Centralized auto-tuned IMC-PI controllers for industrial coupled tank process with stability analysis

Multiple interconnected tanks supplied by more than one pump is a MIMO process which is quite common for industrial applications. Presently, Internal Model Control (IMC) technique is widely used in controlling industrial MIMO processes due to its sole tuning parameter. But, as the IMC technique is entirely based on the process model hence in case of any uncertainty in modeling, nonlinear interacting behavior as well as time varying features of the process parameters impose limitations on the performance of IMC controllers. To overcome such limitations, in the proposed work, a simple auto-tuning scheme is incorporated in the conventional IMC based Proportional Integral controller (IMC-PI). In the proposed auto-tuner, the only tuning parameter i.e., the close-loop time constant is continuously varied depending on the instantaneous process error. Performance of the proposed auto-tuned IMC-PI controller (IMC-API) along with conventional IMC-PI controller are verified on a miniaturized industrial coupled tank process which is a well known interacting MIMO process. Here, control task is aimed towards maintaining the water levels at the respective desired values under both the set point change and load disturbance. Controlling MIMO processes inherently includes uncertainties which have to be taken care through adequate control design. Therefore, stability study of a close-loop system along with robustness of controller has to be ensured against process model perturbation. Experimental results substantiate the performance superiority along with its robustness of the proposed IMC-API compared to conventional IMC-PI controllers.