ScienceAsia 28 (2002) : 365-376 |doi: 10.2306/scienceasia1513-1874.2002.28.365
Use of Globally Linearizing Control with Extended Kalman Filter for pH Control of a Wastewater Treatment Process
Paisan Kittisupakorna, MA Hussainb, N Siripuna, W Daosuda and P Kaewpradita
ABSTRACT: Several chemical industrial plants such as electroplating and metal finishing plants have used strong acids and strong bases in production lines. These acids and bases are then released from the production lines to a wastewater treatment system and then treated to achieve compliance with an effluent standard. It is well known that the pH control of a wastewater treatment process is one of the most challenging control problems due to high non-linearity and time-variance of the pH value during pH titration. A conventional PID controller and an on-off controller are rarely able to handle this nonlinearity resulting in poor control performances. Therefore, advanced nonlinear control techniques are needed. This research presents simulation study of Globally Linearizing Control (GLC) together with an extended Kalman Filter to control pH of the wastewater treatment process of an electroplating plant. The GLC, one of the advanced nonlinear model-based control techniques, has been developed for both Single-Input and Single-Output (SISO) or Multi-Input and Multi-Output (MIMO) nonlinear process systems. Since the GLC is a model-based control technique, it needs measurements and values of states and parameters, which are neither all measurable nor known exactly. Therefore, the extended Kalman Filter, a state and parameter estimation technique, is applied to estimate unavailable or unknown states and parameters, and these estimates are incorporated in the control action determination of the GLC algorithm. Simulation results have shown that in a nominal case, the GLC is able to control the pH of the system to a desired set point and its control performance is equivalent to that of a PID one. In the presence of plant/model mismatch, the GLC is still able to handle this mismatch and gives good control performance whereas the PID gives poor control response; the GLC is much more robust than the PID controller.
a Dept of Chemical Engineering, Chulalongkorn University, Bangkok 10330, Thailand.
b Dept of Chemical Engineering, University Malaya, 50603 Kuala Lumpur.
*Corresponding author, E-mail: firstname.lastname@example.org
Received 6 Aug 2001, Accepted 29 May 2002