TARGETING FOR PROCESS HENs DESIGN AND RETROFIT IN THE CONTEXT OF TOTAL SITES

Chemical processes normally operate on total sites served by central utility systems
that are complex in their operations. The utility systems supply the processes with
requirements such as; process heating, process cooling, power, water supply and
disposal etc. Site expansions such as introduction of new processes into existing sites,
or retrofit projects such as improving the heat recovery of existing processes usually
result in changes in process utility requirements. Introduction of new process into an
existing site normally requires a new heat exchanger network (HEN) to be optimally
designed for integration with the site. On the other hand, retrofit project in an existing
site may require an existing process HEN to be retrofitted to reduce its utility
consumption leading to possible debottlenecking of the process. In normal total site, a
process HEN may consume and/or generate steam utilities at various pressure levels
supplied from or to steam utility system through the steam mains. Thus, during the
incorporation of a new process HEN or retrofitting an existing one, concurrent
changes in the process steam requirement at several pressure levels may occur,
causing concurrent changes in the flow of steam in the respective steam mains.
Therefore, to optimize the utility systems operation and the process HENs design or
retrofit, a better understanding on these interactions in the context of total site is
required. In the previous approach for energy analysis and optimization of total sites,
the impacts of changes in process steam requirement on the utility systems operation
are assessed using steam marginal costs. For steam at a given pressure level, the
marginal cost is defined as the change in the steam utility system operating cost per
unit change in the steam load. The steam marginal costs are calculated based on the
assumption that the changes in the process steam requirements at the various pressure
levels occur successively instead of concurrently. This may not be true when real total
sites are concerned. Also, steam marginal costs are known to vary with the change in
the process steam flow.