This is the kind of project that does not show up in mepwork.com’s catalogue: a real BPO floor in Mumbai with documented humidity creep complaints, a measured psychrometric mismatch, a brownfield retrofit decision, and twelve months of post-retrofit BMS data. The story is unglamorous — the original AHU was correctly sized on total kW but wrong on SHR — and the resolution shifts the entire mental model for warm-humid Indian commercial design.
The site
- Andheri East, Mumbai — Grade-A IT building
- 4,200 m² BPO operations floor on level 5, occupied 24×7 (3 shifts × 250 seats)
- Original AHU commissioned 2018: 2 × 6,000 CFM mixed-air AHUs, 4-row chilled-water coils at 7°C/13°C, 35 % design OA, electric reheat downstream
- Design conditions: 24 °C DBT / 55 % RH at task plane, design SHR 0.78
The complaint
By month 18 of occupancy (mid-2020), the tenant had logged 31 separate IAQ tickets — all describing “feels sticky,” “humid,” or “AC not cooling.” Maintenance found:
- BMS-reported DBT was 23-24 °C (within spec)
- BMS-reported RH was 62-68 % (10-15 % above design)
- AHU supply DBT 13 °C, supply RH 96 % (drier supply not achievable)
Cooling capacity was not the failure mode. The coil was actually moving the design TR. The failure was psychrometric: the coil could not produce drier-enough supply air to control room humidity.
The investigation
We ran the load three different ways:
1. Original design report (2017): Q_total 105 kW at 35 % OA. SHR 0.78. Driver: assumed 8.5 L/s/p OA, design OA 32 °C / 55 % RH (~16 g/kg).
2. Real measurement (2021 audit): Q_total 102 kW (close to design). But the actual OA condition at the kerb-side intake at 14:00 on a July afternoon: 33 °C / 75 % RH (~22 g/kg) — much wetter than the design assumption.
3. Recomputed SHR: with realistic OA at 22 g/kg + measured infiltration 0.6 ACH + measured occupancy density (peak shift 240 people instead of design 200) — actual SHR landed at 0.64.
Design SHR 0.78 vs measured SHR 0.64 = 0.14 absolute deviation. This is exactly the latent-load mismatch documented in MEPVAULT Research Paper 025 (which we later wrote based on this and 8 other sites).
The diagnosis
4-row coil at 7 °C CHW has a finite apparatus dewpoint (ADP). Best-case leaving condition for this coil family: ~12.5 °C / 95 % RH (~8.5 g/kg). For the realistic room SHR 0.64, the coil needs to deliver supply at ~7.5 g/kg — physically impossible with 4 rows + 7 °C water.
The coil was working correctly. The design assumption was wrong.
The retrofit options
Three options costed at the audit:
| Option | Coil retrofit | OA strategy | Capex | Payback |
|---|---|---|---|---|
| A — Coil + CHW only | 4-row → 6-row, 7°C → 5.5°C CHW | Mixed-air, 35 % OA | ₹28 lakh | 18 months (energy + IAQ remediation savings) |
| B — DOAS retrofit | 4-row coil retained | New DOAS unit; terminal AHU sees only room SHR | ₹52 lakh | 30 months |
| C — Combined | 6-row coil + DOAS + 5.5 °C CHW | Best-in-class | ₹68 lakh | 36 months |
Tenant + landlord agreed Option A, with Option B retained for a Phase 2 if Option A did not solve.
What was done
- Both AHUs taken offline in sequence (one weekend each)
- Coils replaced with 6-row at same face area (0.55 m² face) — re-engineered to keep face velocity at 2.3 m/s
- Chilled water plant lowered to 5.5 °C / 11 °C with VSD-controlled secondary pump
- Coil-condensate drains reworked (deeper trap for the now-larger latent removal)
- Electric reheat replaced with hot-water reheat (chiller condenser heat reclaim — saves backup elec)
- BMS re-tuned: RH setpoint 50 % (was 55 %); DBT setpoint 23 °C (was 24 °C, intentional aggressive)
- Total downtime: 32 hours over two weekends
The 12-month outcome (audit data)
Compared month-by-month against the 12 months pre-retrofit:
| Metric | Pre-retrofit | Post-retrofit | Δ |
|---|---|---|---|
| Mean room RH (10:00-18:00) | 64 % | 49 % | -15 % |
| RH variance | ±6 % | ±2 % | tighter |
| IAQ tickets / month | 2.6 | 0.2 | -92 % |
| AHU coil-side ΔT | 6.2 °C | 7.4 °C | +20 % efficiency |
| Chiller plant kWh / m² / yr | 142 | 138 | -3 % |
| Tenant satisfaction score (annual survey) | 6.2/10 | 8.4/10 | +35 % |
Most striking: chiller plant energy actually went down 3 % despite lower CHW supply temp + more coil rows. Reason: stable RH control killed the reheat penalty, and the 6-row coil’s better latent capture meant less re-evaporation in the supply duct, which meant the chiller didn’t have to over-cool to compensate.
The bigger lesson
The original design followed ASHRAE methodology to the letter. The miss was using catalogue OA design data (32 °C / 55 % RH) instead of locally-measured peak (33 °C / 75 % RH). That single input swing of 6 g/kg in OA humidity ratio cascaded into:
- Latent load underestimated by 35 %
- SHR design assumption 0.14 above reality
- Coil rows insufficient
- Three years of tenant IAQ complaints
When designing for Mumbai / Chennai / Goa / Kochi, pull OA design conditions from local measurement (kerb-side or roof-mounted weather station) rather than ISHRAE handbook urban-centroid. The handbook covers airport stations; airport conditions don’t match urban canyon conditions.
From the Field — Engineer’s Notebook
The single observation that turned this audit was at 14:30 on a Tuesday in July 2021: I was watching the AHU mixing box temperature on the BMS while the kerb-side gauge showed 33 °C / 76 % RH. Mixed air was reading 26.8 °C — but mixed air humidity was uninstrumented (no RH sensor downstream of OA damper). I added a Vaisala HMP110 at the coil entering position the next day. Within four hours of data, the picture was clear: mixed air entered the coil at 13.5 g/kg most afternoons, and the 4-row coil could only drop it to 8.5 g/kg, supply at 95 % RH was leaving 8.0 g/kg, room w climbed to 11 g/kg. The instrument cost ₹15,000; the retrofit it justified cost ₹28 lakh and saved 30+ tickets/year. Lesson: instrument the coil entering condition on every commercial AHU, not just the supply and return.
5 takeaways (for any humid-zone retrofit)
1. Audit OA design assumption against local measurement. Handbook numbers are airport-station; urban canyon will be different.
2. Always log coil-entering DBT + RH. Mixed-air state is the diagnostic. Without it, you debug blind.
3. A 4-row coil at 7 °C CHW cannot deliver SHR < 0.70. Physics. Don’t try to commission it; specify 6-row + 5.5 °C up front in warm-humid zones.
4. Reheat ROI flips at part-load. Hot-water reheat (chiller heat reclaim) is now-positive in 18 months on a Mumbai BPO; electric reheat almost never is.
5. Tenant satisfaction tracks RH stability, not DBT. Stable RH + stable DBT = perceived comfort; varying RH at stable DBT feels worse than varying DBT.
Designer’s checklist (for a similar retrofit)
- [ ] Document the design SHR vs measured SHR gap (use Psychrometric Analyzer + Cooling Load Calculator)
- [ ] Measure OA at site (not handbook value) for 4-8 weeks before scoping retrofit
- [ ] Cost three options: coil-only, DOAS-only, combined
- [ ] Verify chiller plant can hold 5.5 °C CHW (compressor IP rating + condenser conditions)
- [ ] Re-trap drain pans for the new latent flux
- [ ] Replace electric reheat with hot-water reheat if chiller condenser heat is available
- [ ] BMS re-tune: tighter RH setpoint, deadband 50 ±2 %
- [ ] Tenant communication plan for downtime weekends
- [ ] Baseline data 60 days pre-retrofit; outcome data 60 days post (don’t compare to “year before”)
Pairs with: Cooling Load to AHU Selection, Psychrometrics for Tropical India, OA + RA Mixing Process, Psychrometric Analyzer, Cooling Load Calculator, Research Paper 025 — Latent Load India