Excessive Deflection of a Post-Tensioned Cantilever Balcony
The owner of a six-storey apartment building in Kandy complains that a fourth-floor balcony 'feels low'. The balconies are 2.4 m cantilevers, 200 mm thick post-tensioned slabs cast four years ago, supported on a core wall at the rear edge. Measured deflection at the tip is 35 mm. The balustrade is cracked at the slab junction, there is a horizontal crack up to 0.4 mm along the top surface near the rear support, and there is a damp patch on the soffit of the same balcony. The as-built information shows 15.2 mm diameter unbonded strands at 1.0 m spacing with the high point at the rear support, concrete cube strength 30 MPa, and no stressing records for three of the fourteen tendons on this floor. A photograph taken during construction shows one tendon end on the fifth floor filled with grout but with no stressing tail, suggesting it was never stressed. The balcony currently carries a tiled finish, and the owner mentions that around thirty people were standing on it during a family function last month.
1. Assess whether the balcony is safe to use now and state the interim measures you would order.
Read the worked answer
Is 35 mm acceptable? No. For a cantilever, the usual total long-term deflection limit to avoid visible damage to finishes and balustrades is of the order of L/250 for appearance and L/180 where partitions or brittle elements are attached. For a 2.4 m cantilever that is 9.6 mm to 13.3 mm. The measured 35 mm is about L/69, roughly three times the limit. The design dead load alone would not produce that on a correctly stressed slab, so something is wrong beyond creep and shrinkage.
The crack at the rear support is the more worrying symptom. A 0.4 mm crack along the top surface near the support in a cantilever is where the negative moment demand is highest. Combined with a damp patch on the soffit, it is consistent with water reaching the anchorage zone of an unbonded tendon system, which can corrode the strand. An unbonded strand with a corroded or broken wire loses its force without warning and without external cracking pattern.
Interim measures
- Prohibit access to this balcony and the one directly below it until the investigation is complete; do not rely on a load posting.
- Erect a temporary prop from below to the underside of the cantilever tip, founded on ground-bearing support or the balcony below with load spread, and take the load gently through a jack with a load cell; do not jack to lift the slab.
- Prevent water ingress: seal the top crack and the balustrade junction as a temporary measure.
- Install deflection gauges at the tip and crack gauges, and record the current tip level relative to the floor slab as a datum.
- Inspect every balcony on every floor for the same defect and check the anchorage pockets at the rear edge for corrosion staining.
Conclusion: not safe for normal use with a crowd; the 30-person event was a near miss on a structure already at three times its deflection limit.
2. Set out the investigation programme to establish whether the tendons are effective, explaining what each test tells you and its limitations.
Read the worked answer
1. Documentation first. Recover the stressing records, the anchorage schedule and the tendon friction/wedge-set data for every floor. Once a strand is stressed and the tail cut, the only record of force is the stressing sheet; where those are missing, assume nothing.
2. Lift-off test at the anchorage. The definitive test. A hollow jack at the live end lifts the anchorage and measures the force at which the wedge releases; the gauge reading at that point is the effective prestress. Do it on at least three tendons per floor, including the suspect one. Limitation: it gives the force at the anchorage, not the force distribution along the tendon, and it disturbs the anchorage, so a new wedge and re-seating are needed afterwards.
3. Expose and inspect the anchorage pockets. Remove the grout or mortar plug at the rear edge. Look for a stressing tail. A tendon with a grouted end and no tail was almost certainly never stressed, which means one of roughly three tendons serving a 1.0 m strip of this cantilever may be inactive, so capacity is about 30 percent down locally and the slab will deflect much more than designed.
4. Ground penetrating radar or covermeter along the tendon line. Confirms tendon position and duct alignment; it does not measure force. Useful to check whether the high point is where the drawings say, because a low high point reduces eccentricity and increases deflection.
5. Tendon condition. This is an unbonded system, so if a duct contains grout at one end only, ask whether the whole system was grouted (making it bonded and subject to a different corrosion mechanism) or whether the grout was just a pocket fill. Open an inspection window at mid-span and at the high point; check for water, voids in the sheathing, and corrosion product on the strand.
6. Level survey over time. A levelling campaign repeated monthly distinguishes a stabilised long-term deflection from progressive movement. Progressive movement plus a damp patch is a corrosion alarm.
7. Concrete strength. Take cores from a non-critical region (not the anchorage zone) to confirm the 30 MPa assumption, using the core results rather than cube results in the structural check.
Order of work: anchorage inspection and lift-off in week 1, because a broken or unstressed tendon changes the whole conclusion; then radar, cores and survey in parallel.
3. Assuming one tendon per balcony was never stressed and corrosion is not significant, outline the structural check and the repair options with their trade-offs.
Read the worked answer
Structural check
Rebuild the model with the confirmed geometry and the measured concrete strength. The key change is that the effective prestress force per metre is reduced by the number of inactive tendons in the strip, and the remaining tendons take a higher force and a longer drape path. Check:
- Serviceability: total long-term deflection including creep and shrinkage, against a limit of about L/180 to L/250. With one of three tendons missing, the section is likely to be cracked under quasi-permanent load, so use a cracked-section analysis, not a gross-section one. If the analysis predicts something close to the measured 35 mm, the model is credible and can be trusted for the next step.
- Ultimate flexure at the support: with reduced prestress, the section relies more on ordinary reinforcement, and the top steel in a PT slab is often only nominal. Check whether the ultimate moment is still adequate; the crack at the support suggests it is close to cracking at service, so the margin matters.
- Punching and the balustrade fixing: a heavy balustrade on a deflecting slab attracts load and cracks at the junction; check the fixings for the imposed deformation.
- Vibration and comfort: a 2.4 m cantilever with reduced stiffness may be uncomfortably lively under crowd loading.
Repair options
| Option | Advantage | Trade-off |
|---|---|---|
| Permanent steel prop from below to ground | Simple, verifiable, cheap | Six storeys of prop stack; obstruction below; new foundations; architectural impact |
| Permanent prop onto the balcony below, with a designed transfer | No ground works | Loads the lower slab and changes its behaviour; needs a chain of checks |
| Add external post-tensioning at the rear edge | Restores stiffness without props | Anchorage detailing into a built slab is difficult; durability; needs a specialist |
| Carbon fibre flexural strengthening at the top of the support | Minimal thickness | Only helps strength if the section is cracked; does little for deflection |
| Demolish and recast the balcony | Certainty | Disruption, propping of the opening, cost |
Recommendation: if the check shows adequate ultimate capacity and the deflection is stable, a permanent prop plus sealing of the crack and re-inspection is the proportionate response. If the check shows the ultimate margin is thin, or if lift-off confirms wider tendon loss, replace the balcony.