Análisis de consumo energético en el laboratorio de industria 4.0 y robótica de la facultad de informática y electrónica para la predicción del factor de potencia mediante técnicas de machine learning.

Abstract

The laboratory of Industry 4.0 and Robotics of the Faculty of Computing and Electronics does not have devices to measure and analyze the quality of energy, as well as other fundamental electrical magnitudes, therefore, the objective of this degree work is to analyze the consumption of energy in the Laboratory of Industry 4.0 and Robotics of the Faculty of Computing and Electronics and predict the power factor applying machine learning techniques. The methodological process of this work includes the identification of the power supply line of the Laboratory of Industry 4.0 and Robotics, the reading of the fundamental electrical magnitudes such as: Voltage, Current, Power and Power Factor, the acquisition and sending of data through an IoT gateway, the creation of a database, the implementation of a web HMI interface for supervision and monitoring of electrical variables and the application of power factor prediction algorithms. With this methodology, the integration of multiple devices was achieved that will allow determining the fundamental electrical parameters, connection to the cloud and providing Internet of Things (IoT) service, in addition, a comparison of three models based on machine learning that allow predicting the power factor efficiently is presented. Through the tests carried out on the prediction algorithms, the mean square error, the score, the prediction limits and the estimate of the Kernel density for each algorithm were determined. In conclusion, it was possible to determine the energy consumption and existing power variations in the Industry 4.0 and Robotics Laboratory and implement algorithms that allow to predict the power factor effectively.