Bibliographic citations
Huamán, A., (2023). Modelos predictivos basados en Machine Learning para optimizar el diagnóstico del índice de salud en los interruptores de potencia de una empresa de transmisión eléctrica [Tesis, Universidad Nacional de Ingeniería]. http://hdl.handle.net/20.500.14076/27298
Huamán, A., Modelos predictivos basados en Machine Learning para optimizar el diagnóstico del índice de salud en los interruptores de potencia de una empresa de transmisión eléctrica [Tesis]. PE: Universidad Nacional de Ingeniería; 2023. http://hdl.handle.net/20.500.14076/27298
@mastersthesis{renati/713038,
title = "Modelos predictivos basados en Machine Learning para optimizar el diagnóstico del índice de salud en los interruptores de potencia de una empresa de transmisión eléctrica",
author = "Huamán Sarzo, Angel",
publisher = "Universidad Nacional de Ingeniería",
year = "2023"
}
The purpose of this research is to lay the foundations for the use of predictive models with artificial intelligence based on machine learning applied to the technical diagnosis of the health index in the power circuit breakers of an electric power transmission company, in arder to obtain a predictive model with artificial intelligence that is effective and efficient, which will be trained with all the expert knowledge that we have from previous studies to obtain an optimal predictive model based on machine learning, Unlike the traditional method used in the company (fuzzy logic) where the machine must be told which are the parameterization rules and have a logical inference criterion for each rule, the predictive model used all the technical parameters (predictor variables) involved in the diagnosis of the health index (target variable) and designed its own mathematical model based on a computational analysis of correlations and principal components. The application of machine learning to the diagnosis of the health index in circuit breakers optimized the entire process that is currently followed, achieving more accurate diagnoses and with better timing. With the design of the predictive model, an effectiveness of 99.27% was achieved with a decrease in man hours of 76.19% with respect to the old methodology.
This item is licensed under a Creative Commons License