Bibliographic citations
Blancas, J., (2019). Implementación del método razonamiento inductivo difuso para el pronóstico de la demanda eléctrica a corto plazo en el SEIN [Tesis, Universidad Nacional de Ingeniería]. http://hdl.handle.net/20.500.14076/18474
Blancas, J., Implementación del método razonamiento inductivo difuso para el pronóstico de la demanda eléctrica a corto plazo en el SEIN [Tesis]. : Universidad Nacional de Ingeniería; 2019. http://hdl.handle.net/20.500.14076/18474
@mastersthesis{renati/710352,
title = "Implementación del método razonamiento inductivo difuso para el pronóstico de la demanda eléctrica a corto plazo en el SEIN",
author = "Blancas Sánchez, Jordan Darwin",
publisher = "Universidad Nacional de Ingeniería",
year = "2019"
}
This thesis will focus on the application of “fuzzy inductive reasoning“ (FIR) to the problem of short term load forecasting. The FIR model, which is based on fuzzy logic, learns the past relationships of the electricity demanda (load) and predicts the behavior of the demand from the last real aggregate data in order to establish the deviations of the programmed demand versus the real demand. The objective of this work is to determine the lowest forecast error by means of the indicator “average absolute percentage error“ (MAPE for its acronym in English), so first of all the foundations of fuzzy logic are developed as a basis to understand the FIR methodology and a previous forecasting method is proposed only using fuzzy logic as a tool. Later, the proposed FIR methodology is developed as an improvement of the methodology with fuzzy logic and which is also complemented by the use of an evolutionary algorithm called “simulated rebound algorithm“ (SRA in English) that will serve as an optimization method (minimization of MAPE) to determine the linear and nonlinear relationships between the variables and thus identify the set of input variables that improve the accuracy of the prediction. Then, both the methodology with fuzzy logic and the FIR methodology complemented with the implementation of SRA, will be applied to the Peruvian energy system through the use of the historical data of the electrical demand of the National Interconnected Electric System to determine the forecast of the electricity demand of the next day (short term). Finally, it is shown that the FIR methodology offers minor errors in the forecast of electricity demand compared to the methodology with fuzzy logic developed initially and with the methodology currently used by the operator of the national electricity system (COES-SINAC), which uses adjustment for least squares as a forecasting tool.
This item is licensed under a Creative Commons License