Advanced Control of a Mixing Process
Jana Závacká *, Lenka Blahová, Monika Bakošová, Ján Dvoran
Institute of Information Engineering, Automation and Mathematics, Faculty of Chemical and Food Technology, STU in Bratislava, Radlinského 9, 812 37 Bratislava, Slovakia
E-mail: * jana.zavacka@stuba.sk
Abstract: Two advanced approaches to control of a laboratory process are compared in the paper. The first approach is robust one and is based on design of robust PI controllers for systems with parametric uncertainty. The design method is based on plotting the stability boundary locus in the controller-parameter plane and the sixteen plant theorem. The stability boundaries obtained for sixteen Kharitonov plants split the controller-parameter plane in stable and unstable regions. The parameters of robust PI controllers are chosen from the stable region common for all sixteen plants. The second approach combines the neural-network based predictive controller and the neuro-fuzzy controller. The neuro-fuzzy controller works in parallel with the neural-network predictive controller and corrects its output in order to enhance the control response. Both methods are applied for control of a laboratory chemical continuous stirred tank reactor that is used as a mixer. NaCl solution with desired concentration is prepared in the equipment. The conductivity of the solution is the controlled variable and the volumetric flow rate of water is the manipulated variable.
Keywords: process control, uncertainty, robust PI controller, neuro-fuzzy controller
Full paper in Portable Document Format: acs_0088.pdf
Acta Chimica Slovaca, Vol. 4, No. 2, 2011, pp. 18—32