Classification of the status of the voltage supply in induction motors using Support Vector Machines

Abstract

Induction motors represent a mainstay in the industry. Its optimum performance is related to the quality of electrical supply. In this sense, when there are variations in the quality of electricity supply, the induction motor can operate in conditions that can affect its performance and shorten its life. In this paper we propose a monitoring of the voltage condition of the supply voltage in induction motors, through a classification using Support Vector Machines (SVM) to detect and classify these events. The necessary data for defining the events associated with the supply voltage are obtained with the use of ATP software, simulating these different conditions in a low voltage motor and the event classification is carried out with the use of SVM. With the use of this mathematical tool, defined areas of disturbs in the voltage will be established from measurements obtained in the simulation of each of the events, which are considered as perturbations in voltage, such as: overvoltage, undervoltage, voltage unbalance, distortion of the voltage wave, voltage sags and simultaneous combinations of some of these events. Assigning tags to each type of disturbance or disturbance combinations allows us to detect and classify the condition of energy supply in induction motors through the SVM. The results show perfect accuracy in the classification of events in the condition of voltage supply of the induction motor, which represents an alternative to the monitoring of these machines in the industry.