Optimal input and decision in multiple model fault detection
| Authors: | Simandl Miroslav, University of West Bohemia in Pilsen, Czech Republic
Puncochar Ivo, University of West Bohemia in Pilsen, Czech Republic
Herejt Petr, University of West Bohemia in Pilsen, Czech Republic |
|---|
| Topic: | 6.4 Safeprocess |
|---|
| Session: | Active fault diagnosis/ Decision system design |
|---|
| Keywords: | fault detection, optimal experiment design, optimal detection, dynamic programming, loss minimization, detection systems |
|---|
Abstract
A design of detection system using closed loop information processing strategy in multiple model framework is treated. The proposed detection system should simultaneously generate decisions and auxiliary input signal which can improve fault detection. The design of detection system is formulated as an optimization problem based on the assumption that the measurements will be available in the future. Then the dynamic programming can be utilized for the minimization of a considered criterion. A significant special case where the relation between the auxiliary input signal and the decision is given a priori is analyzed.
