Geometry of Stage 37
 

Hi guys. After my bloody struggle ;) with Stage 37 I want to share with you results that I've got and, maybe, discuss them (if someone obtained something else)

First of all is 3D geometry of Stage 37.

http://uploading.com/files/1be6418c/Geometry.zip/
Geometry.zip - 421.4 KB
<a href="http://uploading.com/files/1be6418c/Geometry.zip/">Geometry.zip - 421.4 KB</a>

I'm not sure what link should I post so I've put all of them. :o

Hope that this will be useful because when I started this project I haven't found 3D blades geometry. So, geometry of rotor and stator was created in Solid Works 2010 basing on report NASA TP-1337. I took all points from the report, created curve from them and lofted a surface on those curves. You can look for details in files.

Second is performance curve comparison. I've compared experimental curve with the calculated one and found that it had lower pressure ratios and MFRs. And, of course, I thought that I've done something wrong. But, after I found an article "CFD SIMULATION OF FOULING ON AXIAL COMPRESSOR STAGES" from ASME Turbo Expo and there they got similar results (i attached screenshot with comparison of experimental performance with the one from the article and my one). As you can see the article authors got almost the same results as I did.

What do you think about these results? Maybe someone got better coincidence with experiment?

Regards,
Oleg

Well done on your progress. I remember your old posts and now you have improved a lot like Champ;).

Thanks for sharing the geometry.

However you can improve results further as I did. I have compared my results to EXp and LES and they are within 1% to max 2 % for very few cases. You should also compare the pressure, temp and efficiency profiles at outlet to experimental data.

Looks like your question is covered in an FAQ:

http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

Thanks for the praise.

Concerning geometry. I've also tried to use Rotor 37 geometry from Turbogrid tutorial and it gave much better coincidence with the experiment. But I'm not sure how does this geometry was created. That's why I've decided to proceed with the one that I created because I was sure that it was created 100% basing on report data. And what about you? Have you created it by yourself or took it somewhere?

As for pressure, temp and efficiency profiles - I'm doing it now.

Regards, Oleg

Thanks for attention to my thread. But, as i mentioned in my reply to FAR, some incomprehensibility (at least for me) exist in "right" geometry. I've attached map where I added performance curve for case where I used geometry from TG tutorial (all settings were the same except rotor geometry). As you can see it fits experiment much better. If it is not too much trouble say, what do you think about it?

Regards, Oleg

Did you check the CG of each section and what about the stacking line? Did you correctly place the each airfoil section at their respective stagger angle and inlet and outlet angles are also correct? Import both meshes into the ICEM CFD and see if both have same height, same tip clearance, same curvature,same lean and sweep !!!!

Check the flow features and point-out the area of deviation and think why?

From the graph it seems that area is reduced which is evident from the chocking flow rate (reduced in your geometry) and which has direct effect on pressure ratio. As pressure ratio is directly proportional to mass flow, rotational speed and flow angles at in and out.

are you using mixing plane or frozen rotor model? What about the gap between the rotor and stator?

Hi,
I agree with Far's conclusion that the area is lesser, however I would say that blockaeg is higher in the second case. Since in both the models the simulation parameters are exactly the same, I think geometry might be the answer. Since you are using Solidworks, you can import both geometries in it and see the difference ( I think it might be a bit difficult to do it in ICEM) The error might be in either stacking ( As Far said) or in Stagger ( I am a bit biased towards this one:p). Check the geometries of both rotor and stator. Also see CFX help. They might have some info about the source of the geometry, which might be a revised/different versions from your cases.

Hi guys,

I've already done the comparison of both geometries (because only geometries were different) and of course they differs from each other. As for stacking, I took stacking point from NASA report and then put it on stacking axis (radial line). After that i rotate profile on stagger angle from the same report. As for geometry from turbogrid, I' been trying to find out how it was created but still doesn't succeed.

So the difference in geometries is the main cause of differences in results. An I keep trying to understand what have I done wrong.

Or maybe I haven't done wrong anything. Because, as you can see, in the article that I've found, authors got almost the same results as I do. And they are claiming that the geometry was taken from the same report. So as i see it there is 2 options:
1) or the guys from that article from Turboexpo created geometry as wrong as i did.
2) or the geometry in TG tutorial was taken not from this report or maybe was slightly adjusted to fit results better ;)

Unfortunately this article is the only one that i've found about comparison of Stage 37 with experiment (all other ones are about comparison of rotor 37 only).

Dear FAR! What about geometry on which you'd calculated Stage 37? Have you created it by yourself or had taken it somewhere?

Hello,

I'm also trying to simulate the rotor 37 geometry. I used the NASA TP 1337 for the coordinates of the rotor. I compared our both geometries and I noticed, that they differ quite alot, in particular in length (see figure 1)

Figure 1:
http://imageshack.us/photo/my-images...mparisona.jpg/http://img141.imageshack.us/img141/4...arisona.th.jpg

Figure 2:
http://img220.imageshack.us/img220/3...ialflow.th.jpg


Furthermore, I noticed, that the flow path becomes narrower in the axial compressor (see figure 2). However, it looks like this is not considered in the rotor geometry. If I set my stacking line perpenticular to the axis of rotation, my rotor always cuts the shroud. Is there some kind of angle of the stacking line? I still have problems fitting the rotor into the hub/shroud.

Can you share the geometry of the TG tutorial? If someone is interessted, I could post my geometry as well.

With regards

Martin

Hi Martin,

Frankly saying, I'm glad that someone except me has confusion on right geometry of rotor 37. Please find rotor 37 turbogrid geometry attached.

I'm not God of SolidWorks so i still thinking that, maybe, i've done something wrong.

Hope, maybe you'll find the truth.

And of course it would be great if you post your geometry.

Regards, Oleg

Two questions arises from this comment:

1. In first case (from hub side) it may be due to fact that you need to subtract the extra length by the hub surface, since the limiting case is not the blade itself but hub.

2. In turbo grid file (from casing side and tip clearance) , due to algorithm, they have extended the rotor blade by 0.0356 cm (equivalent to tip clearance gap, now in new research it is 0.04 cm, but the overall height of passage remains same). TG algorithm then subtract the blade from the casing to the required tip clearance height. So it makes the tip clearance by subtracting the material from blade and does not change the casing the hub radii.

From Fig. someone may get the feeling that blades are also different, but it is due to misalignment. First match the hub profile for both blades then you can figure out the difference in overall blade. Otherwise it is not possible to comment as it does not make any sense to compare the two geometries with different reference.

Did you compare the rotors only and what is the outcome from this analysis?

So first, I generate the complete blade geometry and delete the parts cutting the hub and shroud (considering the tip clearance)? That's the way I'm currently trying to make the geometry.

I've compared performance calculation of whole stage as i mentioned in the beginning of this thread. But i haven't compared geometries because, unfortunately, i dont have geometry from that article.

did you find the reason for difference in geometry (if any) and results?

You can combine both parts in an assembly in Solidworks.
http://www.solidworks.com/sw/resourc...assemblies.htm

Thanks for advice, but i've already done that. And as i mentioned before they really differs. The thing is that i don't understand why it happens.

I just constructed rotor 37 again using solidWorks only. I assume, that the reason for the differences are the creation of the radii at the leading and trailing edge.

I imported all points into SolidWorks. Afterwards I created an arc at the tailing and leading edge with the last and first point of the manufacturing coordinates as maximum and minimum, respectively. Then, I use splines to connect both arcs using the rest of the imported points:
http://imageshack.us/photo/my-images/854/solidw.png/


Acctually, the length of the rotor is given by the manufacturing coordinates. So there should be no variations.

And have you tried to compare the geometry that you've created with turbogrid's?