文摘
The identification of a process model is a critical step in the implementation of model-basedcontrol schemes. Process identification tests for multiple-input-multiple-output (MIMO) controlproblems can be quite time-consuming and disruptive to the process. Thus, there is significantmotivation to reduce the duration of the identification tests. In this paper, we compare alternativemethods for conducting the plant tests. A detailed simulation case study is used to evaluate theadvantages and disadvantages of various input excitation strategies in terms of the accuracy ofthe estimated parameters, prediction errors, process gains, and relative gain arrays.