Citas bibligráficas
Hernández, N., (2023). Aplicación del Machine Learning para predecir el tiempo de vida útil remanente de un Molino SAG en una empresa minera [Tesis, Universidad Nacional de Ingeniería]. http://hdl.handle.net/20.500.14076/26698
Hernández, N., Aplicación del Machine Learning para predecir el tiempo de vida útil remanente de un Molino SAG en una empresa minera [Tesis]. PE: Universidad Nacional de Ingeniería; 2023. http://hdl.handle.net/20.500.14076/26698
@mastersthesis{renati/712731,
title = "Aplicación del Machine Learning para predecir el tiempo de vida útil remanente de un Molino SAG en una empresa minera",
author = "Hernández Venegas, Napoleón",
publisher = "Universidad Nacional de Ingeniería",
year = "2023"
}
The general objective of this work is to use classification Supervised Machine Learning models to know in advance the number of hours that a SAG Mill has left until it presents a potential failure in order to avoid unscheduled stops and their economic consequences. for loss of profit. The Remaining Useful Lifetime (RUL) indicator is used for this purpose. Regarding data processing, the Python programming language was used on the Colab cloud platform. In the predictive analysis, the relevant predictor variables obtained from the sensors compiled in an information base of one year of operations of the SAG Mill were identified. An exploratory data analysis was performed treating outliers. Oversampling and k-fold cross validation data balancing techniques were used to optimize results. Shown in a comparative structure of performance indicators, it was determined through data processing that with the application of Machine Learning, specifically the XGBoost (XGB) classification algorithm, optimal performance in sensitivity was obtained to predict potential failure. . In the final process, a predictor was developed to estimate, with an accuracy percentage, the potential failure of the SAG Mill. Finally, and taking into account the specific objectives, the reduction of the costs due to lost profits is demonstrated as a result of the decrease in corrective maintenance activities and the increase in the billing of the metric ton per hour generated by the increase in the availability of the SAG Mill.
Este ítem está sujeto a una licencia Creative Commons Licencia Creative Commons