Atrium smoke control and a mezzanine variation
A 30 m × 20 m atrium rises 18 m through a four-storey retail podium in Sha Tin. Each upper level has an open balcony 2.5 m deep overlooking the atrium; shop fronts are protected by 1-hour construction, but balcony edges are open. The performance-based design uses a 2.5 MW steady design fire on the ground floor, a smoke exhaust rate of 45 m³/s at roof level, and natural make-up air through four 2 m × 2 m low-level openings. The CFD predicts a smoke layer interface at 8 m above ground at 600 s, with visibility on the third-level balcony of 9 m and temperature of 55 °C. The tenant wants to add a 300 m² mezzanine café at the atrium edge, reducing the volume by 15%. The fire engineer proposes to keep the same exhaust rate and design fire, arguing that sprinkler control makes the change minor. The approving authority asks for a review.
1. Critique the design fire and tenability assumptions in the atrium analysis. Is the CFD result acceptable as presented?
Read the worked answer
The result is not acceptable as presented.
Design fire. A 2.5 MW steady fire from time zero is a weak basis for a retail atrium. It ignores the growth phase, the shop fuel load and the possibility of a fire at a balcony edge or in a shop opening onto the atrium. If the building is sprinklered, a 2.5 MW controlled fire may be arguable for one shop, but the analysis must show the growth period before control, the location, and the effect of a failed or delayed sprinkler. A sensitivity case at 5 MW, or a fire in the mezzanine or café, is needed.
Tenability. The predicted visibility of 9 m on the third-level balcony already fails a typical 10 m visibility criterion. Temperature 55 °C is below a 60 °C criterion, but the answer must also report CO, radiant heat flux and smoke layer height at occupied locations. An interface at 8 m is not obviously consistent with 9 m visibility on a balcony at 8 m; the model may be mixing the smoke layer with the balcony space. The 600 s snapshot may not be the worst case.
Make-up air. 45 m³/s through four 2 m × 2 m openings gives about 2.8 m/s, which is plausible, but the analysis must show that the openings are not blocked, that the inlet does not disturb the smoke layer, and that the pressure regime is stable.
Acceptance. The approving authority should require a clear ASET/RSET analysis, not just a CFD snapshot: tenability at each occupied balcony, egress travel and queueing, and sensitivity to fire size, location, door openings and exhaust failure. Peer review and commissioning criteria should be part of the acceptance.
2. The mezzanine reduces the volume by 15% and adds 300 m² of café seating. Advise the fire engineer on what must change in the analysis and what evidence the approving authority should require.
Read the worked answer
The mezzanine is a material change, not a minor variation.
What must change:
- Geometry and smoke reservoir. Rebuild the CFD model with the 300 m² mezzanine, its edge, soffit and any downstands. The 15% volume reduction and the new obstruction change plume entrainment and smoke layer descent.
- Design fire. Add a café fire scenario (seating, cooking, electrical) and a shop fire. Do not retain only the original 2.5 MW ground-floor fire. Include the growth phase and a location under or adjacent to the mezzanine.
- Exhaust and make-up air. Recalculate the required exhaust from the plume and the target smoke layer height, not by keeping 45 m³/s. Check that the make-up air path still works and does not pull smoke across occupied balconies.
- Tenability and egress. Check visibility, temperature, CO and radiant flux on the mezzanine and on the balconies. Check occupant load, travel distance and exit widths for the café.
- Compartmentation and structural fire resistance. The mezzanine may need fire-resisting construction, and its support may affect the atrium structure.
Evidence required by the approving authority: a revised performance-based report with sensitivity cases, a peer-review statement, and commissioning tests for the smoke exhaust and make-up air. The argument that 'sprinkler control makes the change minor' is not sufficient without demonstrating that the sprinkler design, water supply and response cover the new café fuel load.