Cold-formed steel design to American standards

What makes CFS different from hot-rolled steel

Behnam Ordoubadi opened by walking through why cold-formed steel requires a different design mindset than hot-rolled steel. He framed the comparison along three axes: material behavior, section geometry, and failure modes.

On the advantages side, CFS is well suited to light loads and short spans where hot-rolled sections would be over-engineered. It has a high strength-to-weight ratio, ships compactly, and can be substituted one-for-one for individual timber framing members in some applications. Fabrication and erection are precise and not weather-dependent the way concrete forming is.

On the challenge side, the core issue is that CFS sections are slender. While compact hot-rolled sections can typically reach their full plastic moment before local buckling, CFS sections buckle before yielding in most cases. This changes how capacity is calculated: you cannot use gross section properties directly. Behnam also highlighted that web crippling - the local buckling of the web at supports or point-load locations - is much more likely to control CFS beam design than it is for hot-rolled steel, and requires careful checking.

He also noted a practical strength of CFS: section shapes are highly customizable. SFIA and SSMA publish more than 700 standard section options, but engineers can also define an arbitrary section shape and design it to the clauses in AISI S100-16, which can be economical for specialized applications.

The three buckling modes and how finite strip analysis finds them

Behnam described the three buckling modes that AISI S100-16 requires engineers to check, and explained why finite strip analysis is the method the code is oriented around.

Local buckling occurs when the flat plate portions of the section (web, flanges) buckle between the corner elements, which remain relatively fixed. Typical half-wavelengths are 4 to 10 inches. Distortional buckling occurs when the lip-web junction rotates, causing the stiffener lip to move relative to the web. It typically governs at half-wavelengths of 16 to 30 inches - intermediate lengths where global lateral-torsional buckling has not yet taken over but the member is too long for pure local buckling to be the critical mode. Global lateral-torsional buckling involves the displacement of the entire cross-section and is capped by the yield moment.

Finite strip analysis is described in AISI S100-16 as the preferred approach for the Direct Strength Method. It runs through every possible half-wavelength for a given cross-section and produces a curve showing load factor against half-wavelength. Two local minima on this curve identify the local buckling capacity and the distortional buckling capacity. Behnam noted that the analysis only needs to be done once per section, and that while alternative methods exist, the code effectively directs engineers toward some form of rational elastic buckling analysis.

The final nominal flexural strength is taken as the minimum of the moment capacities for global, local, and distortional buckling modes, each multiplied by the resistance factor phi equal to 0.90.

Shear, web crippling, and load interaction checks

Behnam walked through the remaining capacity checks in sequence, with particular emphasis on web crippling.

For shear, the key distinction is whether the web has transverse stiffeners or not. Without stiffeners, the yield shear force and a slenderness-based reduction factor determine the nominal shear strength, with phi equal to 0.95. With stiffeners, the shear panel geometry (width and length between stiffeners) governs, and the equations are structurally identical to the local buckling check for flexure because the governing mechanism is local buckling within the shear panel. If stiffener spacing exceeds twice the beam depth, the web is treated as unstiffened.

Web crippling, Behnam said, is really likely to control CFS design. He spent time on the table-lookup procedure required by AISI S100-16, which involves four classification decisions: fastened versus unfastened (bolted versus screwed connection to the support), stiffened versus unstiffened (whether a lip is present on the section), one-flange versus two-flange loading (two-flange applies when the distance between concentrated loads is less than 1.5 times the beam depth), and end versus interior condition (end applies when the distance from a point load to a support is less than 1.5 times the beam depth). He flagged that not all combinations are covered in the standard tables, and engineers encountering an uncovered case must use the most conservative applicable entry.

For load interaction, the required checks depend on whether shear stiffeners are present. With stiffeners, the check is a square-root sum of squares combining the flexural and shear utilization ratios. Without stiffeners, separate reduced capacity values are computed for flexure and shear without global buckling effects, and then two interaction equations are checked.

Design walkthrough using Calcs.com

Behnam ran through two live examples in the Calcs.com CFS beam calculator to show how these checks map to practice.

The first was a simply supported 68-inch beam for an office floor, with a 16-inch tributary width, no transverse shear reinforcement, and lateral unbraced length of 24 inches with the bottom flange torsionally unbraced for the full span. He selected a 362S section from the built-in section library, entered pinned end conditions, applied dead and live loads, and reviewed output showing all code checks with demand-capacity ratios. He highlighted that the calculator shows which of the more than 700 available sections pass and fail for a given loading, making section optimization straightforward.

The second example involved a three-span beam with spans of 72, 120, and 24 inches, more complex bracing conditions, and varying support configurations. A single C-section in that case failed the deflection check. Behnam demonstrated the back-to-back section option in the calculator, which resolved the deflection exceedance. He also pointed to the nested Z-section option as a good choice when web crippling is the controlling limit state, since the doubled web depth directly addresses the crippling check.

He mentioned that custom cross-section properties tools are available in Calcs.com for engineers working with non-standard built-up sections, noting that full design integration for custom sections was still in development at the time of the webinar but that analysis was available.