m_ridzon
Mechanical
- Sep 18, 2020
- 103
Folks, I'm running a Sec. III, Div. 1 analysis in ANSYS. The customer's seismic input is in the form of an acceleration spectrum. I can't speak for other FEA, but the ANSYS response spectrum methodology takes modal data and churns it through a series of arithmetic operations, to include SRSS which strips away plus/minus signs for results.
BPVC requires that stress results of load combinations be done at the component level first (e.g., Sx, Sy, Sz, Sxy, Syz, Szx) before calculating stress intensity (SINT). In other words, in the two contrasting examples below, Option A is code required and B is not...
Option A
SXLevelA + SXSeismic
SYLevelA + SYSeismic
SZLevelA + SZSeismic
SXYLevelA + SXYSeismic
SYZLevelA + SYZSeismic
SZXLevelA + SZXSeismic
...now calculate SINT with the 6 components
...versus
Option B
SINTLevelA + SINTSeismic
With all of this said, how does one handle the fact that ANSYS response spectrum strips away the sign during its arithmetic? And in that case, would Option A produce anything meaningful? Is there another way to tackle this?
BPVC requires that stress results of load combinations be done at the component level first (e.g., Sx, Sy, Sz, Sxy, Syz, Szx) before calculating stress intensity (SINT). In other words, in the two contrasting examples below, Option A is code required and B is not...
Option A
SXLevelA + SXSeismic
SYLevelA + SYSeismic
SZLevelA + SZSeismic
SXYLevelA + SXYSeismic
SYZLevelA + SYZSeismic
SZXLevelA + SZXSeismic
...now calculate SINT with the 6 components
...versus
Option B
SINTLevelA + SINTSeismic
With all of this said, how does one handle the fact that ANSYS response spectrum strips away the sign during its arithmetic? And in that case, would Option A produce anything meaningful? Is there another way to tackle this?
