Comparative Study of Cavitation Problem Detection in Pumping System Using SVM and K-Nearest Neighbour Method

Machine learning is a fast-computational method and data analytics technique by which information be learned directly from the data without any predetermined equation as a model. The algorithms improve their performance as the number of samples available for learning increases. Todays' challenge is energy-efficient technology and conservation of energy. Machine learning technology is the part of artificial intelligence, which helps to keep the devices more energy-efficient and can be applied in various fields like agriculture, industry, medical sector etc. Since pumping system plays a significant role in the agricultural sector and most of the industrial fields, continuous monitoring is necessary to keep the system safe and reliable. The problems like sludge, cavitation, water hammering, rotor bearing fault, and impeller breaking are the significant causes of the damage of the pump. To detecting these problems, machine learning technique can be applied. There are various algorithms of machine learning like Support Vector Machine, Neural Network, Empirical Mode of Decomposition method, K-Nearest Neighbor method etc. This paper concentrates on detecting cavitation fault in pumping system by machine learning algorithm mainly by SVM algorithm and K-Nearest Neighbor Method and on a comparative study of SVM and K Nearest Neighbor algorithm.