-------------------------------------------------------------------
Tid: Fredag 29.nov. kl. 13.00-14.30
Sted: Lunsjrom, institutt for kjemiteknikk (Kjemiblokk 5, 2. etg)
Professor Costas Pantelides, Centre for Process Systems Engineering,
Imperial College, London, UK: "Process Design Revisited"
-------------------------------------------------------------------
Pizza e.l. serveres kl. 13.00.
Foredrag begynner kl. 13.30 (45 min pluss 15 min diskusjon).
Denne gang kan vi by et internasjonal kapasitet innen ommradet design,
simulering og numerisk analyse - Professor Costas Pantelides fra
Imperial College. Hans Ph.D. med Roger Sargent gikk paa videreutvikling
av Speedup, og dette har etter hvert utviklet seg til programmet
gProms (se http://www.ps.ic.ac.uk/gPROMS/).
I foredraget vil han snakke om sin aktivitet innen design, men vi planlegger
en liten rundbordsdiskusjon om gProms etter seminaret (ca. 14.45-16).
Hvis det er noen som onsker aa snakke med Costas mens han er her (fredag
eller evt. lordag) saa vennligst ta kontakt med undertegnede eller
med Invild Lovik (ingvildl@kjemi.unit.no) som er med og organiserer
programmet hans.
-Sigurd
----------------------------------------------------------------
Abstract. "Process Design Revisited" by Costas Pantelides
This seminar presents a recently developed methodology for the design of
continuous processes operating at steady-state. It is shown that a formal
mathematical description of the problem of designing a given process may be
constructed in a completely automatic manner from a knowledge of the potential
available raw materials, the properties of the desired end products, and
mathematical models of the available technology (e.g. unit operations). This
contrasts with earlier approaches in which considerable additional engineering
input is required (e.g. in deriving a suitable "superstructure" for the unit
operations under consideration).
The above approach is directly applicable to any combination of unit
operationfor which suitable mathematical models are available. It is
shown that, provide these models fulfil certain minimum detail conditions,
the combinatorial
complexity of the process design problem is much smaller than previously
thought. Moreover, the use of detailed models that accurately characterise
process performance and economics removes much of the traditional distinction
between "process synthesis" and "process design".
The design of processes from elementary building blocks instead of more
compleunit operations is also considered. It is argued that the main
issues in this
context are the formal characterisation of such building blocks and the
conditions that they must satisfy. Once a suitable set of building
blocks is selected, the remaining design problem can be addressed directly
by the above design methodology.
The seminar is illustrated with several examples of process design problems
involving reactor networks, separator networks, and complete processes.