|
[Why a Consortium?]
[Executive
Summary] [Partnership
Structure]
[Program
Achievements]
[Research Directions of KA Hoo]
Research Directions
The major research directions of our program
are summarized below.
Data Fidelity
(see Statistically Based Techniques)
All model validation and controller strategies require data of high
fidelity. The research is aimed at finding better methods for addressing data
reconciliation, outlier detection, process monitoring models, and statistical
metrics.
Study the Most Challenging Industrial Process
Control Problems
A major emphasis of our program has been to
study the most challenging industrial process
control problems. (see Transition Control).
System Identification Theory
(see Distributed Parameter Systems)
Data driven techniques such as the Karhunen-Loeve expansion, singular value
decomposition, artificial intelligence have the
potential of developing models that are suitable for linear
and nonlinear model-based control strategies.
Unit-Wide and Plant-Wide Optimization
The current industrial approach to
plant optimization is to use rigorous models for
all units. Our efforts are aimed at developing
procedures that use the minimum necessary
complexity in the optimization process models.
Chemical Process Design for Operability & Controllability
Design is a steady-state exercise, however the operation of a plant is
dynamic. The objective is to integrate existing flowsheet designs with operability and controllability
considerations so that optimum steady-state design
conditions can be achieved and real profits can be realized.
[Why a Consortium?] [Executive
Summary] [Partnership
Structure]
[Program
Achievements] [Research
Directions of KA Hoo]
|