I am a graduate Mechanical Engineer and also have a Masters Degree in combustion. I work for a burner manufacturing company and am about to undertake a PhD sponsered by my company. The subject of my PhD is the causes of combustion resonance in multi port burners. The company wants a deliverable of a CFD package to model the flame on a burner in order to predict weather it will resonate or not.

I have two options:

1. I could carry out experimentation to find the causes in this case of combustion resonance. Then obtain a commercially available code such as fluent, and adapt it for my purpose. I understand this would mean that i am not a CFD expert, but i am more interested in doing the right thing for the company.

2. I could write the code myself, as this would enable me to fully understand the workings of the code, and be able to modify it in the future if needs be. Its also cheaper.

I have only basic knowledge of programming fortran and pascal, but am a very capable combustion scientist. I have no idea of the difficulty of writing such a piece of code, could anyone tell me if this would be possible assuming i work at it, or is it just too much? If I do go ahead with it, whats the best starting point for this project?, i.e. good books in CFD, computer programming etc..

I work on burner and furnace design using FLUENT, I'd like to make a comment: beacuse of your expertice on combustion and burner design, and the supportive engineering data you obatined from experiments, you can make a good use of commerical codes for your work. Of course, you can write a CFD code to cover all aspects of flow, turbulence, combustion, emissions, etc., but you have to be very patient for your Ph.D. study period, 4 years, 5 years, maybe 6 years. I am only a CFD code user, I spent three years try to finish my dissertation on enhanced design of burner and furnace, it seems that I have to stay for another 6 months to finish it.

This is just a personal opinion of my own, maybe other people could give their valueable opinions.

Guosheng Kang

I'd say that research in combustion oscillations coupled to resonance requires access to source code and also a good research environment - good supervising, access to people with previous experience from code devloping, access to people with hands on experience from the type of problems you are trying to predict etc. It is not something I'd want to venture into on my own. You'll probably want to experiment with different linearized models in order to efficiently predict eigenmodes of the system etc. This is something that will be very difficult to do in a commercial code.

If Jonas is correct and you need source code, you might wish to take a look at KIVA. This code and its parents have a couple of decades of development. It has been (mostly) developed for internal combustion simulation, first at Los Alamos and lately at U of Wisconsin and Cummins Diesel among others.

I believe that the source is available free, as are a series of LANL reports. There is a user community with active e-mail discussion. Take a look at

http://gnarly.lanl.gov/home.html

Good luck!