Loading...

Low Order Pre-compensator Design Using Linear Matrix Inequalities

Nazmi, Shabnam | 2012

677 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 45223 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Nobakhti, Amin; Thornhill, Nina
  7. Abstract:
  8. One of the most common problems encountered in multivariable systems is the interactions between the system inputs and outputs. Interaction causes difficulties in control of these systems. Hence reduction of interaction in multivariable systems is an important design objective. One way to do this is to employ a pre-compensator in the open loop systems. Several approaches with different design methods and optimization criterions have been proposed thus far. In this study the problem of interaction reduction for a square stable system is considered and a pre-compensator matrix is designed which reduces interactions in the interacting multivariable system.In this case the purpose is to use pre-compensation to obtain a compensated system which is column diagonal dominant. To design the so-called pre-compensator matrix the H2 norm of the interactions is considered, and this norm is transformed into linear matrix inequalities. However, since when using LMIs in design problems, the order of designed pre-compensator is at least equal to the order of augmented input system, this will lead to a high order pre-compensator. In combination with the loop controllers, this will lead to an overall high order controller. Therefore, the objective in this thesis is to design a pre-compensator which has a prescribed low order. For this purpose an order induction method is employed which does not ontain iterations and produces a sub-optimal pre-compensator matrix with the desired low order. This approach may be formulated as an LMI and this offers the chance to combine the problem with other LMIs for a multiobjective design. In the proposed method, the problem of re-compensator design is solved by adopting a novel method to deal with the rank constraint which arises as a result of the order specification. Comparisons are presented between the pre-compensators designed using the proposed method and those which are designed with a full order and subsequently order reduced
  9. Keywords:
  10. Compensation ; Linear Matrix Inequality (LMI) ; Order Reduction

 Digital Object List

 Bookmark

...see more