A quadratic cost function for the optimal control of binary distillation column was minimized using quadratic programming. An input-output model was used and its parameters were recursively updated by instrumental variable method.
A control experiment was conducted for the performance appreciation of the control technique. The performance of the adaptive control was found to be much improved comparing with the conventional PID control. The technique was capable to obtain fast solution and easy to apply for the multi-input-multi-output systems. And also the algorithm was stable for a long term operation which is very critical for the industrial application.