New process integration methods for heat-saving retrofit projects in industrial systems
Doctoral thesis, 2005
New graphical methods for HEN retrofit (further development of existing tools), release of usable excess heat and placement of supply tanks in HWWS systems have been developed. In addition, further development of the matrix method for HEN retrofit has been made.
The methods have been used in case studies, both as stand-alone tools, and in different combinations.
Using the advanced composite curves in combination with the matrix method in HEN retrofit give new insights on which units should be picked for retrofit, and how far the heat recovery should proceed. The extra information that has to be collected to show the placement of existing heaters and coolers give valuable inputs to this method. Calculations showed that heaters and coolers placed close to the pinch are often cheaper by a factor of two and more to retrofit.
A method to show the potential to release usable excess heat at high temperature levels (but still below the pinch) has been developed. This method is based on the full set of stream data from the process, and therefore shows the total potential. Case studies have shown that much of the potential is also possible to release by making changes in the HEN with acceptable investments and PBPs.
Another method, to be used in HWWS systems for different setups of tanks and different values of DTmin has also been developed. This method shows the potential usable excess heat versus temperature in this sub-system for a user-defined number of supply tanks and DTmin.
By studying this sub-system, a large amount of the total potential can be exploited. This leaves the remaining process without any changes at all.
A new optimisation routine has been developed and implemented into the matrix method. This method uses a recursive branch and bound algorithm to explore all relevant paths in the retrofit task. A first solution is made and taken as an upper bound to the problem. All following paths that exceed this bound is directly eliminated. The implementation has proven to be reliable and fast to use, although there are still improvements to be made.
Heat Exchanger Network
Pulp and Paper
10.00 sal KB, Kemigården, Chalmers, Göteborg
Opponent: Professor Truls Gundersen, Institutt for energi- og prosessteknikk, Norwegian University of Science and Technology, Trondheim, Norge