Software to Improve Control System in an Ethanol Distillation Process

This work has developed a predictive control solution based on specific models for the process of ethanol distillation. The advantages of such control are relative to the prediction of the consequences of the disturbances by the model, thus enabling the control action to be done in a previous manner, resulting in the minimization of the variables fluctuation controlled by the process. This results in, among other advantages, energy economy, in the improvement of the ethanol produced and in the increasing production capacity. Another desirable characteristic in this control mode is its capacity to act in non-linear systems as is the case of the distillation columns. Finally, it must be noted that with the application of an advanced control solution, as proposed in this study, it becomes viable, in a second moment, for the ethanol plants to operate in multiple operational conditions, such as: 1) maximum energy economy (scarcity of raw material, for example) and: 2) maximum production condition (for situations with excess of materials to be distilled). The models developed in this project will consist of purely empirical models. Several tests will be done in the different types of models to measure the precision and robustness. The proposed control strategy demonstrated be able to control selected control loops adequately. Steam savings and reduction of product losses were observed.