An adaptive Industrial Internet of things (IIOts) based technology for prediction and control of cavitation in centrifugal pumps

Cavitation is the phenomenon whereby the air bubbles implode on the impeller due to insufficient suction pressure at the inlet section thereby causing noise, vibration and in a long-term, corrosion of the impeller blades. However, the current devices and technologies used to address cavitation are inefficient. The use of intelligence gathered from telemetry data produced via cloud computing and IOT is becoming prevalent in the prediction and control of cavitation. The objective of this research is to simulate cavitation in industrial pumps using the IOT technology. Cavitation was simulated in both good and faulty centrifugal pumps using the Fast Fourier Transform Algorithm (FFT) integrated to an Arduino microcontroller. Subsequently, configuration of the sensors required to capture the data, ingestion into the cloud and numerical analysis of the data were executed using the Artificial Neural Network (ANN) as the machine learning algorithm. The results showed graphs as well as the tables of non-cavitation and cavitation patterns when the ball valve was fully opened (90º) respectively, at the suction of the centrifugal pump. For the non-cavitation patterns, the voltages lie from 0.5-0.75v and for the cavitation patterns, the voltages lie from 0.76-0.95v. Using this technique, the detection, analysis and accurate prediction of cavitation were established.