The AHU That Will Not Dehumidify
A 30,000 L/s air handling unit serves an open-plan office designed for 24 °C and 50% RH with a 500 kW coil load. Chilled water design is 6.5 °C supply, 13.5 °C return, 17 L/s. Commissioning records show 17 L/s through the coil circuit, 6.5 °C entering and 10.2 °C leaving, supply air 15.5 °C, room conditions 24.5 °C and 66% RH, and condensation forming on the diffuser vanes. The chiller's evaporator leaving water is at setpoint and one of its two compressors is loaded to 40%. The coil data sheet gives a design leaving-water to leaving-air approach of 1.5 K. The BMS shows the coil valve at 100% open and the outside-air damper at its 15% minimum.
1. Perform the water-side energy balance and use the approach temperature to locate the fault.
Read the worked answer
Water-side capacity
Q = ṁ · c_p · ΔT = 17 L/s × 4.19 kJ/kg·K × (10.2 − 6.5) K ≈ 264 kW
That is 53% of the 500 kW design capacity, with the coil valve wide open. The plant is not the problem: the chiller is holding setpoint and one compressor at 40% load is producing all the chilled water the coil can use.
Approach temperature
Measured: leaving air 15.5 °C − leaving water 10.2 °C = 5.3 K. Design: 1.5 K. The approach is 3.5 times worse than design.
What that tells you
The approach is the temperature gap the coil needs to drive heat across the water film, the tube wall, the fin bond and the air film. Flow is at design and the entering water is at design, so the driving potential at the inlet is correct. If the coil were simply undersized or the load were low, the approach would stay near design and the capacity would track the load. A collapsed approach with full flow points to a degraded water-side heat path: fouling or scale inside the tubes, air binding in a high point of the coil, a partially blocked strainer or valve on the coil branch, or a bypass around the coil circuit that makes the measured 17 L/s not the 17 L/s actually going through the tubes.
The high space humidity follows from the same fault. A fouled water side raises the coil surface temperature toward the air temperature, so less of the coil falls below the air dew point (about 18 °C at 24.5 °C and 66% RH). Capacity drops and latent removal drops faster than sensible removal - hence a room that is nearly the right temperature and far too humid. The diffuser condensation is the visible symptom of that humidity, not a separate defect.
2. List the diagnostic tests you would run, in order, and state what each result would tell you.
Read the worked answer
1. Coil water-side differential pressure at the measured flow. Compare with the data sheet. A high ΔP means a restriction - fouling, scale, a partly closed valve, a blocked strainer. A low ΔP means the water is bypassing the coil tubes, which changes the whole diagnosis.
2. Strainer and valve differential pressure. The BMS says the valve is 100% open; verify the actuator actually is. A command is not a position.
3. Air venting and water-side pressure profile. Crack the high-point vent at the coil. If the coil is air-bound, the top rows are doing nothing and the effective surface is reduced exactly as observed.
4. Air-side energy balance. Measure entering and leaving air dry-bulb and wet-bulb and the airflow, and split the total into sensible and latent. The water side says 264 kW; the air side should agree within a few percent. If it does not, air is bypassing the coil - check the filter frames, the blanking panels and the coil baffles.
5. Coil face velocity and filter ΔP. Low face velocity or a torn filter gasket can short-circuit air around the coil.
6. Piping arrangement. Confirm counterflow: water in at the air-off end. A piped-in-parallel coil loses mean temperature difference quietly and permanently.
7. Verify the water temperature at the coil, not at the chiller. Long runs in warm plant rooms can pick up heat.
Sequence matters: start with the cheap non-intrusive checks (valve position, vent, strainer ΔP) before pulling the coil for cleaning. On the evidence so far, fouling or air binding is the leading hypothesis and the water-side ΔP will separate them in ten minutes.
3. The building owner proposes lowering the chilled water supply from 6.5 °C to 4 °C to "get the humidity down". Evaluate the proposal.
Read the worked answer
It would work, in the short term, and it is still the wrong answer.
Lowering the supply temperature increases the coil's mean temperature difference, so the same coil with the same fouled water side would deliver more capacity and a colder surface. If the surface drops below the dew point along more of the coil, latent removal recovers and the space RH falls. The improvement might be 15–20% of capacity - worth having during a heat wave.
But:
- It hides the fault. The coil has lost 47% of its design capacity. Lowering the supply temperature buys back part of that at the chiller instead of repairing it, and the fouling will keep growing. You will be back here at 3 °C.
- Evaporator risk. At 4 °C leaving water the refrigerant is only a degree or two colder. On a part-loaded machine, low evaporator temperature plus a low load is the classic route to a freeze-up and a burst tube.
- Efficiency. Chiller COP falls roughly 2–3% per kelvin of lower evaporator temperature. Permanently shifting 2.5 K costs the owner real money in every operating hour, for a defect that is fixable.
- It does not fix the air side. If the investigation shows air bypass or a dry coil, colder water changes nothing at all.
Recommended position: investigate and repair the coil first - verify water-side ΔP, vent it, clean it, correct the piping. If the space cannot be held while that work is scheduled, accept a temporary setpoint reduction to 5 °C with the chiller's low-temperature cut-out verified and a daily log of evaporator approach, and write it down as a temporary measure with an end date. The permanent setpoint stays where the design put it.