12-17-2018 08:36 AM
Dear All,
I am preparing structural analysis for a mechanism. I have to make modal analysis and frequency response (sine) check. Mechanism is very simple (all structure is aluminum) but part of it is rotated so I have definied additional coordinate system to read results in rotated coord system. All nodes are definied in global coordinate system 0, also output coordinate sytem for all nodes is set to 0. For one node (12288) in the middle of top bracket I have definied output coordinate system to 3.
The problems I have observed are as follows:
1. Total translation plots for results from modal analysis (1st mode) and sine (around 1st mode) are different. Basicaly what I have noticed is that for the part of the results are ploted in wrong coordinate system.
As you can see in modal analysis plot upper part is bending in Z axis of cordinate system no. 3. For sine response around resonance frequency the same bracket is bending in direction of Y axis of cordinate system no. 3.
We have made some tests on vibration table and results from test show that real mechanism shakes as presented in modal plot. Comparing results from two 3-axial accelerometers (one placed on hinge 2 and second at node 12288 location) we have observed very good corelation between test results on hinge 2 but at node 12288 it seems axes are incorectly readed from FEM.
2. When I read accelerations for node 12288 using "results to data table" they are also wrongly orienteted.
But when I read data directly from .f06 file the responses are in line with tests results (at least in terms of axes orientation).
My question is - whether in settings of the analysis for frequency response is there any kind of the option I have used incorrectly?
Version of Feamap I am usisng is 11.2.1.
I will greatly appreciate any feedback.
Best regards,
Michał Sidz
12-18-2018 02:08 PM
Dear Michael,
When you run a normal modes/eigenvalue analysis with NX NASTRAN (SOL103) the displacement results you obtain are useless, displacements are normalized, then is NOT correct to use this value to compare with any other analysis type, displacements here in SOL103 is an easy way to show the deformed mode shape, but nothing else. Please note when you run a normal mode/eigenvalue analysis not any load set is aplied, neither force or pressure is applied, this is the reason why displacement results are useless. The usual reason when performing a SOL103 analysis is determining the natural frequencies and mode shapes of the structure with damping neglected. Natural frequencies and mode shapes are functions of the structural properties and boundary conditions.
Best regards,
Blas.
1. Total translation plots for results from modal analysis (1st mode) and sine (around 1st mode) are different. Basicaly what I have noticed is that for the part of the results are ploted in wrong coordinate system.
12-18-2018 02:14 PM
Dear Michael,
When running a SINE VIBRATION MODAL FREQUENCY RESPONSE dynamic analysis is critical to activate in the NX NASTRAN OUTPUT REQUEST the option "relative enforced motion results", read my blog in the following address:
Best regards,
Blas.
12-19-2018 05:30 AM
Dear Blas,
thank you for your answer but it is not exacly solution to my problem. As I was trying to explain when I read data using "results to data table" FEMAP is somehow "mixing" axes. I noticed that when I was comparing FEMAP results and test results we obtained on our mechanism.
For the node I am intrested I set proper output cordinate sytem and using results to data table read accelerations. Below you can find comparison between results to data table and .f06.
I understand that from modal analysis values of displacement are useless. Correct me if I am wrong but if I understood correctly shape of the mode shall not change between modal and frequency response analysis around the same frequency. On the animations I attached in the first post you can observe that behavior.
What I would like to understand is whether I am doing something incorrectly with reading data via FEMAP or I am struggling with kind of the bug.
Best regards,
Michał Sidz