Simulation studies on normalized decoupling control system design for MIMO processes

Abstract

In this thesis, a new engineering oriented decoupling control system design method fore multi-input multi-output (MIMO) processes is presented. By employing the concept of integrated error, gain and phase changes of a transfer function when other loops are closed can be uniquely determined. Consequently, an equivalent transfer function matrix for closed-loop control system can be obtained and its relations with the original process transfer function matrix are derived. Based on the equivalent transfer function matrix elements, parameters of stable, proper and causal ideal diagonal decoupler can then be easily determined by proper parameter assignments of the decoupled transfer function matrix elements. The method avoids the drawbacks of existing decoupling schemes and is very simple which can be easily understood, implemented by field control engineers. These are five multivariable system process models studied to evaluate the effectiveness and simplicity of the proposed normalized decoupling control method. The simulation analysis shows that compared with the existing decoupling control methods, the normalized decoupling control method is very simple in design and effective in control system performance and can be easily understood and implemented by field control engineers.