Study Of Multivariable System Modeling Technique

At present, Single variable control has been already difficult to meet demands of actual production, so has launched the research on multivariable control extensively. If we want to control multivariable system, especially carrying out some advance control algorithms such as internal model control and predictive control that are based on model, system model is prerequisite. This paper is to study the technique of multivariable system modeling.The development and achievement of multivariable system modeling technology are discussed at first, and then the benefit and deficient of the existing modeling methods are compared. Complexity of The traditional algorithm will increase a lot when expanded from single variable to multivariable, mainly because of the decision of the structure of multivariable system. Subspace is a kind of new identification algorithm that base on the state space model. So model structure is simple and only need to know the system order. And there is no difference when applied on single variable and multivariable system. So this text carries on research to this kind of algorithm mainly.