Office-to-Data-Centre Conversion With Low Core Strengths
A 1975 six-storey reinforced concrete office building is being converted to a data centre. The original drawings specify 27.6 MPa concrete and Grade 420 reinforcement. Twelve cores give a mean equivalent 28-day strength of 21.7 MPa with a coefficient of variation of 18 %. A typical interior bay is 6.0 m × 6.0 m with a 200 mm flat-plate slab, and the ground-floor interior columns are 500 mm square with 12 No. 25 mm bars and No. 10 ties at 300 mm centres. The data hall imposes 12 kPa of superimposed load (raised floor, racks and services) against the original 2.4 kPa office live load. Preliminary analysis gives a flexural utilisation of 1.55 in the slab column strip, an axial utilisation of 1.30 in the ground-floor columns, and a punching-shear utilisation of 0.92 at the slab–column connection. The owner wants to retain the existing frame and the programme allows nine months.
1. How should the core results be treated in the structural assessment, and what is the consequence for the strength assumed in the analysis?
Read the worked answer
Cores are the only direct evidence of in-situ strength, and they must be interpreted statistically, not by picking the best number.
With a mean of 21.7 MPa and COV of 18 %, the standard deviation is about 3.9 MPa. For existing-structure assessment the usual approach is to use a statistically conservative value - commonly the mean minus roughly one standard deviation, or the relevant percentile for the number of cores taken - rather than the mean. That puts the design strength for analysis in the region of 17.5–18 MPa, materially below the 21.7 MPa mean and well below the 27.6 MPa on the drawings.
The gap between drawing and reality (about 21 %) is itself significant: it suggests either the specified concrete was not achieved, or the original drawings overstate it. Either way, the original design basis cannot be assumed.
Consequence: the utilisation ratios already quoted must be confirmed as having been computed with the reduced strength. A flexural utilisation of 1.55 computed on 21.7 MPa becomes roughly proportional to 1/√f'c for a steel-governed section but worsens markedly for a concrete-governed one; and punching shear, which scales with √f'c, rises by about 11 % if the strength drops from 21.7 to 17.5 MPa - turning a 0.92 utilisation into an overstress. Punching is the check most sensitive to this and must be re-run before any scheme is chosen.
2. Given the utilisations, compare strengthening the frame against providing an independent structure, and state which you would pursue.
Read the worked answer
The two overloads are different in character, which argues for a mixed strategy rather than one global solution.
The slab (flexure 1.55, punching 0.92). Flexural strengthening of a flat plate by bonded FRP or steel plate is expensive, disrupts the ceiling void, and does nothing for punching unless separately detailed. The pragmatic answer for a data hall is to not carry 12 kPa on the slab at all: install an independent rack-support grid - a steel frame on new pad footings, or a mezzanine - that delivers rack loads directly to the ground or to the columns, and leave the slab carrying only a raised floor and access loads. That removes the flexural problem entirely and drops the punching check back to a comfortable margin.
The columns (axial 1.30). Jacketing a 500 mm column in a live building is intrusive but feasible: a 100 mm reinforced concrete jacket or a steel collar with infill can recover 30–40 % capacity. Alternatively, add new columns on new footings at mid-bay, which also cuts the slab spans.
Recommendation: independent rack-support structure plus column jacketing at the ground floor, retaining the existing frame. Reasoning: it avoids relying on a 1975 slab whose real strength is uncertain, keeps the strengthening work where it is accessible, and fits the nine-month programme far better than a full FRP programme across the floor plates.
3. The No. 10 ties at 300 mm centres and the low concrete strength raise additional concerns. What must be verified before the scheme is committed?
Read the worked answer
Three checks, none of which can be skipped:
- Column hinge/ductility detailing. No. 10 ties at 300 mm in a 500 mm column is roughly 10 mm spacing... specifically 300 mm, which is far in excess of modern confinement requirements and likely exceeds even the older code when the low concrete strength is considered. In a region with any seismic demand, that column is shear-critical and non-ductile. Confirm the site seismicity; if it is significant, jacketing must include proper confinement, and a global lateral-force check of the frame may govern the whole project.
- Development and anchorage. With f'c near 18 MPa, bar development lengths and lap lengths increase substantially compared with the original assumption. Check laps at the column bases and slab reinforcement anchorage; existing laps designed for 27.6 MPa may be inadequate.
- Punching shear with the low strength and no shear reinforcement. A 200 mm flat plate without drop panels or shear studs, at 0.92 utilisation on optimistic strength, has essentially no reserve. If any part of the slab is to carry rack loads, add shear reinforcement at the connections or reduce the tributary load.
Finally, confirm the foundation can take the new column loads and the additional jacketing weight; a 1975 footing designed for 2.4 kPa office load is unlikely to have spare capacity for racks plus a new independent frame.