Which software is best for simulation of turbomachinery based on following three criteria:

1. It should be capable of simulating multistage turbomachinery.

2. It should be capable of simulating the multispool turbomachinery e.g one stage of high pressure turbine and one stage of low pressure turbine on different shafts with differnt rotation speeds. further more having opposite direction of rotation.

3. It should be capable of simulating the real fluid properties including variable Cp, K and viscosity.

Any suggetion would be highly appreciated.

Regards

Rolls-Royce PADRAM and Hydra

Dear Wang,

Thanks for informative reply. now how to get it? Is It availabe to every one in world?

Regards

no

Dear Wang, Then what to do next, any idea. what are terms and conditions to get these softwares? thanks

The codes of John Denton and Bill Dawes are highly regarded in the turbomachinery world. Both of these researchers where at the Whittle Laboratory at Cambridge University so you may be able to contact them through Cambridge University if you want to buy either of their codes.

I don't think that these codes use variable cp, k and viscosity, but if you are a serious user, you will get access to the source code and can add these features.

What about CFX? As far as I know that is regarded as the best commercial package for turbomachinery.

Anyway, good luck!

Thanks TurboGuy,

I think variable cp doesnt matter a lot. As by recent paper (2006 J. turbomachinery , some guy from Rolls Roys) has shown that by constant Cp and Varaible Cp have negligible difference on overall performance as comapred to the assumptions inhernt in Mixing plane model.

I Am using fluent and ther are many papers on turbomachinery who have used Fluent(for example 45th AIAA exhibit at Reno Naveda) but i am getting problem with fluent, It is underpredicting the effecinecy and i have no clue why this is so.

Therefore keeping in above points I have asked about the software which is only for and specilized for turbomachinery.

I am in urget help for my thesis.

Regards SAM

Hi Sam,

Correctly predicting the efficiency of a turbomachine is actually quite difficult.

One of the first things that you should ask yourself though, is how confident are you that the experimental result that you are comparing to is correct, as it is also quite difficult to measure the efficiency accurately unless you take great care.

As for the computations, ask yourself where the losses are coming from and then think about how well the model you are using will capture these features. Typically the losses will come from profile loss (boundary layers), tip leakage loss and losses from secondary flows. Are your walls well enough resolved to get the profile loss? Are you allowing the secondary flows the same amount of distance to mix out in the computation as you are in the experiment? Are you modelling the tip leakage?

Anyway, I would think that Fluent is likely able to predict the efficiency fairly well if you are comparing apples to apples and if your model is set up correctly. I would really make sure that you have tried every possibility with Fluent before looking at another code as the learning curve will be steep for the codes of Denton and Dawes.

If you get within a few percent, you are doing great!

Good luck!

I am resolving the tip clearence gap with 6-12 points and tip leakage flow and mixing of tip leakage flow and secondary flow is evident in computations. So i guess I have resolved the main featurs of flow correctly.

I have tried three different meshes.

1. Boundary layer with prims in the rest of domain. 2. O type mesh 3. C type mesh All meshes have nearly same resolution.

But the first mesh is predicting about 10% higher effecincy than the last two. what could be the problem?

Another problem is:

I am getting problem running two compressor or turbine with opposite rotation direction(counter rotating multi spool) and different rpm, I am not able to interate it, although i have tried all means to do so. Do you have any suggestion so that i would be able to run this configuration in Fluent 6.2 or 6.3.

3rd question: If I have option between Gambit(2.3) and Gridgen (v 15.08) then which mesher I should prefere?

Thanks SAM

Hi Sam,

I have no experience with Fluent, so can't really help you there, maybe someone else can step in?

As for different solutions on the different meshes, the first question that you may have already addressed is "are the solutions independent of any mesh refinement that you do?" Also, is the y+ value the same for all of the computations? This will affect the profile loss.

Secondary flow doesn't really cause any significant loss until the vortices mix out, this is why it is important to model the EXACT same domain as used in the experiments. For example, the downstream boundary of your CFD solution should be at the same axial plane as where the total pressure or temperature (depending on how you calculate efficiency) was measured downstream of the turbine.

I hope this helps.

Hi Sam,

I was actually just curious how Fluent ended up doing on your problem?