(496b) Using Dynamic Flexibility Analysis to Integrate Design and Control under Uncertainty

Abstract:

Daily
process operations are impacted by several short and long-term uncertainties
like daily fluctuations and seasonal variations in production levels, such as
feed compositions or change in services quality. In order for the process to
handle deviations from nominal condition the effect of uncertainties should be
adequately incorporated in the conceptual design phase. In addition to
parametric uncertainties, uncertainty associated with physical properties and
process models add a difficulty to the design problem. Therefore, classic
deterministic design based on information at nominal conditions alone may not
lead to best process performance in the actual industrial plant. In industrial
practice overdesign based in engineering judgment aims at increasing the
robustness of a design.

Recent
progresses in both theoretical approaches as well as computer power have
renewed the interest of many researchers in academia and industry to explore
process design under uncertainty with more mathematical rigor. A problem that
has not yet been well studied is the distinction between process design versus
a controlled process design. The usual practice is design for the worst-case
scenario and in a later stage design a control system to handle the
uncertainty. This practice usually leads to a robust design but the trade-off
between design and control is not exploited leading to non-optimal results. Clearly,
this viewpoint brings together a very important aspect, namely the integration
of design and control. Unfortunately little systematic control and design of
processes under uncertainty is available. Grossmann and co-workers [e.g.
Halemane and Grossmann, 1983; Pistikopoulos and Grossmann, 1988] introduced the
concept of integrating design and control to obtain best trade-offs between
cost and process flexibility considering a steady state assumption.
Pistikopoulos and coworkers [e.g. Pistikopoulos and Dimitriadis, 1995 and Bansal,
Perkins and Pistikopoulos, 2002] have shown that considering a steady-state point of view renders an
unrealistic control scheme and a dynamic analysis is needed.

In this
presentation, we will propose a novel methodology that aim at obtaining best
trade-offs between design and control decisions in a dynamic view of process
control and design. Our methodology will include the concept of flexible design
of controlled systems under uncertainty. We will also demonstrate, with the
help of this dynamic approach, that integration of design and control at
conceptual level yield better cost performance and higher flexibility as
compared to designs, which consider process control separately. In particular
we would like to study the impact of periodical uncertainty and the influence
of their frequency of occurrence. We will compare the advantages and
limitations of our methodology to different deterministic and probabilistic
uncertain design approaches using static-control and illustrate our methodology
with the help of benchmark case studies.

References:

Halemane, K. P; Grossmann, I. E.: Optimal
Process Design under Uncertainty, AIChE J. 1983, 29, 425.

Pistikopoulos, E. N.; Grossmann, I.
E.: Optimal Retrofit Design for Improving Process Flexibility in Linear System,
Comp. Chem. Eng. 1988, 12, 719.

Pistikopoulos, E. N.; Dimitriadis, V. D.: Flexibility
Analysis of Dynamic Systems, Ind. Eng. Chem. Res. 1995,34, 4451-4462.

Bansal, V.; Perkins, J.D.; Pistikopoulos, E. N.: A Case
Study in Simultaneous Design and Control Using Rigorous, Mixed-Integer Dynamic
Optimization Models, Ind.
Eng. Chem. Res. 2002, 41, 760-778.