14th World Congress of IFAC - special session

Robert Alcock (alcockrj@cf.ac.uk)
Sun, 15 Mar 1998 23:32:38 +0100 (MET)

The 14th World Congress of IFAC will be held in Beijing, P R China from
July 5
to July 9, 1999. The Beijing Congress, being the last IFAC
Federation of Automatic Control) Congress this century, will serve as a
forum for the international control community to review the great impact
of automation on the
elapsing century and to look forward to its development in the next
The Congress will cover a wide-range of areas for technical
including plenary lectures, survey papers, lecture papers, poster
sessions, panel
discussions and case studies. For more detailed information, please
check out
the IFAC'99 Web Page at: http://www.ia.ac.cn/ifac99/ifac99.html.

The Intelligent Systems Laboratory, at the Cardiff School of
Engineering, are
planning to organise a special session for the 14th IFAC Congress in
The theme of this session will be on Non-linear Optimisation,
and Control of Chemical Processes (NOICChe). A description of the theme
We welcome researchers in the field to contribute their papers on the

What To Submit ?

* an abstract including a title, 300 - 400 words and 3 -5 keywords
* the IFAC'99 NOICChe Abstract Submission Form (see below)

Please prepare the above in e-mail and hard copy.

Where To Submit ?

Send BOTH by e-mail and post mail to:

Xiu Ji
c/o Professor D T Pham
Systems Division
School of Engineering
University of Wales Cardiff
P O Box 688
United Kingdom

Tel: 0044-(0)1222-874429/874641
Fax: 0044-(0)1222-874003
E-mail: JiX@cardiff.ac.uk

Important Dates

* 31 May 1998
Abstract and IFAC'99 NOICChe Form submitted to Cardiff
* By 10 June 1998
Notification of acceptance of Abstract from Cardiff
* 15 June 1998
Proposal of the NOICChe session submitted to the 14th IFAC
* 30 November 1998
14th IFAC Secretariat Notification of Acceptance of Special Session

Description of the theme for a special session in 14th IFAC Congress

Chemical processes are non-linear, multi-variable, non-stationary and
time-varying. These complexities have long been challenging areas for
research work on
systems modelling, control and optimisation. Research efforts are
because improved process control results in the following benefits: low
product variability, higher value products, increased profits, reduced
utility usage and improved safety. These in turn are economically
important and essential in the competitive industries such as
distillation, refinery, crystallisation, fermentation, bioreaction,
polymerisation, etc.

Modelling, identification and simulation are routinely applied in modern

chemical industries for dynamic analysis and design of control systems.
Precise models of
processes play a key role in understanding dynamics, predicting future
and revising existing designs. Models represented by differential
equations give a good insight into the processes according to chemical
and physical laws governing the processes. Simplified and linearised
versions of the models can be directly used in
model-based control design such as internal model control and model
predictive control. Neural networks are powerful non-linear models. They
have been recently studied and applied to inferential modelling and
process modelling for quality variables in chemical
processes. They can be trained off-line or on-line. On-line
identification creates possibilities for adaptive control and learning.
However, research is still at an early stage. Therefore, papers on
non-linear modelling directly for the purpose of control are most
welcome for the session.

Due to the scale of the processes in the chemical industry, optimisation
can be
difficult and attractive. Optimisation and control are also related here
their ultimate goal is to achieve a possible optimal solution under a
number of constraints. The model-based predictive control strategy has
therefore gained some success in process control because it addresses
optimisation under constraints. Non-linear optimisation algorithms,
which are more efficient or can deal with non-linear
models, are specially useful within the framework of non-linear model or
neural model based control.

Optimisation in the theme of this session has another aspect. As
identification, control and non-linear optimisation together aim to
result in recognised methods for chemical process control, on-line
identification and optimisation become challenging issues. More
efficient algorithms are needed to tackle optimisation both in
identification and control. It is also important if an optimisation
algorithm could address process control and identification together and
solve them in an integrated manner.

With this session, we want to bring together recognised research in the
fields of:

1. Modelling and identification of chemical processes using non-linear
models and
neural networks;
2. Non-linear model based control and non-linear predictive control;
3. Optimisation algorithms and on-line optimisation;
4. Adaptive control and learning;
5. Neural, fuzzy and hybrid systems for chemical process control.

The techniques will preferably be illustrated on real or simulated
processes of the chemical industry.


Title of paper


Corresponding author's title and name

Mailing address