Low Chilled-Water Delta-T in a Quarry Bay Office Tower
Four identical 1,100 kW water-cooled centrifugal chillers serve a 32-storey Grade A office in Quarry Bay. Design conditions are 7°C supply / 12.5°C return chilled water (ΔT 5.5 K), 32/37°C condenser water, and design plant efficiency 0.62 kW/RT. The BMS trend shows that at 70% building load, all four chillers run at about 70% capacity, chilled-water ΔT is 3.1 K, and plant efficiency is 0.86 kW/RT. All AHU cooling-coil valves report 100% open, and the secondary pump VFDs are at maximum speed. The chilled-water supply temperature is controlled at 7°C. A consultant recommends replacing all four chillers with magnetic-bearing machines to recover the 0.24 kW/RT gap. The building owner is preparing a HK$28 million capital request and wants a decision before the next cooling season. There is no sub-metering of individual AHU coil flows, and the last water treatment report notes moderate iron oxide deposition.
1. Diagnose the most likely causes of the low chilled-water ΔT and explain why replacing the chillers alone may not deliver the claimed savings.
Read the worked answer
What the data mean. At 70% load, the plant should need about 3,080 kW of cooling. With a design ΔT of 5.5 K, design flow is roughly Q/(4.186 x ΔT) = 3,080/(4.186 x 5.5) = 134 kg/s. At an actual ΔT of 3.1 K, the same load needs about 238 kg/s, about 78% more flow. That explains why the secondary pumps are at maximum speed. The chillers are not necessarily the primary fault; they are being forced to run at low ΔT and part load because the water side cannot deliver the design temperature rise.
Likely causes. Coil fouling is consistent with the iron oxide report: a fouled coil transfers less heat at a given water flow, so the control valve opens fully and the water returns cooler than it should, reducing ΔT. Three-way valves stuck open or open bypasses allow water to return without passing through coils; this also reduces ΔT. Oversized pumps, poor coil selection and sensor error can produce the same symptom. AHU valves reporting 100% open does not prove design flow; it proves the control loop is saturated.
Why chiller replacement is not enough. New chillers may have better part-load efficiency, but if the system still demands 238 kg/s at part load, the new machines will also run at low ΔT, with poor evaporator heat transfer and unstable capacity control. The plant efficiency gap of 0.24 kW/RT is partly a system flow problem. Replacing chillers without fixing coils, valves, balancing and pump control risks spending HK$28 million to recover only part of the gap. The existing chillers may be fine at design ΔT; the test should isolate chiller performance from distribution performance.
2. Recommend a cost-effective investigation and retrofit sequence, with a rough estimate of the savings you would expect at each stage.
Read the worked answer
Stage 1: diagnosis (about HK$150,000-250,000, 4-6 weeks). Install temporary ultrasonic or clamp-on flow meters on each chiller and selected AHU branches. Trend valve position, differential pressure across coils, supply and return temperatures. Calibrate temperature sensors. Measure coil approach and water pressure drop. Inspect for stuck three-way valves and open bypasses. This is essential because the BMS data cannot separate coil fouling from valve bypass.
Stage 2: low-cost hydronic fixes (HK$0.8-1.5 million). Clean fouled coils and treat the water side. Replace failed three-way valves with two-way valves, close bypasses, rebalance the critical branches, and reset the secondary differential pressure. If pumps are oversized, trim impellers or change the control curve. Expected result: raise ΔT from 3.1 K toward 4.5-5.0 K. Flow falls by about 30-35%; pump power falls with the cube of flow, so pump energy can fall by about 60-70%. Chiller part-load efficiency should improve from 0.86 toward 0.72-0.76 kW/RT.
Stage 3: chiller sequencing and control (HK$200,000-400,000). Re-tune staging so that chillers are loaded to their efficient range, not all four at 70%. Add condenser-water reset and chilled-water supply temperature reset within constraints. Expected additional saving 3-6%.
Indicative saving. At 3,080 kW average cooling for 3,000 equivalent full-load hours, annual cooling is about 9.2 GWh. At 0.86 kW/RT, input is about 2.26 GWh; at 0.74 kW/RT, about 1.94 GWh; saving about 0.32 GWh, worth about HK$370,000/year at HK$1.15/kWh. Pump savings from reduced flow could add 0.1-0.2 GWh. The exact figure must come from the measurements; the point is that a few million dollars of hydronic work can capture much of the gap before any chiller replacement.
3. What decision would you give the owner on the HK$28 million chiller replacement? Justify your answer in terms of risk and life-cycle cost.
Read the worked answer
Recommendation: do not approve the full replacement now. Approve the diagnostic and hydronic retrofit first, with a decision gate after the next cooling season. Reasons:
- The evidence points to a distribution problem, not a chiller capacity problem. The chillers are running at 70% capacity, not at their limit. Replacing them does not fix the 238 kg/s flow demand at part load.
- The claimed 0.24 kW/RT recovery is not a measured chiller-only gap. It includes pumps, towers, controls and hydraulics. A new chiller may improve full-load efficiency but will still be penalised by low ΔT.
- HK$28 million has a life-cycle cost that must be compared with a staged fix. If the staged fix costs HK$2 million and recovers half the gap, it is clearly better value. If it recovers little, the owner still has the option to replace later, with better data.
- Risk: replacing chillers disrupts the building, may require crane access, electrical upgrades and a new refrigerant regime. Doing it before diagnosing the water side risks repeating the same problem with new equipment.
Decision gate. After the hydronic fixes, re-test the plant at 50%, 70% and 90% load. If the existing chillers still fail to meet 0.70 kW/RT at part load after ΔT is restored, replace the two worst-performing units first. That staged approach protects the owner's capital and gives the consultant a defensible basis for the final decision.