Multi-regional fuzzy control for a pH neutralisation process

1 Introduction pH control is the process of adding acid or base to opposing solutions aiming to achieve a solution of pH 7, defined as neutral solution. The function of the controller is to achieve a neutral output of the system. This in itself is not an easy problem to solve as pH control of a neutralisation process is a difficult non-linear control problem. This is largely due to the inherent non-linear behavior of pH measurement, and is further complicated by variable time delays in the process. The relationship between pH measurement and volume ratio of acid to base is non-linear and characteristically described as an S-type curve relationship. Therefore the pH control problem can be considered as posing three different regions of sensitivity for control. For this reason alone conventional methods of control provide unsatisfactory control results, as can only truly accommodate one region scenario. This paper presents both simulated and practical results of applying a unique fizzy logic method to control the pH process, utilising the novel concept of mutli-regional PD type fuzzy logic control. Each region of sensitivity is assigned a unique fuzzy logic model and set of rules. The results are placed in context with the results of conventional means and traditional PD type fuzzy logic control combined with a variable transformation technique. The results demonstrate the accuracy of the methods in achieving the objective, that is an output of pH 7 Fuzzy logc control proposes an alternative method of control for systems where the dynamics of a system cannot be truly defined in mathematical terms. In the case in point pH control is a difficult non-linear problem, that can be defined mathematically for known influents into the system, but for unknown any mathematical model used is not true representation of the systems inputs. This is characterised by the fact that strong acid input has different non-linear characteristics to that of weak acid influent, as shown in Figures 1 and 2, respectively. This is further compounded when weak acid and strong acid influents are combined, the non-linear characteristics change again and are interdependent on the volume and concentration ratio of weak to strong acid within the influent. Therefore the pH control problem lends itself to fuzzy logic control. Fuzzy relational model-based controller has been researched as a solution to the problem of pH control, [I], opposed to the Mamdani fuzzy logic control approach. The Mamdani approach …