Técnica de ciencia de datos para el pronóstico de consumo de gas natural en la industria siderúrgica

Efficient scheduling of natural gas energy resources within the steel industry faces many challenges, as this sector does not have the necessary measures and tools to support its efficient management. In this study, a forecasting approach to natural gas consumption is proposed using the predictive technique of multiple linear regression. For the development of the proposed model, the main variables related to natural gas consumption that will make up the predictive model are established. The evaluation of the model was carried out using data from a steel company in Mexico. The results of mean absolute percentage error, root mean square error and mean absolute deviation of RLM models (MAPE: 10.23%, RMSE: 20492.32, DAM: 20009.03) vs. traditional method (MAPE: 34.05%, RMSE:93055.92, DAM: 65170.91), reflect that the proposed models significantly improve natural gas resource management, providing substantial improvement in natural gas consumption volume estimates for proper energy scheduling.