Tests of hybrid Volterra-type models through NMPC schemes at a batch reactor

Abstract

A simulation of the control of temperature in a batch reactor aimed to prepárate melamine resin, is presented. In contrast to traditional techniques like PID by which the set point turns out to be of importance, a Nonlinear Model Predictive Control (NMPC) by using the Volterra model together to the orthogonal polynomials might permit to control the reference trajectory tracking of the internal temperature of the reactor. This work proposes a NMPC based on the second order hybrid Volterra-type model containing only 6 parameters. It allows to evaluate both advantages and disadvantages of the model as to automate the temperature trajectory tracking from the initial phase of the heating. From the results, interesting and favorable prospects to automate the reactor are expected. It might suggest that the hybrid Volterra models are a key step towards compact technologies containing newest NMPC schemes of high performance.