Chilled Water Delta-T: The Key to Better kW/RT

Delta-T is the lever that decides your kW/RT

Ask why a chiller plant runs at 1.0 kW/RT when its chillers are rated near 0.6, and the answer is usually one number: delta-T, the temperature difference between the chilled water leaving the plant and the water returning to it. Design delta-T for most Malaysian plants is around 5.5°C (a roughly 6°C / 12°C supply-return split). When the real return water comes back only 2–3°C warmer than it left, the plant has low delta-T, and low delta-T quietly wrecks efficiency.

This matters because delta-T and pump energy are inversely linked. The cooling a plant delivers is flow multiplied by delta-T. If delta-T halves, the pumps must move twice the water to deliver the same cooling — and pump power rises sharply with flow. That extra pumping is pure waste, and it lands directly on your plant kW/RT.

Why low delta-T drives kW/RT up

Low delta-T hurts efficiency on two fronts at once.

The obvious one is pumping. To compensate for weak delta-T, operators (or automatic controls) ramp up chilled water flow. On a constant-speed system that means running more pumps; on a variable system it pushes the variable speed drives toward full output. Either way, pump energy that should be a small slice of the plant becomes a large one, and every one of those kilowatts shows up in kW/RT.

The less obvious one is the chillers themselves. Low delta-T usually forces a plant to run more chillers at low load to satisfy the flow demand, even though the actual cooling load is modest. Each running chiller carries fixed parasitic losses, so a plant staging three machines at 40% when two at 70% would do the same job burns energy for nothing. Low delta-T is therefore also a staging problem, not just a pumping one — which is why it is often called "low delta-T syndrome": one fault that cascades through the whole plant.

How each part of the plant shapes delta-T, cooling, and energy

Delta-T is made and lost across the whole plant, not at one component. Walking the chilled water loop shows how each item moves the temperature split — and your energy bill — in turn.

Cooling coils (AHUs and FCUs) are where delta-T is actually created. The coil pulls heat from the air, warming the return water. A clean, correctly sized coil returns water at design temperature, gives full cooling, and keeps delta-T wide. A fouled, undersized, or air-bound coil cannot transfer enough heat, so water leaves barely warmed — delta-T collapses, the space may still be under-cooled, and the plant over-pumps to compensate. The coil is the first place to look when delta-T is low.

Control valves decide whether the coil's work survives. Two-way valves throttle flow through the coil and hold delta-T as load varies. Three-way valves bypass cold water around the coil straight into the return line, diluting return temperature and crushing delta-T regardless of how good the coil is. Wrong valve sizing or weak authority causes the same dilution in miniature. This single choice often separates a good plant from a poor one.

Chilled water pumps set the flow that delta-T is measured against. Oversized or fixed-speed pumps shove too much water through the coils, so each litre spends too little time absorbing heat and returns cold — low delta-T by brute force, with the pump energy to match. Variable-speed pumps that follow delta-T move only the water the load needs; because pump power rises with roughly the cube of speed, trimming flow returns large energy savings while widening delta-T.

The decoupler or bypass in a primary-secondary plant is a hidden delta-T leak. When primary flow exceeds secondary demand, cold supply water crosses the decoupler directly into the return, lowering the temperature the plant sees and faking a low delta-T that no coil caused. Keeping primary flow matched to secondary load closes the leak.

The chillers set the supply temperature and respond to flow. A chilled water setpoint pushed lower than the building needs narrows the delta-T window and costs compressor energy; raising it where humidity and load allow widens delta-T and improves chiller efficiency at once. Staging matters too — low delta-T forces extra chillers online to satisfy flow, so each machine runs lightly loaded and carries its fixed losses for little cooling in return.

The condenser pumps and cooling towers do not change chilled water delta-T, but they decide how much each ton of cooling costs. A fouled tower or high condenser approach raises condensing temperature and compressor lift, so kW/RT worsens even when delta-T is healthy. Resetting condenser water and tower fans against wet-bulb keeps lift — and energy — down. Good delta-T and good heat rejection are two separate levers, and a strong plant works both.

How to tune for the best kW/RT

Improving kW/RT through delta-T is a measure-diagnose-correct loop, not a one-off setting:

1. Meter flow and temperatures first. You cannot manage delta-T you do not measure. Supply and return temperature, chilled water flow, and power on chillers and pumps are the minimum. Without them, every adjustment below is guesswork.
2. Fix the hydraulics. Convert three-way valves to two-way where the system allows, clean or right-size coils, and correct pressure control so coils get the dwell time to do their work. This is where most of the delta-T is recovered.
3. Match flow to load. Move to variable primary flow or properly tuned variable-speed secondary pumping so the plant moves only the water it needs, and let delta-T — not a fixed flow — drive pump speed.
4. Reset setpoints intelligently. Raise the chilled water supply temperature when the load and humidity allow; a higher setpoint widens delta-T and lifts chiller efficiency at the same time. Reset condenser water against wet-bulb to cut compressor lift.
5. Stage on real load. Bring chillers on and off at the load points that keep each running machine well loaded, so low delta-T stops forcing unnecessary chillers online.
6. Hold it with continuous monitoring. Delta-T drifts as valves wear, coils foul, and load changes. A plant tuned once will slip back within months without a system watching delta-T and kW/RT continuously.

Steps 1 and 6 are why this is a controls-and-data exercise. CobiNeural tracks supply and return temperature, flow, and chiller and pump power to compute live delta-T and plant kW/RT, then drives reset and staging logic through the building automation layer. Our chilled-water plant work — including the EPF HQ district cooling plant and the FINAS chilled water plant — turns exactly this kind of low-delta-T drift into a plant that holds an efficient, verifiable kW/RT.

The bonus: lower maximum demand

Because the chiller plant is usually the largest electrical load in a Malaysian building, tightening delta-T does more than cut energy. Fewer pumps and chillers running at any moment lowers the coincident maximum demand, which under TNB's RP4 tariff (effective 1 July 2025) is billed at roughly RM89/kW per month for general commercial supply. Better delta-T pays back twice — on the kWh you consume and on the peak you are charged for. To see your plant's live delta-T and kW/RT, request a plant assessment.