scamplogic
Structural
- Jul 14, 2013
- 5
Hi guys, long time lurker and a first time poster.
I'm writing some analysis software for Australian codes for steel monopoles. The software splits each pole segment into 10 smaller pieces with their own specific diameters and properties. Design shears and moments are calculated using the principle of superposition by determining the mass/load over that segment and then treating it as a discrete point load with a lever arm extending to the fixed end. So far, this seems to have been working just fine.
...Until I tried to add functionality for calculating the natural frequency. The rotations and deflections at each point down the pole are determined using the moment-area theorem after obtaining the "natural frequency moment" for a cantilever (again, using superposition). When I analyse a uniform cross-section I get identical results to Microstran, but when I have two elements in the cantilever (e.g. 10.0m long) with different stiffnesses (I've been testing a 5.0m 508CHS at the base attached to a 5.0m 114CHS) my results are way out. While I believe I've narrowed it down to the way I'm calculating my natural frequency moment (where each "piece" of the pole is treated as a discrete mass with a specific lever arm based on it's height above the base), what I can't explain is that a hand calculation using superposition works just fine if you treat the mass of the two cross-sections as two independent UDLs. On the other hand, using a series of point loads (which is preferable, because the program can analyse the pole the same way regardless of taper) with changing stiffness does not. Is this a consequence of using the method of superposition? Is superposition even applicable in my case? I've come across some literature that suggests it might not be...
Thanks!
I'm writing some analysis software for Australian codes for steel monopoles. The software splits each pole segment into 10 smaller pieces with their own specific diameters and properties. Design shears and moments are calculated using the principle of superposition by determining the mass/load over that segment and then treating it as a discrete point load with a lever arm extending to the fixed end. So far, this seems to have been working just fine.
...Until I tried to add functionality for calculating the natural frequency. The rotations and deflections at each point down the pole are determined using the moment-area theorem after obtaining the "natural frequency moment" for a cantilever (again, using superposition). When I analyse a uniform cross-section I get identical results to Microstran, but when I have two elements in the cantilever (e.g. 10.0m long) with different stiffnesses (I've been testing a 5.0m 508CHS at the base attached to a 5.0m 114CHS) my results are way out. While I believe I've narrowed it down to the way I'm calculating my natural frequency moment (where each "piece" of the pole is treated as a discrete mass with a specific lever arm based on it's height above the base), what I can't explain is that a hand calculation using superposition works just fine if you treat the mass of the two cross-sections as two independent UDLs. On the other hand, using a series of point loads (which is preferable, because the program can analyse the pole the same way regardless of taper) with changing stiffness does not. Is this a consequence of using the method of superposition? Is superposition even applicable in my case? I've come across some literature that suggests it might not be...
Thanks!