AbstractDue to the growth in global demand for biodiesel, there is a tendency for the volume of its main by‐product, glycerol, available on the market to increase. This has led to a search for new uses for this by‐product and the development of new processes that use glycerol as a raw material because it has wide industrial applications. However, to be used in many of these processes, the glycerol must go through a purification stage before reaching the purity required for the application (which can reach 99.7% w/w). Taking this into account, the optimization of the parameters for the main glycerol purification stage (vacuum distillation) is essential to reduce the cost of this stage as much as possible. In this article, optimization of the vacuum distillation column was carried out using an artificial neural network (ANN) and a genetic algorithm (GA) together. With the optimization, it was found that four theoretical stages of equilibrium (separation) is the optimal value, in economics terms, for the distillation column on an industrial scale, leading to better performance (in terms of glycerol purity) than patents that used a smaller number of theoretical stages and an equal performance in relation to the patents that used a greater number of stages, which in this case reduces the cost with the size of the column. This result helps to increase the financially viability of using glycerol as a raw material, which can make several processes cheaper, thus generating greater added value to this sustainable by‐product. © 2023 Society of Industrial Chemistry and John Wiley & Sons Ltd.
Read full abstract