Pattern live loading is the discipline of admitting that live load chooses its own arrangement: dead load is everywhere by definition, but the crowd, the inventory, and the furniture can occupy any subset of spans, and for continuous structures some subsets hurt more than everything-everywhere does. StructLoads keeps the distinction clean by treating the full-load case as the takedown’s backbone, it governs the accumulated column and foundation totals, while flagging where the arranged cases govern instead: mid-span moments, hogging peaks, cantilever see-saws, and the uplift reversals that full loading actually hides. The rule of thumb that organizes the whole topic: adding load to a continuous structure does not always add demand at a given point, and pattern loading is simply the search for the arrangement each point fears most.
Why full loading is not the worst case
On a simply supported beam, more load means more of everything, so the fully loaded case governs and the topic ends. Continuity breaks that monotonicity. Load an interior span of a continuous beam and it deflects downward; its neighbours, pried over the shared supports, bend upward in sympathy, their mid-spans actually relieved. The influence line makes this precise: for any response, a mid-span moment, a reaction, a shear, the influence line maps how load at each position feeds that response, positive here, negative there, and live load placed on the negative regions reduces the response. Full loading therefore understates every response whose influence line changes sign, and the worst case for each response is live load on its positive regions only.
The classic arrangements fall straight out of the sign map. Maximum sagging moment in a span: live load on that span and then on alternate spans outward, skipping its neighbours. Maximum hogging moment at a support: live load on the two spans adjacent to that support, skipping the next ones. Maximum column moment in a frame: unbalanced arrangements, loaded bay one side, empty the other, that twist the joint. Two-way systems generalize this to the checkerboard, alternate panels loaded, which drives the largest panel moments in flat slabs. And the extreme case is the cantilever see-saw: overhang loaded with back-span empty maximizes tip effects and can reverse the far reaction into uplift, a response that the fully loaded structure never shows at all, which is why pattern loading is a safety topic, not an optimization.
| Response sought | Live load arrangement |
|---|---|
| Max sagging in span k | span k, then alternate spans outward |
| Max hogging at support j | both spans adjacent to j, skip next |
| Max column joint moment | one adjacent bay loaded, other empty |
| Max flat-slab panel moment | checkerboard of alternate panels |
| Cantilever uplift at far support | overhang loaded, back-span empty |
What the codes actually require
Codes turn the theory into a duty with thresholds. US practice requires continuous members to be designed for the pattern producing the worst effect, with simplified permissions when the live-to-dead ratio is modest: the well-known concrete simplifications allow just two arrangements, alternate spans for sagging, adjacent spans for hogging, when spans are roughly equal, and the coefficient methods bake typical patterns into their tabulated factors. European practice similarly defines its load arrangements for continuous slabs and beams and lets national annexes trim the case count for ordinary buildings. Two boundaries recur across codes and deserve respect. First, the live-to-dead ratio: when live load is small next to dead, under a third or so, the pattern effect shrinks inside the noise, and codes progressively allow full loading to stand in for the envelope; when live dominates, storage, assembly, heavy occupancies, the patterns are the design. Second, span regularity: the simplified two-arrangement diet assumes roughly equal spans, and a short span beside a long one resurrects cases the simplification dropped, including uplift at the short span’s far support.
One interaction needs naming because it trips people: pattern loading and live load reduction answer different questions and combine carefully. Reduction says the average intensity over a large tributary is unlikely to reach the nominal value; patterning says whatever intensity exists will arrange itself badly. A beam’s pattern cases use the reduced intensity appropriate to each loaded area, and the reduction never erases the arrangement question, a fully reduced live load on alternate spans still beats full loading for mid-span moment whenever the influence line dips negative.
Patterns in the takedown: what changes and what does not
For the vertical takedown, the reassuring half of the story: accumulated column and wall loads are governed, almost always, by the all-spans-loaded case, because a column’s axial influence line is positive nearly everywhere in its tributary stack, so the takedown totals that size columns and footings do not need a pattern sweep, and the full-load run remains the backbone document. The corrections live at the edges of that statement. Reactions on continuous lines shift with pattern, the continuity factors themselves are pattern-dependent, so a first-interior column being sized to the bone deserves the enveloped reaction, not the single-case one. Cantilever-adjacent supports can see their minimum, not maximum, under patterns, and minimum reactions are design numbers too: bearing pads, hold-downs, and overturning checks all care about the lightest credible load, computed with live load absent from the helpful regions, the same logic that makes dead load count at its minimum in governing-combination checks. And transfer structures, which convert many spans’ behaviour into one member’s forces, inherit whichever pattern their collected spans fear, making them the one takedown element that genuinely needs the sweep.
Member design is where the patterns earn their keep daily: continuous beam and slab moments, the hogging steel over supports, the column moments from unbalanced bays. The practical division of labour in StructLoads terms: run the full case for the gravity ledger, run the code’s arrangement set for the members, and let the software envelope them, because the cost of extra cases is a moment of computation, while the cost of a missed pattern is steel in the wrong place.
How many cases is enough
Taken literally, patterning explodes: n spans admit 2ⁿ live arrangements, a ten-span beam offers 1,024. The influence-line insight collapses the count, because only sign regions matter: for each span’s sagging, one alternate-spans case; for each support’s hogging, one adjacent-spans case; so a ten-span beam needs about a dozen arrangements, not a thousand, and the moment distribution era’s coefficient tables were exactly these dozen solved once for equal spans and published. Software re-runs them live, which is why the honest default in any continuous model is the code arrangement set switched on, not because more cases are virtuous, but because the marginal cost is zero and the envelope is the deliverable.
Serviceability rides the same patterns with different numbers: the worst incremental deflection of a span occurs under the same alternate-spans case as its worst sagging moment, evaluated at service rather than factored levels, and a floor checked for vibration or partition cracking under full load only has been checked against the wrong picture. The habit that keeps all of this navigable is labelling: every governing number in the record carries its arrangement’s name, so the checker reproducing the calculation knows that the hogging steel came from the adjacent-spans case and the bearing’s minimum came from the overhang case, and the dead-versus-live split that feeds each arrangement is visible rather than merged.
A worked example: two spans, one number that flips
Take a two-span continuous beam, illustrative round numbers throughout: spans of 8 m, dead load 20 kN/m, live load 15 kN/m. Fully loaded, the mid-span sagging moment in each span is about 0.070 x 35 x 64 = 157 kNm. Now pattern it: live load on one span only gives that span a sagging moment near 0.096 x 15 x 64 + 0.070 x 20 x 64 = 92 + 90 = 182 kNm, sixteen percent above the full-load answer, while the unloaded span’s mid-span moment drops far below its full-load value. The hogging moment over the centre support peaks with both spans loaded, 0.125 x 35 x 64 = 280 kNm, so the two design numbers come from two different worlds: sagging steel from the one-span pattern, hogging steel from the full case.
The reactions tell the takedown side. Fully loaded, the centre support carries 1.25 x 35 x 8 = 350 kN, its maximum, the number the column under it accumulates. But the end supports see their minimum under the one-span pattern: the unloaded span’s end reaction drops toward 0.375 x 20 x 8 minus the prying effect of the loaded neighbour, and with a longer loaded span or a lighter dead load, that number heads toward zero and past it. On this symmetric beam it stays positive; make the loaded span a cantilever’s back-span and it would not. One beam, four design numbers, three arrangements: that ratio is the entire subject in miniature, and it is why the arrangement list belongs in the calculation record rather than in the engineer’s head.
Key takeaways: pattern live loading
Continuity makes load placement matter: influence lines change sign, so live load on alternate spans governs sagging, adjacent spans govern hogging, checkerboards govern flat-slab panels, and overhang-loaded cases create uplift that full loading hides. Codes require the worst arrangement, with simplifications for modest live-to-dead ratios and regular spans. In the takedown, full loading still rules the accumulated column totals; patterns rule member moments, close-run reactions, minimum reactions, and transfer structures. Run both, envelope them, and record which case produced each governing number.
Quick answers
What is pattern live loading?
The practice of placing live load on selected spans rather than everywhere, because on continuous structures the worst demand at each point comes from a specific arrangement: alternate spans for mid-span sagging, adjacent spans for support hogging, checkerboard panels for flat slabs, loaded overhang with empty back-span for cantilever uplift. Influence lines that change sign are the reason full loading is not the worst case, and StructLoads envelopes the arranged cases alongside the full-load takedown backbone.
Why is full loading not always the worst case?
Because loading a span relieves its neighbours: continuity carries an upward prying across shared supports, so the influence line for a mid-span moment is negative on adjacent spans, and live load placed there reduces the moment. Any response whose influence line changes sign is understated by full loading, and some responses, cantilever uplift being the sharpest, exist only in patterned cases and vanish entirely when everything is loaded.
When can pattern loading be skipped?
When the structure is simply supported, where more load means more of everything, or when live load is small beside dead, roughly under a third, where codes allow full loading to represent the envelope on regular spans. The permissions assume near-equal spans: irregular geometry, short spans beside long ones, and any cantilever resurrect the dropped cases. Skipping is a code permission to claim explicitly, not a default to drift into.
Does pattern loading change a load takedown?
Mostly no, at the totals: accumulated column and foundation loads are governed by the all-spans case, because a column’s axial influence is positive across nearly its whole stack. The exceptions matter though: continuity reaction factors shift with pattern, minimum reactions at bearings and hold-downs occur under patterned cases, and transfer structures inherit the worst arrangement of everything they collect. The full run is the backbone; the patterns correct its edges.
When should you not trust the two-arrangement simplification?
Whenever its assumptions break: spans differing by more than the code’s regularity limit, live-to-dead ratios above the threshold, cantilevers anywhere in the run, or transfer members collecting multiple spans. The alternate-and-adjacent diet was calibrated for regular continuous runs, and outside that geometry the full arrangement search, cheap in software, is the only honest envelope, particularly for uplift checks the simplification never generates.