Multi-Domain Modeling Platform For Electrical-Signature-Based Condition Monitoring of Motor-Driven Pumps

Due to the vital role of motor-driven pumps in various industries such as oil and gas, manufacturing, chemical, etc., their continuous monitoring and implementing effective maintenance methods is of crucial importance. periodic inspections and intermittent vibration data collection using accelerometers is among the most common methods. electrical signature analysis is an alternative approach that only uses electrical measurements for the purpose of fault detection. despite the unique advantages of this method, such as its non-intrusiveness and possibility of continuous monitoring of the equipment, there have been limited studies on its underlying theory with majority of the proposed esa-based methods taking data-driven approach towards condition monitoring problem. data-driven methods rely on the experimental data collected from the equipment to train the statistical models. this imposes a serious limitation on the application of electrical signature analysis and makes the generalization harder. in this paper the electromechanical coupling in a motor-driven centrifugal pump is studied in order to demonstrate the effects of different operating conditions of pump on motor electrical signals. lumped parameter approach is employed to derive governing equations of the induction motor and computational fluid dynamics is utilized to analyze the interaction of the centrifugal pump blades and fluid. such a modeling platform presents a physics-based approach towards electrical signature analysis based condition monitoring. a closed-loop hydraulic test rig is built to compare and verify the simulation results.