A Proposal On Plantwide

A Proposal on Plant-wide Control of Industrial Processes Using a Model Predictive Control
(MPC)
ABSTRACT
The connections between the various units of a chemical processing plant need to be put into
consideration in control design because the behaviour of a complete plant is not only given by
its individual units. Though if the units of a plant are connected in series, it is easy to predict the
behaviour of a plant from the behaviour of the individual units, this does not imply that the units
can be operated like individual units. A plant-wide control strategy in which all the controllers in
the system are well integrated is thus proposed.
A plant-wide control strategy based on integrating linear and non-linear model predictive control
(MPC) is proposed. MPC is a very ample range of control methods which makes an explicit use
of a model of the process to obtain the control signal by minimizing an objective function. It
leads to linear controllers which have practically the same structure and present adequate degrees
of freedom. It intrinsically has compensation for dead times, and the resulting controller is an
easy to implement linear controller. The hybrid method is applicable to plants that can be
decomposed into linear and highly non-linear subsystems that interact through mass, heat and
energy flows. Linear MPC(LMPC) would be applied to the linear subsystems and nonlinear
MPC(NMPC) would be applied to the nonlinear subsystems. A simple controller coordination
strategy would be proposed to counteract interactions effect.
Background Information
In my undegraduate studies, I successfully completed my BSc thesis on Plantwide Control of
Hydrodealkylation(HDA) Process Using Aspen Plus and Aspen Dynamics Simulators and PID
control algorithm. Thus I intend to further research in this area, but now using MPC algorithm.
LMPC has been successfully applied to plantwide problems with hundreds of input and output
variables. NMPC has been used as an alternative to LMPC for plants with highly non-linear
behaviour. The key difference is that NMPC utilizes a non-linear model to predict and optimize
process performance.
Proposed Methodology
I intend to use a plantwide control strategy based on integrating LMPC and NMPC. The
motivation for this approach is that most operating units in a typical chemical plant can be
adequately described by linear dynamic models, while a small number of operating units account
for nonlinear behaviour. The plant would be decomposed based on the degree of nonlinearity.
Then the MPC system for the decomposed plant would be contructed.
I intend to present each stage of the research through internal seminars, conferences and finally
through publishing of the completed work.
Background of the work
The connections between the units of a chemical processing plant are important because the
behaviour of a complete processing plant is not only given by its individual units. Though if the
units of a plant are connected in series, it is easy to predict the behaviour of a plant from the
behaviour of the individual units, this does not imply that the units can be operated like
individual units. The output of one unit will act as a disturbance to the other unit; to the extent
that for even a system with simple connection, certain considerations need a perspective above
the unit operation. The issue of plant-wide control seeks to answer the question of how to
combine the controllers of the different unit together.
For instance the presence of mass recycle and heat integration changes the dynamic and steady
state behaviour of the plant in ways that are difficult to predict from the behaviour of the
individual units so that heat integration and mass recycle call for a plant-wide perspective of
control structure design. Plant-wide control simply put, refers to integrating the controllers of
different units of a plant (Larsson, 2000).
A better understanding of plant-wide control will lead to better design of control system. Better
control system will give plants lower energy consumption and better utilization of raw materials.
This is good for the society and the industry. The realization that the field of control structure
design is underdeveloped is not new. In the 1970s several articles were written on the gap
between theory and practice in the area of process control. The most famous is the one of (Foss,
1973) who made the observation that in many areas application was ahead of theory and he
stated that the central issues to be resolved by the new theories are the determination of the
control system structure, which variables should be measured, which inputs should be
manipulated and which links should be made between the sets. The gap is indeed present but on
a contrary view, the theoretician must close it. Many authors have pointed out that the need for a
plant-wide perspective in control is mainly due to changes in the way plants are designed with
more heat integration, recycle and less inventory. Indeed, these factors lead to more interactions
and therefore the need for a perspective beyond individual units. However even without any heat
integration, there is still a need for a plant-wide perspective as a chemical plant consists of units
connected in series and one unit will act as a disturbance to the next one.
The design of a typical plant-wide control structure consists of four major steps:
1. The overall specification for the plant and its control system are stated
2. The control system structure is developed. These steps involve :
i. Selection of controlled outputs ( variables with setpoints )
ii. Selection of manipulated inputs
iii. Selection of control configuration (a structure interconnecting measurements/set

points and manipulated variables)
iv. Selection of controller type
3. Design is followed by a detailed specification of all instrumentation/hardware and software,
cost estimation, evaluation of alternatives and the ordering and installation of equipment.
4. Following design and construction of the plant, plant tests including start-ups, operation at
design conditions and shut downs are carried out prior to commissioning of the plant.
Objectives
The control objectives for this process are typical for a chemical process;
1.
2.
3.
4.
5.
Maintain process variables at desired values.

Keep the process operating conditions within equipment constraints
Minimize variability of product rate and product quality during disturbances
Minimize movement of valves which affect other processes
Recover quickly and smoothly from disturbances, production rate changes or product mix
changes
6. Tune controllers to determine the parameters for the control loops of the process plants
using the auto-tuning approach in the ASPEN DYNAMICS simulator.
Scope of the Work
This work would focus on the plant-wide control of industrial processes using model predictive
control (MPC). The Control application for this work will utilize the first two steps highlighted
earlier in the background of the work.
1.4 Justification for the research
The behaviour of a complete Chemical processing Plant is not given by its individual units, the
connections between the units are equally important. The behaviour of a Plant with units
connected in series is easy to predict from the behaviour of the individual units. This does not
imply that the units can be operated like individual units. The output of one unit will act as
disturbance on the next unit and at steady-state; they must have the same through-put. For a
system with simple connection, certain considerations need a perspective above the unit
operation. An example is the placement of level controllers for a Plant with units in series. It is
exactly such a type of structural question that the field of Plant-wide control seeks to answer. In
addition, the presence of heat integration and mass recycle changes the dynamic and steady-state
behaviour of the Plant in ways which are difficult to predict from the behaviour of the individual
units. Therefore heat integration and mass recycle makes the need for a Plant-wide perspective
much more pronounced when the control structure is designed. A better understanding of Plantwide control will however lead to a better design of control system.
The control structure design problem is difficult to define mathematically both because of the
size of the problem and the large cost involved in making a precise problem definition which
would include e.g. a detailed dynamic and steady state model. An alternative to this is to develop
heuristic rules based on experience and process understanding. This is what will be referred to as
the process oriented approach.
Conclusion
The MPC would be applied to a highly nonlinear process e.g HDA plant or any other chemical
plant wth adequate dynamics, and compare it with the Proportional Integral Derivative(PID)
algorithm.

A Proposal On Plantwide

Uploaded by

Copyright:

Available Formats

A Proposal On Plantwide

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Proposal On Plantwide

Uploaded by

Copyright:

Available Formats

A Proposal on Plant-wide Control of Industrial Processes Using a Model Predictive Control

iii. Selection of control configuration (a structure interconnecting measurements/set

Maintain process variables at desired values.

You might also like