Mejora del pronóstico de carga a corto plazo en el sistema de distribución de la Empresa Eléctrica Ambato Regional Centro Norte S.A. (EEASA) basada en criterios de Big Data

Abstract

This technical project focuses on the improvement of the process to obtaining the demand forecast of electrical energy in the Empresa Eléctrica Ambato Regional Centro Norte S.A. (EEASA) using Big Data criteria, through the elaboration of a computer model in PYTHON programming language 3.7.

A graphical analysis module is linked to the development of the program that allows evaluating the electric charge values present in the EEASA database and additionally an electrical load projection module in which algorithms of machine learning or automatic learning are integrated, such as the Autoregressive Integrated Moving Average, Linear Regression, Support Vector Machine and Multilayer Perceptron.

Different study cases were proposed as the most frequent scenarios for the generation of load forecast such as: holidays week, scheduled and unscheduled incidence weeks, dry season week and rainy season week, which allowed to test each implemented algorithm, evaluate their results and compare them with each other by calculating the indicators of the absolute average error percentage (PEMA), the quadratic average error (EMC) and the forecast accuracy (EP).