Use of Autoassociative Neural Network for Dynamic Data Reconciliation
| Authors: | Thibault Jules, University of Ottawa, Canada
Bai Shuanghua, University of Ottawa, Canada
McLean David D., University of Ottawa, Canada |
|---|
| Topic: | 1.1 Modelling, Identification & Signal Processing |
|---|
| Session: | Nonlinear System Identification II |
|---|
| Keywords: | Measurement noise, data reconciliation, dynamic neural network, controller performance |
|---|
Abstract
The technique of dynamic data reconciliation has been previously studied in literature and shown to be an effective tool to validate process measurements corrupted by measurement noise, using information from both measured values and process models. Real-time implementation of dynamic data reconciliation involves solving complex optimization problem, leading to large computation time. This paper presents a study on the use of Autoassociative Neural Network (AANN) for dynamic data reconciliation. Once trained, the AANN can be directly used for online signal validation. Closed-loop performance of the AANN was evaluated for two storage tank processes.
