Learning Forum

[simon.stadler]

I need help with general interpretation of Response Spectrum Analysis (RSA) results in ANSYS.

The problem is explained using a simple single beam model with a point mass at one end. The other end has a fixed support.

Model description:

• beam of length L=1000mm, with rotationally symmetric circular cross section with radius R=10mm.
• the beam is assumed to be massless, with elastic modulus of 200000MPa
• orientation of the beam axis is in global z axis
• one end of the beam has a fixed support
• the other end has a point mass m=1kg

The Modal analysis gives two orthogonal bending modes with identical natural frequency.

The used response spectrum is constant for 1Hz-1000Hz with acceleration of 1000mm/s².

The response spectrum is applied in both horizontal directions X&Y.

For Modal Combination CQC is used with 2% damping.

The orthogonal combination technique is SRSS (default in ANSYS).

The RSA results in a normal stress of 1.2759MPa at the fixed end of the beam, constant around the circumference of the circular cross section.

The bending moments at the fixed end of the beam are Mx=1000Nmm, My=1000Nmm.

In total Mtotal=SQRT(1000²+1000²)=1414.2Nmm.

Taking this moment Mtotal for stress calculation via Mtotal/W gives a normal stress of 1.8006MPa.

So there is exactly a factor of sqrt(2) between the directly received stress results and the stress result calculated via the bending moment.

1.8006/1.2759=sqrt(2)

This raises the question: What is the correct approach for verification of the cross-section? Should I use the stresses calculated directly in the RSA, or should I use the bending moment? Where is the mistake / misinterpretation?

Thanks for help.

[dlooman]

The beam element is only evaluating the stress at the extreme fiber locations relative to the principal axes.  It's not evaluating the stress at a location 45 degrees away.  I think you would see this same issue if you applied two forces to the beam in a static analysis.  The stress being reported is based on a moment of just 1000 N-mm, 1414/sqrt(2).