Page 62 - Sustainability in the Process Industry Integration and Optimization

Pro c ess O p timization 39

Another popular technique in process design and software
development is modularization or encapsulation. The complexity
management efforts in information technology and process modeling
led to development of the concepts of object-oriented modeling (see,
e.g., Modelica, 2009a; UML, 2010) and object-oriented programming
(e.g., C++, C#, Java, Delphi). With these modeling concepts, a number
of related objects or operations can be grouped together and
represented as a single object or operation. Similarly, flowsheeting
and simulation software may offer the option of representing a
distillation column as a single operating unit at the level of the entire
flowsheet while still allowing simulation of the column at the local
level; such a facility is available, for example, in gPROMS (PSE, 2009)
and HYSYS (HYSYS, 2010).
The key to object-oriented thinking is the principle of information
hiding, according to which every object should conceal as many details
as possible about its own functionality and provide to other objects
only the information relevant to interactions with them. Hence, the
interface describes the bundle of information items that are defined
as being available to an external object. An example of applying
information hiding in water network design is the abstraction of the
detailed information about water-using operations into only three
relevant pieces of information for each operation and contaminant:
the limiting inlet and outlet concentrations and the limiting flow rate
(or contaminant load). This bundle of information thus constitutes the
operation’s interface to the water minimization problem.
Efficiently managing the model and problem complexity is greatly
facilitated by the practice of documenting the complete modeling
process and optimization results, including their interpretation. It is
best to document all stages, starting from the conceptual modeling
and ending with the computational implementation and obtained
results. All this documentation should be systematic so that the
reasoning and results can be clearly understood and traced back to
their roots. This style of documentation is extremely useful and makes
the work of teams and individual engineers smoother and more
efficient in the long run.
Another important tool for managing complexity is targeting,
which reveals limitations in the underlying design or operation task
to which the optimization is being applied. With proper targeting it
is possible to obtain an upper bound on system performance and/or
a lower bound on system cost. In fact, it is also possible to calculate
practically achievable targets:
• For water systems, current targeting practices mainly yield
the first type of estimate: the maximum possible amount of
water reduction.
• For HEN synthesis, the Maximum Energy Recovery (MER)
targets can be established, and HENs achieving them also