Steel by diameter, per element and per floor — with an optimised cutting plan from standard 12 m commercial bars. Cutting waste between 2% and 5%, written out by floor and by diameter.
Ordering rebar from rough estimates means 10-15% extra stock, on-site scrap, and cash-flow risk. Cubica extracts steel directly from the model by diameter, element and floor, and generates an optimised cutting plan — each commercial bar used to the maximum, minimal remainders, and orders calibrated to reality.
Excerpt from Armaduras.xlsx — simplified example.
Reference project: residential building · 44 units · 7 floors + basement · 2 145 m² plot.
| Item | Unit | Quantity |
|---|---|---|
| Steel by origin | ||
| Modelled (specified by the designer) | kg | 87 800 (40%) |
| Inferred (base mesh where no rebar is drawn) | kg | 131 600 (60%) |
| TOTAL steel required (stays in the building) | kg | 219 400 |
| Ø | Required (kg) | To buy (kg) | 12 m bars |
|---|---|---|---|
| Ø 8 | 18 600 | 18 950 | 3 999 |
| Ø 10 | 26 700 | 27 210 | 3 675 |
| Ø 12 | 35 200 | 35 870 | 3 366 |
| Ø 16 | 64 100 | 65 320 | 3 447 |
| Ø 20 | 46 100 | 46 980 | 1 587 |
| Ø 25 | 28 700 | 29 250 | 632 |
| TOTAL | 219 400 | 223 580 | 16 706 |
"To buy" already includes splices (50Ø rule) and the 12 m mother-bar cutting waste — this is the column you order from the supplier. The +1.9% in this sample comes from a cutting plan solved floor by floor; the plan we deliver solves the whole building at once and measures 2% to 5% waste, depending on the mix of bar lengths in the project. Includes: rebar for slabs, foundation slabs and footings (specified + base mesh). Also includes: columns, walls, beams and foundation walls — modelled rebar where it exists; EC2 minimums (§9.2, §9.5, §9.6) only on an explicit decision by the estimator, and always flagged as an estimate.
The "Inferred" portion follows the minimum-reinforcement rule for that element — it is not the structural engineer's design calculation. Where reinforcement has not yet been drawn in the model, confirm this quantity with the engineer before ordering.
The distinction between "Modelled" and "Inferred" follows documented industry practice, not an internal convention: the BIMForum Level of Development specification only classifies rebar as explicit geometry from LOD 400 onward — at earlier levels, more common in construction-issue design, rebar usually isn't drawn bar by bar. A recent technical article estimates that less than 10% of rebar ever gets modelled before fabrication in real practice. That's why most models we receive include an "Inferred" portion — it isn't a limitation of our engine, it's the industry's normal state.
The split by chapter (ordinary rebar, starter bars, mesh) and the logic of "what counts under each item" is not a rule we invented: it follows the convention documented in LNEC's Regras de Medição na Construção (CS 26) — the most widely used Portuguese reference for what belongs, and what doesn't, in each measurement chapter. It's not a guarantee any given provider follows it; it's an external yardstick any client can ask to check against.
| Floor | Steel (kg) |
|---|---|
| Basement | 34 200 |
| Ground floor | 28 400 |
| Floor 1 | 27 900 |
| Floor 2 | 27 300 |
| Floor 3 | 26 700 |
| Floor 4 | 25 900 |
| Floor 5 | 25 100 |
| Floor 6 | 23 900 |
| TOTAL | 219 400 |
Bar left over from cutting one floor can often still be used on the next — instead of every floor buying new bar from scratch, usable offcuts carry over into the following phase.
| Floor | Without optimisation | With phasing | Savings |
|---|---|---|---|
| Basement | 2 605 | 2 605 | — |
| Ground floor | 2 162 | 2 119 | 43 |
| Floor 1 | 2 124 | 2 060 | 64 |
| Floor 2 | 2 079 | 1 996 | 83 |
| Floor 3 | 2 033 | 1 931 | 102 |
| Floor 4 | 1 972 | 1 854 | 118 |
| Floor 5 | 1 911 | 1 777 | 134 |
| Floor 6 | 1 820 | 1 656 | 164 |
| TOTAL bars | 16 706 | 15 998 | 708 |
This sample needs 708 fewer bars — 4.2% of what it would take without a consolidated plan. It is the difference between the two columns, checkable row by row. The cut is planned over the real set of pieces, rather than improvised as the job advances. Where the saving lands depends on the project's mix of lengths.
| Ø | l0 EC2 (m) | 50Ø splice used (m) | Check |
|---|---|---|---|
| Ø 8 | 0,40 | 0,40 | OK |
| Ø 10 | 0,50 | 0,50 | OK |
| Ø 12 | 0,60 | 0,60 | OK |
| Ø 16 | 0,80 | 0,80 | OK |
| Ø 20 | 1,00 | 1,00 | OK |
| Ø 25 | 1,25 | 1,25 | OK |
Assumptions: C30/37 concrete, A500 steel, 50Ø splice rule. Informative check — does not replace the structural engineer's validation.
Every value is auditable back to the source element — filter by floor, by diameter, by element. At each design revision it reprocesses with no manual remeasuring.
1
We read the building's Revit/IFC model, including the reinforcement the structural engineer has already explicitly drawn on some elements.
2
We distinguish "Modelled" steel (what the engineer specified) from "Inferred" steel (the base mesh we fill in automatically wherever no reinforcement is drawn, following the minimum-reinforcement rule — a code-based estimate, not the engineer's calculation). Inference applies to slabs, mat foundations and footings; on columns, walls, beams and foundation walls we measure the modelled reinforcement where it exists.
3
We total the steel by diameter (Ø8 to Ø25) and calculate how many 12m bars to order, already accounting for splices and cutting waste, and checked against Eurocode 2.
4
You receive a quantity takeoff by element, auditable back to the source element in the model, with the cutting plan already consolidated across the whole building and the purchase sequenced floor by floor — actual savings vary per project, depending on geometry and available offcuts.
A percentage with nothing to compare it to means nothing. We simulated seven cutting regimes over the same real set of pieces, extracted from real structural designs, and compared them with the plan we deliver.
And what this is not. It is not our algorithm cutting better than a good steel fixer: against someone who sorted all those pieces by diameter and length with every offcut in plain sight, the gain would be near zero. What nobody does by hand is that — keeping that inventory across a whole building, before anyone cuts the first bar. That is where the value is, and that is how we put it.
These figures come from simulation, not from site observation: seven cutting regimes compared over the same pieces of real projects. We did not measure sites.
Every number in this section comes from real measurements and can be substantiated on request. What we have not measured, we do not claim.
Cubica measures every modelled element and returns the 3D/DXF of everything measured, so each quantity can be checked against its geometric origin. The scope of the measurement is, by definition, the content of the model supplied: elements that were never modelled, or modelled incorrectly upstream, cannot be measured by any extraction method — which is why we report in writing the omissions and inconsistencies we find along the way.
The service is delivered from the model and documentation supplied, without a site visit: we measure what that information represents. Existing conditions it does not represent therefore fall outside the scope — changes made on site, deterioration, concealed elements and pre-existing works. Where a divergence between the model and what is built is identified, we reprocess the measurement on the corrected information.
Send us your BIM model — we'll come back with a sample in 1–5 working days.