Study On Two-degree-of-freedom Internal Model Control And Decoupling Internal Model Control

Internal Model Control which is evolved from the1950’s Smith predictorcontroller has been developed more than half a century and formed a relativelycomplete theoretical system. Actually, Internal model control is a kind of robustcontrol. It has the following virtues, such as the speed of response is fast, the amountof calculation is small, the performance of robustness is excellent, be good at dealingwith the process interference. The main contents of this thesis are twodegree-of-freedom internal model control and multivariable V-norm decouplinginternal model control. The main topics of this thesis involve:1. Based on a review of internal model control, we gather the design of controller,the improvement for filter, the two degree-of-freedom structure, single-variableinternal model control, multivariable decoupling internal model control andcombination with others advanced control algorithms.2. A two-degree-of-freedom IMC control structure is used for the first-orderdelay unstable process. At first, a feedback proportional controller is applied to meetthe objective of stable generalized second-order plus delay time with over-damp. Andthen, two-degree-of-freedom strategy is used to design the tracking controller andrejected controller. The set-point response and the disturbance response arecompletely decoupled and adjusted independently to achieve the satisfying controleffect. Simulation examples show that in the case of the large parameter perturbations,it still can achieve the desired performance quickly.3. A new V-norm decoupling internal model control method is proposed with theemphasis on the matrix form used to express the decoupled part, take advantage of thesmall amount of calculating V-norm. As for the delay unit, the use of first-order padeapproximation is a solution. Compared with existing methods, it shows that the program has obviously advantage in load disturbance rejection and anti-overshoot.4. For the non-square multivariable system, generalized inverse method and mainmanipulated variable method are used. In the first method, the generalized inverse ofnon-square system is used to design IMC controller. In main manipulated variablemethod, one input is selected to be a master, while the other inputs influence to theoutput is compensated.