Energy simulation and variable analysis of refining process in thermo-mechanical pulp mill using machine learning approach

Data from two thermo-mechanical pulp mills are collected to simulate the refining process using deep learning. A multilayer perceptron neural network is utilized for pattern recognition of variables. Results show impressive capability artificial intelligence methods in energy simulation so that correlation coefficient 98% accessible. comprehensive parametric study has been made investigate effect disturbance variables, plate gap and dilution water on simulation. The generated model reveals non-linear hidden between which can be used optimal control strategy. Considering variables’ simulation, accuracy could increase by 15%. Removing gape predictive variables reduces determination up 25% both mills, while mentioned value removing 9–17% mill 1 about 35% 2.