Abstract
This paper investigates and improves a technique known as Nonlinear Dynamic Data Reconciliation (NDDR) for a real industrial process. NDDRS is a technique for data reconciliation that requires an objective function to be minimised subject to both algebraic and differential, equality and inequality constraints. These constraints are obtained from the mathematical description of the process and ensure that the measurement data can be optimised to conform as closely as possible to the true behaviour of the process. One of the difficulties of using the original NDDR is that a rigorous process dynamic model is required as a constraint. Unfortunately it is very hard to establish a rigorous dynamic model for a complex industrial process, particularly for data reconciliation purpose. A transfer function matrix model has been introduced in this new NDDR method. Therefore the rigorous dynamic model is avoided. The real industrial data from FCCU is used to illustrate the efficiency of the new NDDR method. Copyright © 2006 IEEE.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call