Low ΔT Syndrome in a Two-Chiller Plant on a Design Day
An 800-ton chilled water plant in a 22-story Midtown office tower has two 400-ton centrifugal chillers, constant-speed primary pumps, and a variable-speed secondary loop designed for 1,200 gpm at 44°F supply and 60°F return. On a 93°F August afternoon the BAS trend shows secondary flow at 1,150 gpm, 44°F supply, 57°F return, the lead chiller at 55% and the lag chiller at 29% with two surge alarms logged since morning. South-side tenants have complained of warm space for two weeks; the VAV boxes serving them read 100% airflow with hot water reheat valves open. The plant engineer proposes renting a third chiller to "keep up." The building has a demand-response commitment, and the property manager wants to know why the electric bill rose 18% year over year while occupancy fell 4%. The coil-cleaning contractor's latest report notes "several valves not stroking."
1. Calculate the plant's actual delivered load and compare it with design flow and design ΔT. What does the comparison tell you about where the energy is going, and does it support renting a third chiller?
Read the worked answer
Load: Btu/h = gpm × ΔT × 500 = 1,150 × 13 × 500 = 7,475,000 Btu/h → ÷ 12,000 = 623 tons.
Hmm - recheck the trend rather than trusting the first arithmetic. Supply 44°F, return 57°F gives ΔT = 13°F, not the 7°F implied by a badly calibrated return sensor on some days. If 13°F is real, the load is 623 tons and both machines are legitimately needed. Before anything else, verify the sensors - a single calibrated thermometer strapped to the return header, compared with the BAS value, decides whether this is a control problem or a load problem.
If 13°F is confirmed:
- Load 623 tons = 78% of the 800-ton design.
- Flow 1,150 gpm = 96% of design, and design flow corresponds to 800 tons at 16°F. At 623 tons the loop should be near 780 gpm at a 16°F ΔT. Running 1,150 gpm for a 623-ton load means the secondary pumps are doing roughly 47% more work than necessary - that is the bill increase.
- Chiller 2 at 29% is the smoking gun: it is running below its stable range, which explains the surge alarms, and a surging centrifugal is inefficient and abusive.
Renting a third chiller would make it worse, adding parasitic load and another machine at low load. The correct first move is a sensor check, then coil and valve investigation.
2. Assuming the temperatures are accurate, list the four most probable causes of the low ΔT and describe the specific field check that confirms or eliminates each.
Read the worked answer
Low ΔT means too much water is circulating for the heat actually being absorbed. The classic causes, in order of how often they are the answer:
- Terminal unit valves stuck open or failed actuators. Confirmed by a valve-position survey: walk the south-side AHUs and VAV boxes and read commanded vs. actual position, or feel the leaving-air and leaving-water temperatures across each coil. A coil with a small air-side ΔT and a small water-side ΔT is passing water it does not need. The contractor's "not stroking" note makes this the leading candidate.
- Three-way valves or a bypass/decoupler left open. Confirmed by temperature at the decoupler and by checking the valve schedule against the as-built. Any continuous flow through the bypass returns 44°F water directly to the return header and manufactures low ΔT.
- Differential-pressure sensor error or a defeated DP reset. If the DP sensor reads low, the secondary VFDs ramp to maximum and stay there. Confirmed with a temporary gauge at the sensor location and by comparing the DP trend against valve position - pumps at full speed with valves mostly closed is the signature.
- Fouled or air-bound coils. Confirmed by coil ΔT and by a strainer/air-separator check. A fouled coil cannot transfer the load even at full flow, so the zone stays warm while the plant runs cold.
Sequence the checks cheapest-first: sensor calibration, decoupler, valve survey, then coil condition.
3. State the operating decision you would make today and the 30-day corrective plan, including what you tell the property manager and the demand-response coordinator.
Read the worked answer
Today (if sensors verify):
- Drop the lag chiller only if the lead machine can carry the measured load with margin and the lead is not already near 90%. At 623 tons it cannot, so keep both online but hand the load deliberately: load the lead to roughly 85–90% and let the lag trim, rather than running two machines at 55%/29%. Two machines at low load surge; one loaded machine plus a properly staged second is stable.
- Reset the secondary DP setpoint down in small steps (e.g., 2 psi per hour) while watching the worst-case valve position. Stop when any critical valve reaches about 90% open. Expect flow to fall toward 800–900 gpm and the ΔT to climb.
- Suspend the rental quote.
30-day plan:
- Calibrate the CHW supply and return sensors against a reference and document the result.
- Complete the valve survey on the south side; replace failed actuators and correct any three-way valves.
- Verify the decoupler is closed and the DP sensor reads true.
- Re-trend at the same outdoor condition and compare gpm, ΔT, kW/ton, and pump kW.
Communications: Tell the property manager that the bill increase is pump and chiller energy from excess flow, not occupancy, and give a kW/ton before-and-after. Tell the DR coordinator that committed curtailment capacity is unaffected, but the plant's baseline kW may drop, which changes the settlement baseline - flag it now, not after the event.