Chilled-Water Plant with Low Delta-T: Add a Chiller or Fix the Loop?
An office campus chilled-water plant has three 400 TR centrifugal chillers and primary-secondary pumping; design supply/return is 6/13 °C at 520 m³/h. On a design-day survey at 85% of design load, the secondary loop measures 750 m³/h with a 4.1 K ΔT, and two secondary pumps run at full speed with the third on standby. All three primary pumps run continuously whenever any chiller is on. Most AHU coils have three-way valves, and many are found at 100% open at part load; the two-way fan-coil valves are also nearly full open. Chiller 3 starts and stops several times each afternoon, and the operator reports that the last chiller cannot be loaded above 60%. The owner's budget request proposes a fourth 400 TR chiller, citing 'no capacity left', rather than investigating the loop.
1. Check the heat balance at the surveyed condition. Quantify the flow and delta-T deviation from design, and estimate the pumping consequence.
Read the worked answer
Design heat flow: 1200 TR × 3.517 kW/TR = 4220 kW. At ΔT = 7 K, design mass flow = 4220/(4.186 × 7) = 144 kg/s, or about 518 m³/h - consistent with the stated 520 m³/h.
At 85% load: Q = 0.85 × 4220 = 3587 kW. With measured ΔT = 4.1 K, implied flow = 3587/(4.186 × 4.1) = 209 kg/s = 752 m³/h. That agrees with the measured 750 m³/h, so the instrumentation is internally consistent.
Flow is therefore about 750/520 = 1.44 times design, while ΔT is only 4.1/7 = 0.59 of design. The loop is moving roughly 45% more water than it should for the heat it is carrying.
Pumping consequence depends on control. If the secondary pumps are constant-speed and excess flow is throttled or bypassed, pump power is at least 45% above design and probably more. If the pumps are on variable speed with a differential-pressure setpoint held constant, power still scales roughly with flow for the same head, so about 45% excess. If the flow could be reduced to design by removing the excess and slowing the pumps on a closed-loop system, power would fall roughly as (520/750)³ = 0.33, a two-thirds reduction. The exact number requires the pump curve, but the direction and order of magnitude are clear.
2. Explain how three-way valves, over-pumping, and chiller staging interact to produce the symptoms, and why a fourth chiller will not restore design capacity.
Read the worked answer
Three-way valves at the AHUs are the central problem. At part load they bypass water around the coil, so the coil sees less temperature rise while the loop still carries full flow. The return water is diluted with cold bypass water, the plant ΔT falls, and the chillers see a low entering-to-leaving temperature difference. To meet the load, the control system keeps more chillers and pumps running, which pushes even more water through the loop and depresses ΔT further - the classic low-delta-T spiral.
The two-way fan-coil valves being nearly full open suggests the loop is also poorly balanced or the coils are undersized/fouled; the hydraulically nearest coils receive excess flow while remote coils are starved. High flow also raises pressure drop and pump energy.
Chiller staging responds to the symptoms. If staging is based on return-water temperature, the diluted return looks warm enough to bring on another machine, but once it starts it cannot load because the evaporator ΔT is small and the leaving-water setpoint is easily met. Chiller 3 short-cycling and being unable to exceed 60% is a sign of poor staging control and low ΔT, not a shortage of nominal capacity.
A fourth chiller adds nominal tons but not useful capacity: the limiting factor is the low ΔT and the hydraulic distribution, and the existing machines are not fully loaded. It would also add capital cost and likely make staging worse. Before adding capacity, the plant should prove that the existing chillers cannot deliver design tons at design flow and temperature.
3. Recommend a prioritised corrective programme and state what must be verified before any capacity addition is approved.
Read the worked answer
A practical sequence:
- Verify the survey instruments and the chiller load readings. Confirm flow meter calibration and that the ΔT sensors are matched; a 0.5 K error matters at these small differences.
- Fix the terminals first. Convert AHU three-way valves to two-way where the coil and control permit, and check that fan-coil valves are not stuck open. Rebalance the secondary loop so remote coils receive design flow. This is the highest-value change and directly restores ΔT.
- Fix the pumping control. Reset differential pressure to the critical zone rather than holding a fixed high head, and stop running all three primary pumps when fewer chillers are on. Primary-secondary decoupling must be checked: a common header with excess primary flow can cause low ΔT and temperature blending.
- Check coils for fouling, air binding, and control-valve authority. A coil that cannot achieve design ΔT at design flow is a terminal problem, not a chiller problem.
- Repair chiller staging: stage on actual load or on a combination of load and leaving-water temperature, with adequate deadband and minimum run times. Investigate the short cycling on Chiller 3, including minimum load and condenser conditions.
- Re-survey after these changes. Only then calculate the true available capacity at design conditions and compare it with the load growth projection.
Approval of a fourth chiller should require evidence that, after the terminals and pumping are corrected, the existing plant cannot meet design load with one standby machine and acceptable staging. Otherwise the owner will pay for capacity that the distribution system still cannot deliver.