Beam Torsional Stiffness by STAAD

Chris,

Can you help me this out?

There is a simple beam resisting a torque as shown in the file attached. The beam is fixed at both ends. Calculated by STAAD, the torsional angle at the loading point is 0.121 rad. But when I do the hand calculation (using formulas in AISC Steel Design Guideline Series 9), the angle is 0.067 rad (about 50% for this calculation case).

I checked the sectional properties of the beam given in STAAD, everyone looks good except the warping constant Cw is not given.  Coicidently, when I ignore Cw in the calculation the rotational angle is close to the value from STAAD. Does STAAD consider Cw?

Regards,

j1d

 

Parents
  • On page 61 of the Guide, the plot of the torsional function Theta''*(GJa/T) is zero only at the 2 ends of the beam. At all intermediate locations, it is non-zero, indicating that warping resistance is mobilized at interior points along the span. It is allowed to warp freely only at the ends.

    In the STAAD analysis, warping stiffness is not considered at any location along the longitudinal axis of the member. By default, the torsional resistance anywhere along the length of the beam is solely the St.Venant's torsional resistance.

    From a broader perspective, the reason why STAAD and probably most other structural analysis programs do not consider the stiffness associated with warping is perhaps the following.

    The equations contained in the AISC Design Guide for computing the warping resistance are for idealized end conditions - Free, Fixed or Pinned. In a general framed structure where the type of connections between members determine the amount of fixity at the ends, the guide recommends a more advanced analysis (see the left half of page 108), which usually means modeling every component of the cross section - flanges and webs - using finite elements.

    The other aspect is the torsional loading. The exact position of the loading decides the amount of twist the member undergoes. But in a generalized program like STAAD based on the stiffness method, the torsional load is converted to equivalent "fixed end actions" at the nodes of the member. So, the rotation is calculated on the basis of uniform torsion on the full member span, not a torsion that varies from one end of the span to the other.



Reply
  • On page 61 of the Guide, the plot of the torsional function Theta''*(GJa/T) is zero only at the 2 ends of the beam. At all intermediate locations, it is non-zero, indicating that warping resistance is mobilized at interior points along the span. It is allowed to warp freely only at the ends.

    In the STAAD analysis, warping stiffness is not considered at any location along the longitudinal axis of the member. By default, the torsional resistance anywhere along the length of the beam is solely the St.Venant's torsional resistance.

    From a broader perspective, the reason why STAAD and probably most other structural analysis programs do not consider the stiffness associated with warping is perhaps the following.

    The equations contained in the AISC Design Guide for computing the warping resistance are for idealized end conditions - Free, Fixed or Pinned. In a general framed structure where the type of connections between members determine the amount of fixity at the ends, the guide recommends a more advanced analysis (see the left half of page 108), which usually means modeling every component of the cross section - flanges and webs - using finite elements.

    The other aspect is the torsional loading. The exact position of the loading decides the amount of twist the member undergoes. But in a generalized program like STAAD based on the stiffness method, the torsional load is converted to equivalent "fixed end actions" at the nodes of the member. So, the rotation is calculated on the basis of uniform torsion on the full member span, not a torsion that varies from one end of the span to the other.



Children
  • Now I got it. Thanks!

    No warping constant included in the member stiffness sounds like a loss for me. Just as the calculation example shows, the torsional angle increased about 100%. Warping deformation  is beyound the 6 degrees-of-freedom of a node in a member, but I don't know if the warping stiffness Cw can be somehow incorprated into the term of pure torsion (St.Venant's Torsion).

    Regards,

    j1d