Space Frame Structures: Three-Dimensional Truss Systems
A conventional planar truss carries load in one vertical plane; a series of them, linked by purlins, does the work of an ordinary roof. A space frame instead arranges members in a three-dimensional lattice, usually two parallel layers of grid connected by diagonal web members, so that load applied anywhere on the top layer finds multiple paths through the grid to the supports. That redundancy is the whole point: a space frame spans large, column-free areas − exhibition halls, gymnasiums, aircraft hangars, stadium roofs − using members that are individually light because each one carries almost pure axial force rather than bending.
Geometry and Member Behavior
The most common configuration is the double-layer grid: a top chord layer and bottom chord layer, each typically arranged on a square or triangular grid, offset from one another and connected by diagonal web members that form pyramidal units between the two layers. Because every joint is treated as effectively pinned and every member is expected to carry axial force only, the same idealization used in planar truss analysis extends into three dimensions, just with equilibrium now written in three force directions at every joint instead of two.
This axial-only behavior is what makes space frames so material-efficient for their span: a member loaded in pure tension or compression uses its full cross-section, unlike a bending member where material near the neutral axis contributes little to capacity. The trade-off is joint complexity. A double-layer grid node can have eight or more members converging from different directions at different angles, and the connector at that node has to transfer all of those axial forces without introducing the bending or eccentricity that would violate the pin-jointed assumption the whole analysis relies on.
Proprietary connector systems (ball-and-socket hubs threaded to receive tubular members, or bolted node plates) exist specifically to standardize this joint problem, letting a manufacturer mass-produce a single node type used at every intersection regardless of how many members frame into it, rather than engineering a bespoke connection for each unique joint on the project.
Load Distribution and Redundancy
A concentrated load on a double-layer grid does not stay local the way it might on an isolated planar truss; it distributes across many members in the surrounding grid, so peak member forces under a point load are lower than a simple tributary-area calculation would suggest. This behavior also gives space frames a meaningful degree of structural redundancy: the loss or failure of a single member, from fabrication defect, impact damage, or fire, does not necessarily bring down the whole system, since load can redistribute to adjacent members through the grid's multiple paths, unlike a determinate planar truss where losing one critical member can trigger a local collapse.
That said, the redundancy is not unlimited, and edge and corner regions of a grid, along with the members immediately around a support, typically carry higher forces than the interior field and deserve closer attention during design; support reactions in a double-layer grid tend to concentrate through a small number of members near each column, which is often the controlling design case rather than the more uniformly loaded interior.
Supports and Erection
Space frames are usually supported on a relatively small number of columns compared to the area they cover, which is the entire architectural appeal, but that concentration means support reactions and the connection details at each column are disproportionately important relative to the lightness of the typical interior member. Some designs taper the grid depth or add extra diagonal members near supports specifically to manage this concentration without oversizing every member in the field to match.
Erection method also shapes the design. Many space frames are assembled at ground level in modular units or even as a complete grid, then jacked or lifted into final position as a single rigid piece, which avoids the need for extensive temporary shoring at height but requires the grid to be checked for a temporary lifting condition − different support locations and often different member forces than the final in-service condition, and sometimes the case that actually governs member sizing for lighter, longer-span grids.
Deflection, not member strength, frequently governs proportioning for long-span space frames, since a grid stiff enough to meet serviceability deflection limits under snow or live load is usually already well within stress capacity; depth-to-span ratios for double-layer grids commonly fall in the range of 1/15 to 1/20, somewhat deeper relative to span than a comparable two-way concrete slab system but far shallower than a solid planar truss of the same span would require.