NON-LINEAR MODEL REFERENCE CONTROL OF pH PROCESS: AN EXPERIMENTAL STUDY

The control of pH is important in many processes including wastewater treatment, chemical processes and biological processes. This paper considers a model reference non-linear controller developed by Jayadeva et al. (1990a). The method is tested using a 7-litre continuously stirred tank reactor to neutralise a strong acid using a strong alkaline solution. The method is first realised using a simulation of the process. Subsequently it is demonstrated on an experimental rig using real-time control. Experimental results confirm that a robust control of the process is achievable. INTRODUCTION In many processes, pH neutralisation is a very fast and simple reaction. In terms of practical control, it is recognised as a difficult control problem (Shinskey 1973; Pishvaie et al. 2000; Wright et al. 1991). The difficulties arise from high process nonlinearity (the process gain can change tens or hundreds of times over a small pH range) and from changes in the pH characteristics due to changes in influent concentration. Various techniques have been developed to control process pH. Young and Rao (1986) presented a variable structure controller (“sliding mode control for a neutralisation process”) involving strong acids and bases. Parrish and Brosilow (1988) used nonlinear inferential control in a simple simulated neutralisation process, using static estimation of the concentration of a single monoprotic weak acid. Kulkarni et al. (1991) presented non-linear internal model control for a simulated system of sodium hydroxide (NaOH) and hydrochloric acid (HCl). Li et al. (1990a) and Li and Biegler (1990b) presented non-linear feedback methods for a simulated neutralisation process. In the present work, a non-linear controller design is implemented. It uses a design procedure presented by Jayadeva et al. (1990b). The controller is implemented practically on a 7litre reactor.