A potential user asked:

  • I received a message from my E-mail about ASCEND.
  • I would like to know what it is:
  • Whether it can be used in developing
  • Chemical Engineering Process Modeling.
The following response is the opinion of Ben Allan. It
should in no way be construed as reflecting the views of
Carnegie Mellon, Arthur Westerberg, or anyone else involved
in the ASCEND project.


ASCEND is an interactive modeling environment, modeling
language, and solver interface for creating, debugging,
solving, and exploring nonlinear equations. The language
allows models to be built up from smaller models.

The primary use of ASCEND to date has been as a research
tool at Carnegie Mellon. Students have used it since 1989
to model nonideal distillation, batch distillation, ammonia
plants, ordinary and partial differential equations,
architectural structures, membrane reactors and so forth.
We make available free many of these models, particularly
the basic thermodynamic and distillation models.

Creating new mathematical models is still an art.
Discovering the correct specification of variables, good
initializations, good scaling of equations, and methods to
analyze results are supported by a variety of tools in
ASCEND. As such, ASCEND is suitable for developing
"Chemical Engineering Process Modeling" -- that is, after
all, what we do here for a living.

More information on the ASCEND approach to process modeling
can be found in a paper available by ftp as:

ASCEND Modeling

We will be updating our main web site Documentation section
soon to bring all the related theses and papers on line.

ASCEND is not (yet) a drag-n-drop flowsheeting tool, as
many of todays simulators are. These simulators hide (free
the user from?) too much detail to be efficiently used in
developing unusual models. ASCEND is probably not yet a
suitable tool for routine production modeling where the
user just plugs together icons for predefined units. The
underlying ASCEND solver and modeling language are designed
to handle just such a user, but the ASCEND GUI has no tools
for this sort of modeling. (It's pretty hard to get a ChemE
degree for working out graphic display algorithms).

Our primary goal as developers of ASCEND has been to show
what is really possible, to be a greenhouse of ideas for
the simulation software enterprise. The latest tools out of
several companies fairly reek with ideas pioneered in
ASCEND III and ASCEND II.

What we lack in user interface features we make up for in
portability and price: ASCEND runs nearly anywhere and it
is free.

ASCEND is an academic project, as such we encourage the
free exchange of ideas and reasoned, critical commentary on
those ideas. Many of the issues addressed by ASCEND revolve
around how best to address user needs when dealing with
complex information. To better obtain information design
insight, we need users who are willing to share the
adventure and offer thoughtful suggestions and criticism. I
believe that the user should never be _limited_ to the
tools provided by a GUI designer. Users are invited to
examine and modify, if needed, the ASCEND source code at or
(slightly beyond) their level of programming skill. Most
users will want to stay in our Tcl layer. Some will want to
venture into our interface (ick!) layer, some into our
solver layer, and (we expect) almost none into our compiler
layer.

If nothing in the list above fully answers your question,
please refine your question and try us again.

Thanks for your interest,

Ben Allan













Home | Download | Documentation | About | Contact


Copyright © 1997 Carnegie Mellon University. All rights reserved.
Send all questions and comments regarding this site to
ascend+www@edrc.cmu.edu.