| 總結: | Most of the process industries use variable area tank processes such as conical and spherical tank processes. However, their nonlinear dynamics and interactions between such processes along with measurement noise, make controller design challenging. In spite of their widespread use, the development of a unified control strategy that ensures reliable performance across the entire operating range remains an open research problem. Hence, this paper proposes Non-Linear Model Predictive Control (NLMPC) with Unscented Kalman Filter (UKF) for a nonlinear lab scale interacting two tank hybrid process, which provides dynamic model predicting capability and better compensation towards nonlinearities in the presence of continuous measurement noise. To enhance the performance, NLMPC is designed using the estimated states with the help of UKF. Initial simulation is performed by the proposed controller with/without estimator under servo and regulatory operations in the presence of disturbance and noise. Further, NLMPC with UKF is implemented in a lab scale setup and results are compared with existing controllers. The robustness of the controllers is assessed by computing various performance indices such as integral absolute error, integral square error and control effort.
|
|---|
