The decision between Primary-Secondary (P-S) and Primary-Only-Variable (P-O-V) chiller plant configuration determines pump energy, control complexity, and “low ΔT” operational risk for the building’s chilled water system. P-S has been the industry default since the 1970s; P-O-V has gained ground since 2010 with reliable variable-frequency-drive (VFD) and chiller-flow control.
For Indian commercial cooling towers and chiller plants, the choice carries 15-30% differential in pump-energy operating cost across a typical year. This pillar walks the comparison: when each makes sense, the design checks each requires, and the operational pitfalls.
Primary-Secondary architecture
The traditional plant: a primary loop with constant-speed pumps (CSPs) circulating water through chillers at near-design flow, and a separate secondary loop with variable-speed pumps (VSPs) responding to building cooling demand.
A bypass pipe (“decoupler”) connects supply and return of the two loops, allowing primary and secondary flow rates to differ. Excess primary flow returns to chillers via decoupler when secondary demand is low; excess secondary flow draws from decoupler when secondary demand is high.
Chillers Loads
│ │
Primary CSP → ▼ (Decoupler) ▼ ← Secondary VSP
│
(return to chillers)
Pros:
- Constant chiller flow; chiller stable, low risk of low-flow trip
- Chillers stage independently of building demand
- Well-proven; almost any HVAC contractor knows it
Cons:
- Two pump sets (primary + secondary) means ~1.4× pump kW vs single set
- Constant-speed primary pumps run at design flow regardless of load
- Decoupler can develop “negative flow” (secondary draws from supply, defeating chiller staging)
- Low ΔT syndrome (return water hotter than design) common in chilled beam / underfloor systems
Primary-Only-Variable architecture
A single set of variable-speed primary pumps. No decoupler, no secondary loop. Pump speed modulates to match building demand directly.
VSPs → Chillers → Loads → return → VSPs
The chillers must be tolerant of variable flow. Modern centrifugal chillers (post-2010) typically accept 50-100% of design flow without performance degradation; some go to 30%. Older chillers often have minimum 80% design flow requirement, making P-O-V impractical.
Pros:
- Single pump set; ~30-40% lower pump kW
- No decoupler complexity; no negative-flow issue
- Simpler controls
- Lower capex (one less pump set, fewer valves)
Cons:
- Chiller flow varies; risk of low-flow trip if building load drops fast
- Requires chiller flow staging logic (when to add/drop chillers)
- Needs minimum-flow bypass valve to protect chiller during transient low-load
- Less margin for unplanned chiller flow excursion
When each architecture fits
| Project characteristic | P-S | P-O-V |
|---|---|---|
| Chiller minimum flow ≥ 80% design | OK | NO — too restrictive |
| Chiller minimum flow ≤ 50% design | OK | YES |
| Building cooling load swings rapidly (5%/min) | OK | Risky without smart staging |
| Building cooling load is steady | OK | Optimal |
| Chiller plant operator experience: low | OK | Risky |
| Operator experience: high | OK | Optimal |
| Capex priority | OK | Lower capex |
| Opex priority | Medium | Lower opex |
For Indian projects, P-O-V is gaining ground for buildings ≥ 1,500 TR (where 30% pump-kW saving justifies the slight complexity). Below 1,500 TR, P-S remains common because the savings are smaller in absolute rupees.
Low ΔT syndrome — the operational risk
Both architectures suffer from low ΔT syndrome, but it manifests differently. Symptoms:
- Return water temperature rises (e.g. design 12 °C, actual 9 °C)
- Building temperature still maintained → chillers run more
- Chillers cannot reach target leaving water → chiller compressor at full speed at part load
- Plant kW per TR worse than design
Causes:
- Coil 3-way valves (instead of 2-way) → constant flow at part load
- Oversized 2-way valves → poor authority, partial bypass
- Primary-secondary bypass during low-load → “false load” mixed with secondary
- Coil bypass at part load (cooling coil with no reheat)
Cures:
- Specify only 2-way valves, sized for proper authority (Cv selection)
- Ensure coils are sized for design ΔT at design flow
- For P-S: add VFD to primary pumps to vary primary flow with secondary
- For P-O-V: chiller flow staging that responds to actual ΔT
Decoupler design (for P-S systems only)
The decoupler pipe must:
- Be sized for at least 50% of secondary peak flow
- Have minimum length 10 × pipe diameter (for proper mixing)
- Be located so primary supply enters near chiller side, primary return enters near load side
- Be uninsulated (mixing should be visible/observable through pipe colour)
A common error: undersized decoupler. Secondary pump pressure exceeds the decoupler capacity → forces flow direction reversal in decoupler → primary supply mixes back into primary return → chiller “sees” warm supply → unstable operation.
VFD primary pumps (for both architectures)
Whether P-S or P-O-V, modern best practice includes VFDs on primary pumps. VFDs allow:
- Speed reduction at low-load (saves 25-40% pump energy at 60% flow)
- Better matching to actual chiller flow needs
- Soft start/stop (reduces water-hammer events)
Capex: 5-8% of pump capex. Payback typically 2-3 years from energy savings.
Worked comparison: 1,500 TR plant in Mumbai
Plant configuration:
- 3 × 500 TR chillers, water-cooled centrifugal
- Design supply 6.7 °C, return 12.8 °C, ΔT 6.1 K
- Annual cooling-mode hours ≈ 6,500
- Average load factor 60%
- Chilled water flow at design 410 m³/h per chiller, total 1,230 m³/h
P-S configuration:
- Primary: 3 × CSP (one per chiller), each 410 m³/h at 25 m head, ~125 kW total at design
- Secondary: 2 × VSP (one duty, one standby), 1,230 m³/h at 35 m head, ~155 kW at design but ~70 kW at average
- Total pump kW: 125 + 70 = 195 kW
- Annual energy: 195 × 6,500 × 0.85 = ~1.08 GWh
P-O-V configuration:
- 3 × VSP (one per chiller branch), 410 m³/h at 50 m head (combined primary + secondary), ~85 kW per pump at design
- Average pump kW at 60% load: 3 × 85 × 0.6 × 0.6 (cube law) = ~55 kW total
- Annual energy: 55 × 6,500 × 0.85 = ~0.30 GWh
Savings: 0.78 GWh/year × ₹10/kWh = ₹78 lakh/year
Capex difference: P-O-V saves ~₹15-25 lakh capex (no separate primary CSP). Total economic advantage: ₹78 lakh/year operating + ₹15-25 lakh capex saved. Compelling case for P-O-V at this size.
Five common hydronic plant design mistakes
1. 3-way valves at coils. Forces constant flow; defeats variable-flow plant. Use 2-way only with VSP control.
2. Oversized 2-way valves. Poor authority → bypass at low load. Spec valves for 50% authority at design.
3. Decoupler too small. Negative flow in decoupler corrupts chiller staging.
4. Single chiller with no staging. Plant operates at fixed flow regardless of load; pump kW always at peak.
5. No VFD on primary pumps. Constant-speed primary pumps consume 30% more pump kW than necessary.
Quick checklist
- [ ] Chiller minimum flow capability documented (% of design)
- [ ] P-S vs P-O-V decision based on chiller capability + plant size
- [ ] All coil control valves are 2-way with proper authority
- [ ] VFDs on primary pumps regardless of architecture
- [ ] Decoupler (if P-S): sized ≥50% of secondary peak, length 10× pipe diameter
- [ ] Chiller flow staging logic specified (P-O-V only)
- [ ] Minimum-flow bypass valve (P-O-V) to protect chiller during transients
- [ ] Annual energy comparison documented for owner
References: ASHRAE Handbook HVAC Sys & Eqp 2024 Ch 13 (Hydronic Heating and Cooling); ASHRAE 90.1-2022 §6.5 (Pumping); ISHRAE Handbook 2024 Ch 11; AHR Expo presentations on chiller plant optimization.
[Disclosure block, Legal notice — auto-included by article template]