(11b) Condition Monitoring of a Centrifugal Pump with Magnetic Signatures

Centrifugal pumps are an important asset in a wide range of industries. Generally working in severe environments and carrying various fluids from water to oils, caustic acids, and sewage, they are subjected to severe deterioration and mechanical failures (accounting for more than 60% of their breakdowns). Cavitation, a phenomenon which can cause significant damage to the impeller and other moving parts of a pump, can result in a pump breakdown if gone unnoticed. Current condition monitoring techniques are mainly focused on the development stage of these pumps to control product quality or ensure the integrity of these assets. Other widely used techniques such as vibrational and acoustic analysis have to deal with mechanical and electrical noise which can affect the accurate diagnosis or require extensive computation to monitor the health of these assets. In this paper we present a new methodology of condition monitoring the health of a centrifugal pump using an array of magnetic and infrared sensors. The magnetic sensors capture magnetic signatures in the electromagnetic field emitting from the centrifugal pump, which are able to identify abnormalities such as cavitation. The Infrared sensor is able to pick up abnormal temperature rises resulting from cavitation or any other defects during the operation of the pump. The captured data when sent to an Artificial Neural Network would be able to accurately predict faults in the asset’s lifetime thus allowing the user to make operational changes to increase the life of the asset and prevent a breakdown.