Hello world!
Welcome to WordPress. This is your first post. Edit or delete it, then start blogging!
Hacking ICEM CFD
I love CFD and I love Python. At my current job, we are using ICEM 10/11 a lot, so I really wished that ICEM and Python spoke the same language *pun*. Being that I’ve got thousands of spare hours on my hands *another pun*, I’ve started a project to develop a Python module that provides an interface to control ICEM. Some of the highlights are:
- Program ICEM via Python
- You don’t have to deal with TCL (whether this is an advantage or not is based on personal taste but TCL isn’t my favorite language in the world)
- Your script runs ICEM instead of ICEM running your script
- Interfaces are separated into classes: ICEM, geometry, mesh…
- Values are objects, not strings
- Installs via Pythons distutils
Currently the module supports import/export of geometries/meshes, translation/rotation of geometry families, unstructured mesh building (including prisms), smoothing of unstructured meshes. This is basically all that I currently need in my current job and it has really improved my workload. I still do geometry cleanup manually, but when I am happy with the cleaned geometry, I can run a python program to automatically generate a series of tweaks to the original geometry (a range of control surface deflections for example) and then build/smooth unstructured meshes. I can then go home for the night or be productive on something else at work.
I’d love to hear if anybody else is interested in this capability. My current work is owned by my employer (being developed on their time), but if enough are interested, I would be willing to re-implement it in my spare time and making it publicly available. I would especially love to have collaborators to help create a totally comprehensive library of functions in Python.
Back From The von Karman Institute
Just got back from a week long “Intro To CFD” course at the von Karman Institute in Brussels, Belgium. It was a great refresher since I haven’t looked at this stuff since graduating with my Masters in Aerospace in 1995. One of the coolest things about the course was having the opportunity to talk with John D. Anderson at the course. If you have been enrolled in an university level aerospace program, I guarantee that you have read at least one of his books (and will also bet that you actually liked the book).
The course of course covered the basics of Finite-Difference and Finite-Volumes methods as well as spatial/temporal discritization schemes. An interesting lecture presented finite element methods for CFD applications, touting the ability to advance towards higher orders of accuracy. For the time being it seems to be more academically interesting than practically useful in an industrial setting.
In the beginning…
We’ve heard that a million monkeys at a million keyboards could produce the complete works of Shakespeare; now, thanks to the Internet, we know that is not true.
-Robert Wilensky, speech at a 1996 conference
With this quote at heart, lets begin the curiously sparse blog!