Kiln Predictive Modelisation for Performance Optimization

Abstract

Exploiting the power of AI in the heavy industry brought back satisfactory results at the kiln level, machine learning techniques allowed predictive modeling of the baking process using powerful Machine Learning models which had a great impact on the energy consumption and kiln production rate among all the models used. Large amounts of historical data were used after analysis and preparation, for which several methods were applied such as preprocessing and feature selection. All models were tested on 20% of the data, using Mean Absolute Error and Root mean squared error as metrics to evaluate our models in order to identify the influencing variables that contribute most to the increase in energy consumption and kiln production rate.