Model tuning for controller design
Conference Paper
Overview
Identity
Additional Document Info
View All
Overview
abstract
Practical systems are usually of high dimension. To precisely represent these systems, large models are needed. We shall call these large models evaluation models. The size of an evaluation model is normally too large for controller design. These evaluation models are developed for the purposes of analysing the systems and evaluating the controller designs. Model reduction techniques are normally used to reduce the models for controller designs. We shall call these reduced models design models. The process of model reduction introduces errors into design models which may not be ignored in the controller design. This paper presents a method to appropriately tune the design models and to properly adjust controller gains to tolerate these errors. This method is a systematic tuning scheme for the combined procedure of Modal Cost Analysis and Output Variance Constraint (MCA-OVC) model reduction and controller design. The scheme uses Modal Cost Analysis (MICA) model reduction method (Skelton et al., 1983, Yousuff et al., 1985) and Output Variance Constraint (OVC) controller design method (Skelton et al., 1985, Hsich et al., 1989, 1990) as the basic tools to tune the design models from the evaluation model and to tune the controller for high performance controller design. By tuning the procedure according to certain intuitive guideline rules derived from some properties of the MCA and OVC algorithms, the scheme literatively searches for high performance controller design. The paper is organized as follows. Section 1 presents the MCA model reduction method and the OVC controller design method. We will discuss tuning parameters of this scheme in section 2. The information needed for the tuning of the OVC controller design will be presented in section 3. In section 4, we will discuss how to tune the entire MCA-OVC procedure. A brief conclusion is given in section 5.
