Learning to Control pH Processes at Multiple Time Scales: Experimental Application Assesment

S. SYAFIIE; F. TADEO; E. C. MARTÍNEZ

Kuala Lumpur, Malasia

Conferencia; Asian Pacific Conference of Chemical Engineering; 2006

Asian Pacific Confederation of Chemical Engineering

This article proposed a solution to pH control based on model-free learning control (MFLC). The MFLC technique provides a general solution for acid-base system which is simple enough for implementation in existing control hardware. MFLC is based on reinforcement learning (RL), which is learning by direct interaction or experience with a system/plant. The MFLC algorithm is model free and allows handling input and output constraints. A novel solution of MFLC using multi-step actions (MSA) is implemented: using abstract actions at multiple time scales consisting of several identical primitive actions. This temporal abstraction strategy helps solving the problem of determining a suitable fixed time scale to select control actions so as to trade off accuracy in control against learning complexity. An application of MFLC for pH process at laboratory scale is presented, showing that the proposed MFLC is able to control adequately a typical neutralization process, and maintain the pH in the desired goal band. Also, the MFLC controller generates a smoothly varying control signal to drive pH towards the goal.