Data Analysis and Model-based Control of Multi-effect Evaporators for Energy Saving

Abstract

Background: In the production of sugar, optimizing energy usage in multiple-effect evaporators (MEEs) is a major difficulty. This study offers a thorough method for enhancing energy efficiency and control performance in a five-stage MEE system at a sugar mill that is currently in operation. We significantly increased control accuracy and steam efficiency by automating formerly manual procedures and putting sophisticated control techniques into place. With correlation values above 0.96, Support Vector Regression (SVR) was shown to be the most successful statistical regression technique for simulating system behavior out of seven approaches. A third-order state-space model appropriate for controller design was produced by subspace identification techniques. When compared to manual operation, the Model Predictive Controller (MPC) increased steam economy by 1.36% and decreased control errors by almost 99%. Additional results indicate that using the best vapor bleed techniques resulted in a 10% increase in steam economy. This study offers a framework for comparable energy optimization initiatives in process industries and illustrates the useful advantages of advanced control in industrial evaporation systems.