Gain and order scheduled fractional-order PID control of fluid level in a multi-tank system

Gain scheduling is widely regarded as an effective nonlinear control technique, and its extension to fractional-order control is a natural step. In this paper, we investigate a particular method based on a gain and order scheduling approach for fractional-order PID controllers. The method is applied to the control of a real-life laboratory model of an industrial multi-tank system. Gain and order scheduling is realized by means of a control law comprising two static PID controllers and an appropriate control blending rule providing this way means for stability analysis of the control system. The design of controllers for level control in the first tank is carried out by considering linear fractional-order approximations of the nonlinear model of the process with locally applicable frequency-domain robustness specifications. The controller for the second tank is obtained using time-domain optimization of the transient response. In addition, an extended Kalman filter is designed to reduce measurement noise propagation into the control law thereby enhancing the performance of the pump. The majority of necessary computations, including those related to controller design, are performed numerically in the FOMCON toolbox for MATLAB.