Control of recycle process particularly with the design of recycle compensator accompanied by the controller, arises with some issues. As sufficient evidence shows the necessity of recycle compensation for the desired performance of recycle process but the model-based recycle compensator design happens to be problem-dependent and non-realizable, which makes cancellation of recycle effect problematic. In such a scenario model-based controller design accompanied by recycle compensator has to accommodate both uncertainty and uncompensated recycle effect, which becomes a hard task. To overcome these control design issues in recycle system, formulation of control law in two possible ways with the estimates obtained from the Extended State Observer’s (ESO’s) design has been proposed. The reason for two possible ways of proposed control law formulation depends on the way recycle effect gets cancelled which based upon the order of the ESO’s design selected. Selecting the appropriate order for the ESO’s design and tuning parameters required for the proposed formulation, considered to be available from model knowledge of recycle system represented in a specific recycle structural form. However, this paper focuses only on the way the control law could be formulated for recycle system based upon the availability of recycle path data and has not considered providing an optimal selection of ESO’s order as well as tuning parameters. In simulation results, the performance of the proposed formulation under different strategies have been provided using the performance index of Integral Square Error (ISE) under nominal, perturbed, and external disturbance cases.
Read full abstract