The designs of model-based state space controllers for industrial twin shaft gas turbines, presented at last year’s conference [1], were enhanced by a limiting function for selected state variables. In order to avoid the disadvantages of common controller concepts involving abrupt structural changes, the limitation was realised by a parameter-variant state space controller. To reduce the sensitivity of the full state space controller to parameter changes, a reduced order controller was developed, taking into account only the dominant state variables of the system. As in previously presented designs, a PI state space controller was designed for the reduced system using the pole placement method. Subsequently, this reduced controller was adapted to the original nonlinear system. With appropriate pole placements for the reduced order state space controller, a high quality of control, comparable to the behaviour of a full state space controller, can be obtained. The resulting controller also shows a reduced sensitivity to variations of the feedback parameters. The intended state variable limitation of the original nonlinear system to defined thresholds has been achieved by applying floating functions between different controller parameter sets.

This content is only available via PDF.
You do not currently have access to this content.