Showing results for 
Search instead for 
Do you mean 

Correlation of cylindrical shell



is there any possibility to do find out MAC for cylindrical shell between FEA and experimental modal analysis file with cylindrical coordinate system.


I have tried this but when i am importing the made shape, FRF, Geometry UNF file the mode shape is getting distorted due to the change in the coordinates system.


plz help me 


thank you



Re: Correlation of cylindrical shell



So from the sim, you right-clicked a New Solution, set the Solver to Modal Test Data and when you displayed the mode shape in Post, it was distorted?


If so, where do the geometry and shape .unv files come from, are they from the same system and can I please see them?



Philippe Tremblay

Re: Correlation of cylindrical shell

Thank you  for replying 


The unv files are generated using model testing software[Mescope] and iam adding the geometry, FRF, and Shape in the Model Test data solution window.


i have performed the model testing using a cylindrical coordinate system.


if i see the mode shape in MEscope they are matching to the FEM mode shape


but when i import them in the Model test solution window there i am seeing an distorted mode shape which is not matching at all


i have tried this for a simple plate model it is working fine. 


the shell is of 350mm OD and 1.5mm thick. 864mm length 


encl: Mode shape.unv, geometry.unv,FRF.unv, Mode1.Jpeg, Mode2.Jpeg




Thank you

Re: Correlation of cylindrical shell



Your mode shape results don't make sense. From the analysis shapes, you have a constraint at one end of the cylinder, which should be at sensors 1 thru 6. Hence there should be no radial acceleration there.


Comparing the response at sensors 1 and 19:




The accels are complex. There are results in X only, I compute the magnitude of the complex results directly from the UFF file:




For mode 1, the mag of sensor 1 is half that of sensor 19. And for mode 2, its about the same, in fact for this mode the max magnitude is 0.32. So for some reason, MeScope generated poor data.



Philippe Tremblay

Re: Correlation of cylindrical shell

Dear sir
I would like to add few clarifications regarding the discussion on MAC corelation of cylindrical shell between FE analysis results and Experimnetal modal analysis results
1. The Boundary conditions considered in the FE model were Free-Free. In experimental modal analysis these boundary conditions were simulated by elastic rope.
2. The mode shpaes in FE matched with Experimental modal analysis mode shapes when viewed indepndently, however to correlate i want to estimate the MAC number bewtween the two. In this process i have exported Experimental modal analysis results from ME Scope to NX CAE, in uff format. The experiment was performed by Roving Hammer method and using cylindrical coordinate system. The excitation was in raidal direction and response was mounted at the point no:37. in Z-direction (totally there are 48 points condidered)
Now on comparison of the two results, in NX CAE, i am seeing distorted mode shape for the experimental mode shape and not matching to the one observed in MESCOPE prior to importing in NX CAE.

I am attaching the imported uff files and NX simulation files. Kindly request to help me..

Thanking you
with regards
Karthik .A

Re: Correlation of cylindrical shell


Can you import the uff you created, back into MESCOPE and try to animate those shapes? If so, do they look good?



Philippe Tremblay

Re: Correlation of cylindrical shell

Dear sir

i have imported the uff files back into ME Scope and required to define the animation equation, indicating the direction of excitation at each point. Then animation is good and as expected similar to FE results. However importing the uff files back into ME Scope without assigning the animation equation does not match to the FE Results. The probable reason is that coordinate system used for experimental analysis is cylindrical coordinate system. In FE Analysis the default coordinate system is cartesian. when Experimental results are imported into Nx CAE, i am not able to assign the coordinate system at each measured point, hence the mode shape is not matching. Kindly suggest what to do

Re: Correlation of cylindrical shell



You have defined a cylindrical coordinate system at each sensor location. This does not make sense: There should be a single cylindrical system on the longitudinal axis, which is assigned to the sensors and from which the software derives a series of cartesian systems, where each cartesian system is defined from the master cylindrical system and the sensor locations.


I have manually modified the csys types from cylindrical to cartesian.


Still the results don't make sense. Will follow up




Philippe Tremblay

Re: Correlation of cylindrical shell

I saw this issue a couple of months ago, the official I-DEAS Universal File Format always stored nodes in the Global Coordinate System, the nodal records include a reference to a displacement coordinate system, that could be cylindrical, but the nodal values themselves never are.  I think some 3rd parties have interpreted the coordinate system number incorrectly.


Name:   Nodes - Double Precision

Status: Current

Owner:  Simulation

Revision Date: 23-OCT-1992

----------------------------------------------------------------------------Record 1:        FORMAT(4I10)

                 Field 1       -- node label

                 Field 2       -- export coordinate system number

                 Field 3       -- displacement coordinate system number

                 Field 4       -- color

Record 2:        FORMAT(1P3D25.16)

                 Fields 1-3    -- node coordinates in the part coordinate


Records 1 and 2 are repeated for each node in the model.




       121         1         1        11

   5.0000000000000000D+00   1.0000000000000000D+00   0.0000000000000000D+00

       122         1         1        11

   6.0000000000000000D+00   1.0000000000000000D+00   0.0000000000000000D+00


Re: Correlation of cylindrical shell

As masherman has indicated, the problem is that the shapes are expressed in local coordiante systems. Mode shape dataset 55 expects the displacements or accelerations to be in the global coordinate system, and NX FE Correlation does not attempt to make any transformation. MeScope has exported the shapes in local systems: Is there an option that would allow export in the global system?

Philippe Tremblay