The proposed method utilizes a frequency-domain optimization technique to find the optimal simplified z-transfer functions for a given higher order s-transfer function. The optimum values for the parameters of the prespecified reduced order discrete model are obtained as a result of minimizing the sum of magnitude-squared error function. Examples are considered to illustrate the application of this method, compared to other methods, in system identification, discretization, and model reduction, for both small and large sampling rates.
discretization, continuous control systems, model reduction
For both simulation studies and design with implementation of computer control systems, discrete models with reduced order for complex continuous process control systems are frequently required.
There are two general approaches to implement discretization and model reduction starting from higher order continuous modelsL In the first approach, higher order continuous models can be reduced using one of the available techniques 2, and then the resulting loworder models can be discretized using one of the well-known existing methods, such as numerical integration, pole-zero mapping, or the hold-equivalence method. In the second approach, higher order models can be discretized using one of the above methods and the discrete model obtained can then be reduced. Obviously both approaches are equivalent and yield identical results. The first approach would seem to be computationally more efficient. However, the second approach seems to be more amenable to practical use.
Several methods have been developed in the literature using the above approaches for computing equivalent