Development and Prediction of Cavitation in Centrifugal Pumps using Industrial Internet of Things Technology

Abstract

Abstract The enormous cost of downtime and safety hazards caused by failure of pumps due to cavitation are widespread in the oil and gas industry - upstream, midstream and downstream sectors. Current devices and technologies used to address cavitation are either used after it occurs or very expensive to procure. A cheaper and more efficient system was proposed, capable of addressing cavitation in centrifugal pumps. A test rig of the centrifugal pump was constructed, integrated and programmed the sensors using C++ and Arduino microcontroller. Subsequently, the telemetry data from the sensors and predicted the likelihood of cavitation or no cavitation by varying the ball valve angles were analyzed. Although preliminary results are very promising from current methodology, but currently working on using python and machine learning techniques. It is expected that these and more will further improve pump cavitation prediction accuracy.