Mathematical models for the prediction of consumption and generation of electrical energy through generator sets that use oil and diesel from the Petroecuador company

Abstract

This study develops a mathematical model to predict electricity consumption and generation in Petroecuador's generator sets using oil and diesel. By implementing advanced machine learning techniques, specifically decision trees and ARIMA models, it was possible to capture the complex relationships between fuel consumption and energy production. Analysis of historical data revealed a strong linear relationship between crude oil consumption and power generation, while diesel showed a non-linear and weaker relationship. The model achieved a determination coefficient of 0.97 in validation with real data, demonstrating exceptional predictive capability. This precision was achieved using detailed daily data, underlining the importance of maintaining frequent and thorough records. The developed model not only allows accurate predictions but also offers a versatile tool for strategic planning, inventory optimization, maintenance scheduling, and improving Petroecuador's overall energy efficiency. This study establishes a solid foundation for future research in the field of energy prediction and efficient resource management in the oil and energy industry.