(79e) Integrated Design and Control of Heat Exchanger Networks (HENs) for Multiperiod Operation

Optimal synthesis of Heat Exchanger Networks (HENs) allows energy integration in a chemical process plant and reduction in energy costs. Poor HEN design or ineffective control strategies can result in the potential savings not being realized. Typical HEN design algorithms are tailored for a single set of operating conditions which include flow rates, target and source temperatures. Flexible HEN design on the other hand deals with designing HENs that can accommodate changes in the operating conditions [1]. Control objectives include deciding controlled variables and their optimal target values, tracking them by manipulating bypass valves and splitters [2]. Other issues include process dynamics, valve saturation, constraints on utilities, and transition between modes of operation.

Certain operating conditions are truly multi-period, especially if the operating or ambient conditions change in a seasonal or periodic manner. The design objectives include feasibility and optimality for the multi-period operation, while the control objective is to effect the transition optimally from one mode to another. The aim of this contribution is to study the interplay between design, control and operation of HENs for multi-period operation.

Depending on how the feasibility and optimality of transition is addressed, there are two approaches possible: 1) Decouple the design and control problems, or 2) incorporate control requirements at the design stage itself. Within the first approach itself, there are several design methods possible. One method of flexible HEN requires the HEN to be feasible for a certain operating range, by ensuring feasibility of corner points [3]. This can be extended to true multi-period operation by stipulating the true multi-period operating points to be the corner points of a feasible region. The control problem of optimally transiting from one mode to another and implementing the same can be solved independently. Thus, the problem of optimally transiting is decoupled: all paths that lie in the interior of the feasible region are required to be feasible and an optimal path is selected at the control stage.

On the other hand, in the second approach, the existence and optimality of a feasible transition path between the different operating modes is imposed upfront at the design stage itself. In addition, one can stipulate a particular form of control structure and mode of transition. These constraints are thermodynamic, structural and network dependent.

In the first approach, the final design might be overly conservative, since all that is required is the existence and optimality of a transition path rather than the more stringent requirement that all paths be feasible. While the second approach might be more complex, there is potential for substantial savings as the resulting HEN is not over-designed. These questions have not been addressed satisfactorily in literature. The aim of this contribution is to address the shortcoming in the current approaches through the case-study of a crude pre-heat train in a refinery and derive general conclusions.

References

1. Aatola, J., Simultaneous synthesis of flexible heat exchanger network, Applied Thermal Engineering 22, 907?918 (2002). 2. Lersbamrungsuk, V., Srinophakun, T., Narasimhan, S., and Skogestad, S. Control structure design for optimal operation of heat exchanger networks, AIChE J., 54 (1), 150-162 (2008) 3. Floudas, C.A., and Grossmann, I.E. Synthesis of flexible heat exchanger networks with uncertain flowrates and temperatures, Comput. Chem. Engng, 11 (4), 319-336 (1987).