Dear all,
I am writing to you to ask for your personal preference on which CFD
software package I should use for all the following (as one) and your
reason why;
A) Large Scale City Analysis of wind turbulence and velocity
B) Internal and external analysis (combined)
C) CFD for Urban Design.
Currently we use IES-VE for CFD analysis but from what I've heard, the
general consensus is that, it may not be near accurate enough. In
addition, the computational time is far too long.
Regards,
Ronan Carney
JRR wrote;
Look up Blue Ridge Numerics in Charlottesville VA. They use
bi-linear bricks (elements) to
get near quadratic performance with minimal effort..... you're going
to need a Linux Cluster;
There was a DOE bid due Monday at NREL for developing models of wind
flow better than
the Raleigh ?? distribution with height that we have now.
CFD AnalysisHi, Carney
If you want to do those jobs, you'd better to have a cluster or at least a work station to minize the computing time. There are a lot of commercial CFD software which can fulfill your purposes, like FLUENT and PHOENICS etc.. I do not ever try IES-VE though.
Hope this can help.
Zhuolun Chen
Carney,
I'm also among the group of people who are not too comfortable with CFD
results from tools like IES-VE and Designbuilder. These are primarily energy
modeling tools with built in CFD applications. However they provide a
seamless environment (for Energy modeling and CFD) and have easy to
understand user interface but I feel the main problem is with 'automatic
meshing' that these software offer. Meshing is a critical component of the
CFD process and it would be hard to automatically generate right kind of
mesh for infinite variations of geometry and operating conditions.
To obtain correct CFD results we must check mesh convergence, may be these
tools have algorithm which create very fine mesh to start with and hence
they take a long time to produce results. Improving hardware capacity like
JRR and Zhoulum suggested is essential but its also critical to create right
kind of mesh to get quick results. This is especially critical for building
application since the volume of analysis zone is usually huge, if we mesh it
improperly then we can end up waiting for an year even on 32 machine cluster
or a powerful Xeon server.
In my opinion also stand alone CFD tools like Fluent, Phoenics, CFX, StarCD
etc. are much better CFD tools than IES-VE. Even if you would like to
use IES-VE it would be prudent to test a few cases of your analysis and
compare results with established stand alone CFD tools.
There are also CFD tools specific for building applications like
Fluent-Airpak and Flovent which are customized for building applications.
These tools are easy to understand for building engineers and produce good
results. Another good option can be OpenSource CFD libraries like OpenFoam,
FDS etc. However the problem there is that learning curve with OpenSource
platforms is rather slow.
Regards
Shishir