چكيده به لاتين
Abstract:
Pumps are one of the most widely used machines in various industries and centrifugal pumps are the most common type used pumps, therefore, this type of pump maintenance is very important and in some sensitive industries need to implement a monitoring system for this type of pump is unavoidable, one of the major defects in the centrifugal pump is displayed and the damage is irreparable is cavitation. cavitation caused the loss of head, flow and efficiency of centrifugal pumps and cause corrosion and serious damage to the blades surfaces, this study aims to design and test a monitoring system for diagnosing cavitation in centrifugal electro-pumps combine information on vibration signals and electric current signals, a laboratory for the simulation of cavitation in centrifugal pumps designed and built and at different flow rates at the pump cavitation and fault simulated.Pump vibration signals and signals from the electric pump at any rate before the start of cavitation to start and spread cavitation are measured.the measured signals are analyzed and processed and then extract critical parameters are sent to the classification, the decision to enhance the accuracy and reliability of the monitoring system designed, combining data using bayesian methods is done. Finally, the results of the monitoring system fused with vibration method alone monitoring system and method alone compared electrical current analysis .
Keywords:Centrifugal pump, Cavitation, Signal processing, Support vector machine, Data fusion
