SDC Torsion Design

office (847) 816-3816
fax (847) 816-3971
toll-free (888) 425-2475

Crane Girder Design

SDC has spent 20 years studying the design of crane girders for applications in the steel industry.  The section on ../SDC Torsion Design/SDC Torsion Design.htm goes into some of the details of our design approach including a sample output from our proprietary computer program.  The computer program has allowed SDC to achieve a greater understanding of crane girder failure than any of our competitors.  SDC has been able to directly correlate our crane girder inspection findings with the analytical results from the computer program.  The program successfully predicted the failure in a component of a new, heavier replacement crane girder with a thrust plate and new tie-back connections.

SDC has confirmed that torsional thrust is not a secondary effect and when combined with fatigue stress reversal, causes most of the crane girder failures in steel mills.  Attached is a photograph of a 112 foot long crane girder that is 11'-6" deep.  While the girder was designed to support two (2) 50T EOT cranes, a single, stationary crane moving plates to a nearby piler crane caused the warp shown in the photograph.  The crane girder was replaced by a torsion-resistant girder because the crane rail could not bear flush on the girder flange.

2005 AISC SPECIFICATION

The 2005 AISC Specification integrated both the ASD (Allowable Stress Design) and LRFD (Load Resistance Factor Design) Specifications into a single set of rules for the design of steel structures. The new specification unified the design standards with the most current knowledge and design practices.

For crane runway design, the section regarding plate girders found in the 9th edition of the code has been eliminated. Section F12 has been added to include the design of unsymmetrical structural shapes such as built-up plate girders. For standard structural sections, meeting the new requirements is not that difficult. However, for crane girders constructed from built-up sections, meeting the requirements of Section F12 is extremely challenging. The design engineer needs to determine the shear flow in the girder that conforms to zero warping shear stress at the terminal ends of the girder cross section and calculate the torsional warping constant (Cw) in order to apply the code equations.

Below are some crane girder design tips that may be of interest:

  • Welded crane girders should be designed with heavier web plates so that the intermediate stiffeners can be eliminated.  The fillet welds from the intermediate stiffeners create fatigue issues that result in base metal cracks in the girder webs.  The base metal cracks are not due to direct flexural bending or shear.  They are the result of the stress reversal due to the combined effect of flexural shear, pure torsional shear, and warping shear.

  • Crane girders should be designed with a crane rail offset from the centerline of the girder of at least 0.25% of the girder length.  There are several reasons for this crane girder design requirement.  First of all, older mill building are not square.  Maintaining straight and parallel crane rails in accordance with CMAA requirements can be very difficult.  It is SDC's experience that realigning crane girders is not practical and very expensive.  Moving the crane rails off the girder centerline is the most efficient way to realign crane rails.  The second reason to design the crane girders with a crane rail offset is because cranes rarely make square lifts.  The momentum created by a swinging load applied to the top of the crane rail as a trolley moves laterally and crane bridges down a runway induces large forces which are not clearly understood.  The crane rail offset helps account for the P delta affects due to both lateral deflection and girder rotation from the twisting or warping of the girder. 

  • Gantrex Tie-Back systems work well but can be expensive to install correctly.  The connections to the girders or columns usually fail due to under design.  Design engineers need to make sure that the maximum total design thrust on the tie-backs corresponds with the same load/force combination that induces the maximum torsional stress in the girder. 

  • Both girder stress and thrust on the tie-backs must be performed for each wheel of the crane as it bridges down the crane girder.  To apply the proper structural properties (gross section modulus, effective section modulus, moment of inertia, normal distance to the shear center, etc.) and to manage arrays of response data (shear, moment, deflection, rotation and derivatives of rotation) due to various load combinations form multiple crane wheels is a tremendous challenge from an accounting perspective.  Database technology is perhaps the only way to keep track of the loads and (+/-) signs when combining the torsional and bending stresses. This task is even more challenging when considering fatigue and non-fatigue conditions when the effective cross section is calculated.  It is our experience that a spread sheet is very limited and a poor tool to track the intermediate results of the analysis.

  • Torsion-resistant crane girders need to be sized with the proper balance between its length (L) and its torsional inertia property ratio (Beta).  Both the Saint-Venant torsional constant (J) and the warping constant (Cw) need to be determined to calculate the (Beta) value.  The optimum cross section for a torsion-resistant crane girder needs to have its (Beta*L) product fall in the torsional response range of "mixed torsion" between dominating warping torsion and dominating Saint-Venant torsion.  If the (Beta*L) value is not optimized, the design of a shorter girder tends to be controlled by warping while a longer girder tends to be controlled by pure torsion. 

  • Calculation of the warping constant (Cw) for an unsymmetrical crane girder cross section can be very tricky and complicated.  Properly performing the linear integration for the shear flow about the shear center was the most difficult aspect in development of our crane girder design computer program. 

  • The cross sectional properties of a crane girder are calculated differently for a flexural and a torsional analysis.  Within the flexural analysis, properties for bending stress and for deflection also require different effective girder cross sections.  Essentially the calculation of the properties for a given girder profile for regular flexure alone needs to be done four times (once for the gross section, once for effective section, once for orthogonal axes, and once for the principle axes).  For a non-symmetrical girder profile, one trial section for a preliminary crane girder design could take a man month to calculate by hand if we are lucky and make no calculation mistakes!  This does not include time spent on review.  A girder profile configuration fitting the optimized criteria can rarely be achieved in one trial and the cost to design a torsion-resistant crane girder without automation is prohibitive.

 112 Foot Long Crane Girder with 5" Warp in Bottom Flange.

112 Foot Long Crane Girder with 5" Warp in Top Flange and Backup Truss Causing Wear on New Crane Rail.  Note:  The Top Flange Warp is Not Visible in this Photograph.

Crane Runway with Torsionally Reinforced Crane Girders.  The rail misalignment due to girder warp caused the crane to jump off the rail.