A vertical extension is a load takedown run twice: once to discover what the existing building actually carries, and once to prove it can carry more, and StructLoads earns its keep here because the entire project is decided by the comparison between those two runs, column by column, wall by wall, footing by footing. Adding storeys is the one job where the structure’s history is a design input: the reserves you hope to build on hide in old design decisions, unclaimed live load reductions, a heavy roof about to be removed, conservative old span tables, and the deficits hide in the same archive, corrosion, undocumented alterations, a garage that was never meant to hold up apartments. The workflow that keeps the project honest has a fixed order: establish the existing loads, establish the existing capacities, place the new storeys, and only then argue about materials.
First run: what the building carries today
The existing takedown starts from evidence, not from the original drawings’ optimism. Measure or verify the real construction: slab thicknesses cored or scanned, wall materials confirmed, finishes and build-ups weighed as found, because a 1960s office that has been through three fit-outs carries superimposed dead loads the original engineer never met, and sometimes lighter ones, screeds removed, partitions gone. Then run the multi-storey accumulation with today’s occupancies and today’s code loads, storey by storey down to the foundations, so every column and wall carries a defensible current demand.
This first run already produces the project’s most important map: the margin map. Old buildings are rarely uniformly stressed; they have lazy columns and busy ones, and the extension wants to know which is which. Three reserves appear so often they are worth hunting deliberately. Live load reduction, if the original design never claimed it, the code’s own reduction provisions applied to the large tributary areas of lower columns can free 10 to 20 percent of live demand legitimately. The roof swap: an extension replaces the existing roof build-up, and a heavy gravel-ballasted or tiled roof being removed refunds real permanent load before the first new beam arrives. And occupancy truth: a floor designed for storage now used as office has been holding a reserve for decades. None of these reserves exists until the takedown demonstrates it; all of them together often decide whether the extension is one storey or three.
Second run: the extension placed, honestly
The new storeys then land on the model as real construction, and the material choice is the biggest load decision on the project. Conventional concrete adds roughly 7 to 10 kPa of dead load per floor; steel framing with metal deck perhaps 3.5 to 5; light steel or cross-laminated timber construction 1.5 to 3, which is why timber has become the default vocabulary of rooftop extensions: two timber storeys can weigh less than one concrete one, and sometimes less than the roof being removed. The comparison that matters is per column: new demand at each existing support versus demonstrated capacity, and the pattern of the answer often redraws the architecture, a lightweight setback storey where the margin map is thin, full floors where it is fat.
Alignment is the quiet decider. New walls and columns that land on existing column lines send their loads straight down the strong paths; a new layout that ignores the old grid manufactures transfer structures at the interface, heavy beams doing expensive work, adding exactly the dead load the project cannot afford in exactly the wrong place. The interface storey deserves its own drawing: how each new post meets an old column head, how the new lateral system’s forces, wind on the taller building, not just the new storeys’ weight, enter the old core and walls, since the extension raises the sail area as well as the gravity, and the bearing walls that quietly held the old roof may now anchor a taller building’s overturning.
| New storey construction | Added dead load per floor | Two storeys add roughly |
|---|---|---|
| Reinforced concrete flat slab | 7-10 kPa | 14-20 kPa |
| Steel frame + composite deck | 3.5-5 kPa | 7-10 kPa |
| CLT / light timber | 1.5-3 kPa | 3-6 kPa |
| Removed old roof build-up | refund of 1-3 kPa | credited once |
The deficits: what the archive holds against you
The same history that hides reserves hides liabilities, and the second half of assessing an existing frame is looking for them on purpose. Undocumented alterations lead the list: openings cut through slabs for services, walls removed in a 1990s refit, a retrofit that strengthened one line by quietly re-routing load onto another. Material condition follows: carbonation and chloride attack in the parking levels, corroded reinforcement at leaking roof edges, timber decay at bearings, each one a local capacity subtraction at exactly the resolution the margin map needs. And code drift cuts both ways: a frame designed to a 1965 code may carry generous gravity margins and, simultaneously, lateral details no current code would accept, so the extension that is fine by gravity can still trigger a seismic assessment the moment the building’s mass and height change.
The practical consequence is a survey scope written by the takedown, not before it: core and scan where the model says the demand is highest, open up the connections the interface storey will rely on, and test materials where a demonstrated strength would formalize the old conservatism into usable numbers. Investigation money spent on the busy columns buys extension capacity; the same money spent uniformly buys a report.
The ground gets the bill
Every kilonewton the extension adds arrives, eventually, at soil that was loaded decades ago and has finished settling under the old building. The foundation reactions from the second run are the document the geotechnical engineer needs: per-footing increases, not a percentage hand-wave, because the increases are as uneven as the margin map above them. Old foundations hold surprises in both directions: pad footings sized generously by rule-of-thumb eras carry real reserves, while foundations that have already used the soil’s easy capacity respond to new load with new settlement, and differential settlement between a heavily extended line and an untouched one cracks the building along the difference.
Where the ground account runs short, the remedies are ranked by disruption: redistribute the extension’s loads toward the lazy columns, lighten the construction further, enlarge footings where access allows, and, last, underpinning, real money and real risk, which the takedown either justifies with numbers or, better, designs away. The same accounting settles the basement columns and any transfer structures on the way down: a vertical extension is not a roof project, it is a foundations project that happens to start in the sky.
A worked example: two timber storeys on a four-storey frame
Take a four-storey 1970s concrete frame, illustrative round numbers throughout: columns on a 6 m grid, each interior column carrying 36 m² per level. The existing takedown, with verified finishes, gives dead 6.5 kPa and live 2.5 kPa per floor, and a heavy roof at 5.0 kPa dead: current interior column demand at the foundation about 1,250 kN. The original design, generously, never claimed live load reduction; claiming it now on four floors of accumulated live load frees roughly 60 kN. Removing the old roof refunds 5.0 x 36 = 180 kN. The margin ledger opens with 240 kN of found capacity before anything is built.
The extension: two CLT storeys at 2.5 kPa dead, 2.0 kPa live, plus a light roof at 1.5 kPa. New demand per interior column: (2.5 + 2.0) x 36 x 2 + 1.5 x 36 = 378 kN, and the live share shrinks under reduction. Net increase after the refunds: roughly 140 kN on 1,250, about 11 percent, the kind of number old pad footings frequently absorb within their original conservatism, subject to the geotechnical check. The same extension in concrete would have added nearly 700 kN net, five times the ask. That ratio, not any preference for timber, is the argument: the material choice moved the project from underpinning territory to paperwork. Whether this building says yes is then a member-level question, and the checked takedown, both runs side by side in StructLoads, is the document that says it defensibly.
One caution closes the example: the 240 kN of found reserves and the 140 kN net ask are averages over the interior columns, and the edge and corner columns run their own ledgers, smaller tributary areas, smaller refunds from the removed roof, and a larger share of the new lateral demands. A vertical extension that passes on the interior grid and fails at a corner is normal, and the fix is usually architectural, pulling the new storeys back from that corner, rather than structural. The margin map has edges, and the extension should be shaped to them.
Key takeaways: load takedown for a vertical extension
Run the takedown twice, existing and extended, and let the comparison govern: the margin map of lazy and busy columns, the reserves found in unclaimed live load reduction, removed roof build-ups, and occupancy truth, against per-footing increases the soil must accept. Choose the new construction by weight, timber’s 1.5 to 3 kPa per floor against concrete’s 7 to 10, align new supports with old column lines to avoid manufacturing transfers, and give the interface storey and the raised wind demands their own designs. The extension is decided at the foundations, and the two takedowns are the evidence.
Quick answers
How do you do a load takedown for a vertical extension?
Twice: first establish the existing building’s real loads from verified construction, materials, finishes, and current occupancies, accumulated storey by storey to the foundations; then add the proposed storeys as real construction and compare demand to capacity per column, wall, and footing. The margin map from the first run, plus reserves from unclaimed live load reduction and the removed roof, decides how much extension the building buys, and StructLoads keeps both runs comparable member by member.
Where do existing buildings hide reserves for extra storeys?
In three recurring places: live load reduction the original design never claimed, worth 10 to 20 percent of accumulated live demand on lower columns; the old roof build-up, whose removal refunds 1 to 3 kPa of permanent load before the extension arrives; and occupancy truth, floors designed for heavier uses than they now serve. Old material conservatism adds unofficial margin that testing can sometimes formalize. None of it exists until a takedown demonstrates it with numbers.
Why are rooftop extensions usually timber or light steel?
Weight arithmetic: CLT and light steel add 1.5 to 3 kPa of dead load per floor against concrete’s 7 to 10, so two timber storeys can weigh less than the heavy roof being removed, keeping per-column increases in the 10 to 15 percent range that existing frames and footings often absorb. The lighter construction also trims the seismic mass and the foundation ask, frequently converting an underpinning project into a verification exercise.
Do the foundations always need strengthening for an extension?
No, but they always need checking: the extension’s per-footing increases, delivered unevenly according to the margin map, land on soil that finished settling decades ago, and the answer ranges from full absorption within old conservatism to underpinning. The ranked remedies are redistribution toward lazy columns, lighter construction, footing enlargement, and underpinning last. Differential settlement between extended and untouched lines is the failure mode to design against, not just total capacity.
When should you not trust the original drawings for an extension?
Always verify before relying: fit-outs change finishes and partitions, alterations cut structure without documentation, corrosion and carbonation erode capacity, and as-built construction deviates from drawings in both directions. Core, scan, and inspect enough to confirm what the first takedown assumes, because the whole project rests on the existing run’s honesty. Drawings are the hypothesis; the building is the evidence, and the takedown should be built from the evidence.