Design of heat exchanger networks with good controllability

Rapport, 2018

One important way to improve the energy efficiency of chemical process plants is to improve the heat integration within and between industrial processes. This is accomplished by recovering excess process heat at high temperatures and using it to replace primary heating at lower temperatures through a heat exchanger network. However, as a process becomes more heat integrated, process control may become more difficult. Poor control performance can, in turn, easily lead to increased costs that outweigh the predicted energy cost savings. It is therefore essential to model and analyze the effect of the process changes and address the identified potential control challenges. However, the majority of existing methods for controllability assessment of heat exchanger networks only consider steady-state properties, and not the dynamic aspects, which in reality can seriously affect process control characteristics. With better methods for controllability analysis alternative design options could be evaluated and compared more reliably at an earlier design stage.

This report proposes the basic structure of a step-wise approach for integrating dynamic considerations into the design process for heat recovery improvements in process industry, and suggests suitable methods and tools to be used for the different steps of the proposed framework. As part of this, recent work that has been performed to evaluate and improve the methods used in controllability assessment is outlined. Additionally, a number of areas are identified in which significant further efforts are required before a complete controllability assessment framework can be specified and a toolbox for integrated design and controllability analysis can be developed.

One central area requiring continued research and development is to define an adequate controllability index for use in heat exchanger network design. For example, it is relatively easy to argue that some of the commonly applied controllability measures are insufficient since they are based on steady-state system interactions only. However, as illustrated in the report, alternative measures of system interactions that take dynamics into account suffer from other drawbacks, of which one is scaling dependency. Nevertheless, these are interesting for further development of a new controllability assessment method, since the issues with scaling can possibly be dealt with using an approach evaluated in this project. Another area where further work is needed is to develop tools with some level of built-in support for formulation of dynamic models of heat exchanger networks. Model simplifications, or other means of handling the large model sizes typically resulting from dynamic modelling of heat exchanger networks may also be needed in order to overcome difficulties in model simulation and analysis. In addition to the development needs related to individual assessment steps, there is an apparent need for appropriate protocols for information transfer and conversion of models between different tools.

This report gives an overview of insights revealed in recent research with respect to the controllability of heat exchanger networks. Through this research, the knowledge for continuing the effort to define a better controllability index for heat exchanger networks has been improved.