An Industrial Perspective on Control-Relevant Identification

System identification is of critical importance for implementing advanced control in the refining and petrochemical industries. As a result, control-relevant identification, which views the identification problem in terms of its interaction with control system design, is a topic that possesses significant practical impact. The circumstances that motivate control-relevant identification are discussed in this paper. These include the demanding conditions under which identification experiments are executed in process systems, as well as the requirement for sophisticated technology that does not require a highly skilled user to carry out the identification. Specifically, we examine prefiltering and uncertainty estimation as examples to demonstrate the requirements of control-relevant identification for process control.