Walk onto almost any modern steel-framed construction site and you'll see hollow sections doing work that open profiles like I-beams and channels simply can't do as efficiently: standing as slender exposed columns in a glass atrium, forming the diagonal bracing in a seismic frame, or carrying the torsional loads in a curved footbridge. The preference isn't aesthetic. It comes down to a handful of structural engineering principles that most procurement teams never see discussed, but that quietly determine why a specification calls for SHS, RHS, or CHS in the first place.
This article looks at the actual engineering reasons hollow sections outperform open profiles in specific load cases - and, just as importantly, where an I-beam or channel is still the more efficient and economical choice. Understanding both sides of that comparison is what separates a genuinely informed specification from a default habit.
The Geometry Advantage: Equal Resistance in Every Direction
The single biggest structural advantage of a hollow section comes down to one number: the radius of gyration.
An open section like a wide-flange I-beam has two very different radii of gyration - one about its strong axis, one about its weak axis, often differing by a factor of two or more. That means an I-section column is only as strong as its weakest direction of buckling, and if a member is exposed to load or wind pressure from multiple directions, the engineer has to design around the weaker axis regardless of how much stronger the other axis might be.
A square or circular hollow section, by contrast, has an equal (or very close to equal) radius of gyration in every direction. There is no "weak axis." For a column that needs to resist buckling under compressive load applied from unpredictable or multi-directional sources - which describes a large share of real building columns, especially in open floor plans without bracing walls to define a dominant load direction - this geometric property alone often makes a hollow section lighter than an equivalent-capacity open section, because none of the material is "wasted" providing strength in a direction the structure doesn't need as much.
This is also why hollow sections dominate as the compression members in space frames and lattice trusses: every diagonal and vertical member in a 3D truss system experiences load from a different orientation, and a section with uniform resistance in all directions is structurally efficient in a way an I-section fundamentally cannot match without added material.
Show Image
Torsional Stiffness: An Order-of-Magnitude Difference
If radius of gyration is the subtle advantage, torsional stiffness is the dramatic one - and it's arguably the most underappreciated reason hollow sections get specified for certain members.
Open sections resist torsion primarily through warping and a relatively weak St. Venant torsional constant, because an open thin-walled shape simply doesn't have a continuous load path around its perimeter to carry shear flow efficiently. A closed section - any hollow section - resists torsion through shear flow running continuously around the closed perimeter, which is a dramatically stiffer mechanism. In practical terms, a closed hollow section can have a torsional constant an order of magnitude higher than an open section of comparable size and weight.
This matters directly in members that see meaningful torsional load: curved bridge girders, crane runway beams subject to eccentric wheel loads, spandrel beams supporting eccentric floor loads along a building perimeter, and any member in a structure where load isn't applied cleanly through the shear center. Specifying an open section for a genuinely torsion-critical application usually means adding stiffeners, diaphragms, or oversizing the member to compensate - extra fabrication and material cost that a hollow section often avoids by geometry alone.
DRILLING HOLE PROCESS
HOLLOW SECTION PROCESSING
Round steel pipe processing
Local Buckling Classification: Where Hollow Sections Have Limits
None of this means hollow sections are unconditionally superior, and this is where a more nuanced, less commonly discussed point comes in: local buckling behavior.
Design codes - EN 1993-1-1 in Europe, AISC 360 in the US - classify structural sections (open and closed alike) by the width-to-thickness or diameter-to-thickness ratio of their individual plate or wall elements, into classes ranging roughly from "compact" (able to reach full plastic moment capacity and sustain rotation for plastic design) through "slender" (where local buckling occurs before the section reaches yield, and the design capacity has to be reduced accordingly).
For hollow sections, this classification is directly tied to the width-to-thickness ratio of the flat faces (for SHS/RHS) or the diameter-to-thickness ratio (for CHS). A thin-walled, large-diameter CHS or a wide, thin-walled RHS can fall into a slender classification just as readily as a thin-flanged open section can - the closed profile does not automatically exempt a section from local buckling checks. Engineers specifying larger, thinner-walled hollow sections for architectural or long-span applications need to verify the section's classification explicitly rather than assuming that "hollow" implies "compact." This is one of the more common oversights in preliminary sizing, caught only when a detailed check is finally run.
Fire Performance: An Advantage Rarely Mentioned in Sales Material
One genuinely underdiscussed advantage of hollow sections shows up in fire engineering, not gravity or lateral load design.
Structural fire protection requirements are driven substantially by a section's "section factor" - the ratio of heated perimeter to cross-sectional area (often written Hp/A). A section with a lower Hp/A ratio heats up more slowly during a fire, because there's proportionally less exposed surface area driving heat into a given mass of steel. Open sections like I-beams expose surface area on both sides of the web and both flanges, all separately heated. A hollow section of comparable load capacity often has a lower Hp/A ratio simply because its geometry concentrates the same steel mass behind a more compact, continuous exposed surface.
The practical consequence is that hollow section members can sometimes achieve a required fire resistance rating with a thinner intumescent coating or less fire protection material than an open section carrying the same design load - a real cost and weight saving that rarely appears in general marketing material because it requires a section-specific calculation to demonstrate, not a general claim that applies uniformly across every size and application.
The Connection Design Trade-Off
This is the point where hollow sections lose some of the advantage they gain elsewhere, and it deserves more attention in specification decisions than it usually gets.
Open sections offer flat, accessible flange and web surfaces that make bolted connections straightforward - a connection plate can be bolted directly to a flange with standard clearances and standard detailing. Hollow sections, by their closed nature, don't offer that same accessible surface for a conventional bolted connection without additional hardware: through-bolting requires access to the tube interior (often impossible once the structure is assembled), and surface-mounted bolted connections typically need flange plates, cap plates, or blind-bolt systems specifically designed for closed sections.
Welded connections carry their own complexity. Where hollow section members meet at a truss node - a K, T, Y, or X joint in lattice girder construction - the connection isn't just a matter of welding two tubes together. The chord face at the joint has to be checked for punching shear and local plastification under the concentrated load from the incoming brace member, following design provisions such as those in EN 1993-1-8 Section 7 or the CIDECT design guides developed specifically for tubular joint design. These checks are genuinely more involved than a standard open-section connection design, and underestimating this complexity at the design stage is one of the more common sources of late-stage redesign on tubular truss projects.
None of this makes hollow section connections impractical - tubular trusses are built successfully worldwide - but it does mean the fabrication and connection design cost has to be weighed against the material efficiency gained from the section itself, rather than assuming hollow section is a straightforward cost win in every case.
Where Open Sections Still Make More Sense
Given everything above, it's worth being direct about where an I-beam or channel remains the better choice: primary bending members where load is applied predominantly about one axis, and where that axis is known and fixed by the structure's geometry - a typical floor beam supporting a slab, for instance, bends about one axis only, and an I-section's material, concentrated in flanges far from the neutral axis, is more efficient in pure uniaxial bending than a hollow section's material distributed around a closed perimeter.
Open sections are also generally cheaper to connect, as discussed above, and are more widely available in a broader range of standard rolled sizes for very large, deep members than hollow sections typically are. For a straightforward floor beam, an I-section usually remains the more economical specification; for a column exposed to multi-directional load, a bracing member under compression and possible torsion, or a truss chord in a space frame, a hollow section typically wins on both material efficiency and, in many cases, fire performance.
Making the Right Call
The choice between hollow and open sections isn't a matter of one being categorically superior - it's a matter of matching section geometry to the actual load case the member experiences. A structural engineer weighing this decision is really asking: does this member see load from a single predictable direction, or from multiple directions and possibly torsion? Is local buckling classification going to govern the design at the wall thicknesses under consideration? What does the connection design actually cost once flange plates, cap plates, or tubular joint checks are factored in, rather than just the per-ton price of the steel itself?
Getting this right at the specification stage - rather than defaulting to habit or to whatever section type a past project happened to use - is where real material and cost savings come from on a structural steel project, and it's a decision worth revisiting on every new structure rather than assuming the previous project's section choice still applies.
