Curva de demanda de energía eléctrica en sector residencial estrato tres de Bogotá D.C. mediante caracterización por redes neuronales artificiales

Abstract

The electric energy demand curve shows how the demand or electric load varies in a certain time interval; based on the forecast of the average consumption of electric energy by the load (residential, commercial and industrial), the country's electric energy dispatch is programmed. At present, there is no specific load curve for the residential load that specifically characterizes the consumption of each socioeconomic stratum, which is why this degree project seeks to characterize the demand curve of electric energy of the residential sector for the socioeconomic stratum 3 of the city of Bogota D.C. This requires the collection of data taken from field measurements; but because this degree project was realized during a pandemic (COVID-19), the data was collected through a survey, which was intended to gather information on the consumption of electricity of users throughout the day, based on the number of appliances they own, hours of use, number of inhabitants in the house, among other questions that contribute to the calculation of electricity consumption for certain hours of the day of the inhabitants of the residential sector stratum 3 of the city of Bogota DC. The survey was formulated to 271 people, a significant sample of the total population. In order to analyze the data obtained, it is first necessary to organize them, and then use statistical methods to detect and solve the presence of outliers in the answers obtained. If the outliers are not detected, they can cause the sample mean to not adequately represent the population sample, but on the contrary, to obtain biased information and in some cases erroneous information. With the data obtained from the surveys, an average curve is determined for the stratum 3 residential sector, complemented with technical data sheets and manuals of the household appliances present in the survey, in order to assign the specific power to each appliance. By means of the powers obtained, the values of time of use of the equipment, utilization factors, consumption in standby mode, among others, a daily consumption curve per equipment and per user is established, based on these results and through the application of robust statistical methods, a model of electric energy demand curve of the residential sector stratum 3 of the city of Bogota D. C. is presented. C, in this curve are presented results of monthly consumption equal to 145.51 kWh per month, value validated with the help of Matlab® software, computational tool that allowed the creation of an "Artificial neural network perceptron multilayer", which provides a curve with a monthly consumption equal to 144.89 kWh per month.