Análisis predictivo de consumo y tarificación dinámica de energía eléctrica mediante simulación de red AMI y Máquinas de Aprendizaje para el sector residencial.

Abstract

This document develops the curricular integration work “Predictive analysis of electricity consumption and dynamic pricing through AMI network simulation and machine learning for the residential sector.” Chapter 1 describes the background of the research, the definition of the problem, the general objective, the specific objectives, and the research methodology used. Chapter 2 describes the theoretical framework of electricity generation, electrical power systems, smart metering systems (AMI), machine learning, simulation tools, and the project survey, including the analysis of the consumption profile and the current pricing scheme. Chapter 4 presents the system modeling using Python, integrating the AMI network simulation. Chapter 5 presents the analysis of the results obtained, showing that predictive models, especially LSTM, are able to estimate demand with good accuracy in the short term. Likewise, the comparison between TOU and RTP tariff schemes shows that the price structure directly influences the final cost to the user, demonstrating the potential of dynamic pricing as a tool to encourage more efficient residential consumption management. Finally, conclusions and recommendations are presented.