Warm Upper Floors and a Low Plant Delta-T: Capacity or Distribution?
A 12-storey office tower has levels 5–8 served by one AHU, design 24,000 m³/h through a 105 kW chilled-water coil at 7/12.5 °C, with two-way control valves throughout. Occupants on the upper floors complain of afternoon heat and stuffiness. Measured at 15:00 on a design day: supply air 23,500 m³/h at 15.5 °C, return air 25.0 °C, return CO₂ 1,100 ppm, AHU fan at 50 Hz on the VFD. The chiller plant reads a running ΔT of 3.2 K against a 6 K design, with two of three chillers running. The operator has recently reset the chilled-water supply from 7 to 8 °C 'to save energy'. Testing and adjusting was never completed for levels 5–8, the AHU chilled-water strainer has not been cleaned since handover, and the outdoor-air damper appears to be sitting at its minimum position.
1. Quantify what the coil is actually delivering and identify the most probable cause, distinguishing between the candidate mechanisms.
Read the worked answer
Actual coil duty (sensible):
- Mass flow = 23,500 × 1.2 / 3600 = 7.83 kg/s
- Q = 7.83 × 1.005 × (25.0 − 15.5) = 74.8 kW, about 71 % of the 105 kW design
- Design chilled-water flow = 105/(4.19 × 5.5) = 4.56 kg/s ≈ 16.4 m³/h
So the coil is short by roughly 30 kW while the fan is at full speed. That is a heat-transfer failure, not an airflow failure - extra air would show as a smaller air-side ΔT with more total capacity, and 23,500 m³/h is essentially design.
Mechanisms and the measurement that discriminates them:
- Chilled-water starvation (fouled strainer, air-bound coil, failed or throttled two-way valve, control valve stuck) - check ΔP across the strainer and coil, and the coil water ΔT. If the coil ΔT is well above 5.5 K at low flow, it is starvation.
- Coil fouled externally (filter bypass, dirty fins) - check the coil face velocity and the air-side ΔP; a high air-side ΔP with a normal water ΔT points here.
- Supply-temperature reset - going from 7 to 8 °C costs roughly 4 % of coil capacity per kelvin of supply temperature at fixed flow, i.e. about 4 kW here. Real, but not the 30 kW gap.
- Over-flow elsewhere - a 3.2 K plant ΔT with cold supply means water is passing somewhere without doing work (bypass, decoupler, over-open near-zone valves). This explains why the far coils starve while the plant looks 'full'.
Diagnosis: the plant ΔT and the coil shortfall together rule out insufficient chiller capacity. This is a distribution and commissioning fault, aggravated by the premature reset.
2. The client's consultant has submitted a variation for an additional 200 kW chiller to 'restore resilience'. Advise the client.
Read the worked answer
Reject the variation as premature and technically unjustified.
Reasons:
- Two of three chillers are running with a 3.2 K plant ΔT against 6 K design. The plant is delivering roughly half the useful cooling per unit of water it pumps; adding a chiller adds capital to a symptom and will lower the ΔT further.
- The served coil is at 71 % of design duty with the fan at full speed - a coil that is not being fed is not evidence of a shortage of source capacity.
- The 1,100 ppm CO₂ shows the outdoor-air damper is at minimum, so the space is under-ventilated as well as under-cooled. That is a further symptom of an unbalanced/never-commissioned system, not of chiller capacity.
- There is no measured load data supporting 200 kW. Nameplate capacity is not load.
What I would instruct instead: complete the testing and adjusting for levels 5–8, clean the strainer, prove the coil water flow and ΔT against the 16.4 m³/h / 5.5 K design, verify the outdoor-air damper position and the CO₂ response, and restore the 7 °C supply with a proper reset strategy. Then re-measure on a design day. If, and only if, the plant is shown to be capacity-limited at design conditions, revisit the chiller - with data, not with a complaint log.
If the client insists on spending money now, the same sum buys far more useful cooling in hydronic balancing, PICVs and controls.
3. What permanent design and commissioning changes would you specify so this does not recur?
Read the worked answer
- Pressure-independent control valves at every coil with a stated maximum flow, so one coil cannot rob the riser. Two-way valves alone do not guarantee a design ΔT.
- A design-ΔT specification with a contractual consequence: coils specified at 6 K, and the TAB scope required to demonstrate plant ΔT at design and part load, not just terminal airflows.
- Chilled-water supply reset governed by the critical-zone valve position, not by a schedule or a fixed offset. Reset that is not feedback-controlled from the worst-case zone will always trade comfort for a paper saving.
- A decoupler/bypass with a flow meter and alarm, or elimination of the decoupler in favour of a variable-primary arrangement, so over-flow is visible rather than hidden as a low ΔT.
- Strainers with differential-pressure switches and a maintenance regime, plus air vents at coil high points.
- Outdoor-air measurement (flow station) with a CO₂-based demand-controlled ventilation limit, so minimum fresh air cannot silently drift.
- A commissioning gate: practical completion of levels 5–8 should not have been certified without the TAB record - the certification process, not the hardware, was the first failure here.