Canned motor pump is a common sealless pump,which is widely used in important fields such as civil,chemical,military and aerospace,such as liquid delivery systems in the chemical and pharmaceutical industries,and rocket launch fuel refueling systems for aerospace.The health status of canned motor pumps has a huge impact on production.Therefore,strengthening the status monitoring and health assessment of canned motor pumps is of great significance to ensure production safety.Rotor-graphite bearing wear is the most common failure of canned motor pumps,and it can also indirectly lead to rotor-system failures such as rotor imbalance,rotor misalignment,and mechanical loosening.At present,the deflection yoke potential difference method is often used to detect graphite bearing wear failure,but this method has low sensitivity,it is difficult to detect a slight degree of bearing wear,and it cannot detect bearing symmetrical eccentric wear failure.Based on this problem,this paper proposes a method of using signal processing technology to extract health parameters from the electrical and vibration signals of the deflection yoke,and to detect bearing wear based on the extracted results.This article designs experiments to verify the feasibility of this method,and develops a canned pump Io T intelligent monitoring system based on this.In this paper,the typical structure of the canned motor pump is firstly studied,and the structure and working principle of the key components such as graphite bearing and canned motor are studied.Then summarize the failure mechanism and signal characterization of rotorgraphite bearing wear,rotor imbalance,rotor misalignment,and mechanical loosening.Then,the acquisition methods of the deflection yoke electric signal and the vibration signal of the canned motor pump are introduced in detail,and the signal processing methods in the time domain and frequency domain are summarized.The rotor-graphite bearing wear monitoring experiment was designed to screen the time domain and frequency domain indicators.The peak value and effective value of the electrical signal of the deflection yoke were selected as the calibration parameters of the graphite bearing wear degree,and a method for judging the location of the bearing wear based on the vibration signal was proposed.It has been verified by experiments that this method is more sensitive than traditional methods and has a better diagnostic effect on symmetrical wear of bearings.Finally,the method was applied to the field of health assessment of canned motor pumps,and an intelligent monitoring system of canned motor pumps was developed based on Java14.0language,which realized the evaluation function of the health status of canned motor pumps.