On identification of a coupled four tank system

In this work, the parametric identification of a multi-input multi-output (MIMO) four tank system is investigated. The four tank system contains four coupled tanks, two pumps, and a water reservoir. The system outputs are water levels of tanks while two pumps operate as control inputs. Four steps are employed to complete the identification process: experimental data acquisition, data filtering, parametric identification, and verification. Initial input step responses are used to choose proper sampling time, filtering requirement and Pseudo Random Binary Sequence (PRBS) inputs design. Then, PRBS inputs are applied to the system for identification and verification data collection. Two parametric identification methods, transfer function and state space, are employed for system identification. Finally, the performance of the identified models is compared with actual verification data. The results validate the good accuracy of the estimated models.