Chiller Sizing for Malaysian Humidity: Why Latent Load Breaks Imported Selection Software

Let me start by correcting the title slightly, because the precision matters.

A chiller does not know whether the load it is serving is sensible or latent. It makes cold water. Hand it a duty in kilowatts and a set of water temperatures and it will do the job. Nothing about moisture “breaks” a chiller selection program directly.

What breaks is the chain of assumptions that produces the number you type into that program. Load calculation software ships with temperate defaults. Coil selection carries those defaults forward. By the time an air-cooled chiller duty emerges at the far end, the moisture load has usually been understated, the required chilled water temperature has usually been set too high, and the resulting plant will be simultaneously oversized on paper and incapable of keeping a Malaysian building dry.

That combination — too big and still not doing the job — is the specific failure this article is about. It is extremely common, it is expensive, and it explains most of the “cold but clammy” buildings in the Klang Valley.

The number that changes everything

Here is the calculation that ought to be on the wall of every M&E office in Malaysia.

Take a Malaysian outdoor design condition of 33°C dry bulb at 27°C wet bulb. That air carries a humidity ratio of about 20.1 g of moisture per kg of dry air, an enthalpy of roughly 84.8 kJ/kg, and a dew point around 25°C.

Now take an indoor design condition of 24°C at 60% relative humidity, comfortably within the 23 to 26°C and 50 to 70% RH range that MS 1525 recommends for Malaysian buildings. That air holds about 11.2 g/kg, at an enthalpy of roughly 52.6 kJ/kg, with a dew point near 15.7°C.

Every kilogram per second of outdoor air you bring in must be dragged from the first state to the second:

• Total load: about 32 kW per kg/s
• Sensible portion: about 9 kW
• Latent portion: about 23 kW

Roughly 72% of your ventilation load in Malaysia is moisture, not temperature.

Now run the same calculation for a temperate design office — say 28°C dry bulb at 20°C wet bulb outside, 24°C at 50% RH inside:

• Total load: about 9.4 kW per kg/s
• Sensible portion: about 4 kW
• Latent portion: about 5.4 kW

The same ventilation rate imposes about three and a half times the coil load in Kuala Lumpur, and the moisture component alone is more than four times larger.

That single ratio is the source of nearly every sizing error that follows. Software written around the second set of numbers, with defaults calibrated to the second set of numbers, applied to the first set of conditions, will be wrong in ways that are not obvious until the building is occupied and someone complains about the smell of damp carpet.

There is a related observation worth sitting with. A GBI seminar note on MS 1525 made the point that average outdoor air temperature in Malaysia sits only about 4 degrees above the comfort range. On the sensible side, our climate is genuinely mild. It is the moisture that is brutal. Any design intuition built on dry-bulb temperature will systematically underestimate what Malaysian air conditioning actually has to do.

Where the temperate defaults hide

Four places, in the order they cause damage.

1. The design day is the wrong design day

This is the most technically interesting failure and the one fewest people check.

ASHRAE publishes climatic design conditions in separate sets: cooling conditions as dry bulb with mean coincident wet bulb, evaporation conditions as wet bulb with mean coincident dry bulb, and dehumidification conditions as dew point with humidity ratio and mean coincident dry bulb. They are different design points, and they occur at different times.

The critical detail, stated plainly in the engineering weather data documentation: the highest moisture values typically occur when dry-bulb temperatures are lower, peak humidity ratio is what you use for sizing dehumidification systems, and peak moisture conditions usually represent a higher total enthalpy than peak dry-bulb conditions.

Read that again, because it inverts the usual assumption. The worst hour for your cooling coil is not necessarily the hottest hour. In Malaysia it is frequently a wet afternoon: dry bulb drops to 28°C after a downpour, dew point sits stubbornly at 25 or 26°C, and total enthalpy is higher than it was at noon.

Most load software defaults to the cooling DB/MCWB design day. If nobody changes it, the dehumidification design condition never gets checked, and the coil is sized for an hour that is not the hard one.

2. Sensible heat ratio defaults

SHR is the fraction of total load that is sensible. Temperate office design typically lands somewhere around 0.85 to 0.95. Malaysian systems carrying realistic outdoor air quantities can sit at 0.65 to 0.75 at the coil.

If the software assumes a high SHR and reality is much lower, the selected coil has too few rows, too little contact time, and too high an apparent dew point. It will hit the supply air temperature target and miss the supply air moisture target completely. On the commissioning sheet everything looks fine, because commissioning sheets record temperature.

3. Outdoor air quantity, and the assumption underneath it

Ventilation rates in the region tend to run above ASHRAE 62.1 baselines — for retail spaces, for instance, MS 1525 practice sits at around 6 L/s per person against ASHRAE 62.1’s 5. Higher occupancy densities and sealed buildings drive this.

Every extra litre per second is 32 kW per kg/s of pure penalty, three-quarters of it moisture. Outdoor air that would be a modest line item in a temperate load calculation becomes one of the largest single loads in a Malaysian one.

Then there is infiltration, which imported software models badly for our conditions. Door openings in retail and F&B, loading bays, poorly sealed façades and lift lobbies all admit air at 25°C dew point. In a temperate climate that leakage costs you a little sensible load. Here it costs you moisture that the system was never sized to remove.

4. Chilled water temperature

To hold a room at 24°C and 60% RH, room dew point is about 15.7°C. Supply air must arrive below that dew point to remove moisture at all — realistically 12 to 14°C — which requires a coil apparent dew point around 10 to 12°C, which requires chilled water entering at roughly 6 to 7°C.

That is not much margin. And it collides directly with one of the most widely recommended energy-saving measures in the HVAC world: chilled water temperature reset. Raising chilled water supply temperature genuinely does improve chiller efficiency, and in a dry climate it is close to free money. In Malaysia, raise it two degrees and you can push the coil ADP above the room dew point, at which point dehumidification stops almost entirely.

The building gets more efficient at making a space humid. Occupants respond by lowering the thermostat, which puts the load straight back on — and then some.

What actually goes wrong on site

The failure sequence is consistent enough to be predictable.

The building runs humid. Space RH sits at 70% or above. Every relevant standard treats 70% as the ceiling precisely because mould growth sets in above it, and long exposure above that level creates a genuine indoor health hazard rather than just a comfort complaint.

Occupants compensate with temperature. Nobody can feel relative humidity directly; they feel clammy and reach for the thermostat. Setpoints drift down to 20°C, sometimes 18°C. This is the origin of the Malaysian office where people wear jackets indoors — a phenomenon the MS 1525 seminar material specifically calls out.

The plant now runs colder and harder for a worse result. Lower setpoint means more sensible cooling, more running hours, and higher energy cost, while the humidity problem remains unsolved because the coil still cannot get below the dew point it needs to.

Oversizing makes it worse, not better. This is the counterintuitive part. Engineers who suspect the load calculation is optimistic often add a safety margin. But an oversized chiller satisfies the sensible load quickly and then unloads or cycles. Short run times mean less time with air passing over a wet coil, which means less dehumidification. The oversized plant is worse at moisture removal than a correctly sized one. You have paid more to get a damper building.

Part load is where it really bites. At 50% sensible load on a rainy afternoon, the latent load has barely moved. SHR collapses. A plant sized and controlled around the design-day sensible peak has no strategy for that hour, and that hour happens hundreds of times a year in Malaysia.

And now it costs more than it used to. Since the RP4 restructure took effect on 1 July 2025, the Peninsular base tariff rose from 39.95 to 45.4 sen/kWh, and the old Maximum Demand charge was replaced by separate Capacity and Network charges. A plant running longer hours at a depressed setpoint to chase a humidity problem it cannot fix is now paying a meaningfully higher price for the privilege.

Designing around it

None of this is difficult. It requires refusing a handful of defaults.

Use Malaysian climatic data, and use the right design condition. Check what outdoor design conditions your software has loaded. Then check the dehumidification design condition — peak dew point and humidity ratio with its mean coincident dry bulb — separately from the cooling design condition, and confirm the coil works at both. MS 1525:2019 includes design wet-bulb temperature data for Malaysia; use it rather than whatever the software shipped with.

State the SHR you are designing to, and sanity-check it. If your load output shows a system SHR of 0.9 for a Malaysian building with real ventilation rates, something is wrong upstream. Find it before ordering equipment.

Calculate the required supply air dew point explicitly. Work backwards: target room condition, room dew point, required supply dew point, required coil ADP, required chilled water temperature. Write those numbers into the specification. If the chilled water temperature that falls out is lower than the plant can deliver, you have found your problem at design stage, which is the cheap place to find it.

Treat outdoor air separately where you can. A dedicated outdoor air system that dries ventilation air to a low dew point before it reaches the space removes the largest and most variable moisture load from the main coil entirely. It lets the terminal units handle sensible load, which is what they are good at, and it decouples humidity control from sensible load control — which is precisely the failure mode described above. For high-occupancy buildings in Malaysia this is usually the right answer, and it is still under-used here.

Specify turndown, not just capacity. Ask what the chiller can actually do at 25% load, and what happens below that. A plant that can only step in large increments will cycle through exactly the part-load conditions where our humidity problems live. This is where inverter-driven and multiple-compressor machines earn their premium in this climate, and where the argument for them is much stronger than the raw efficiency numbers suggest.

Be careful with chilled water reset. If you implement it, interlock it to space humidity, not just to outdoor temperature or load. Reset upward only when moisture conditions permit.

Resist blind safety margins. If you do not trust the load calculation, fix the load calculation. Adding 20% to a chiller because the numbers feel light produces a machine that dehumidifies worse and runs less efficiently. If you need resilience, get it from staging and redundancy — multiple smaller machines rather than one oversized one.

A note for process cooling

Everything above concerns comfort cooling, where humidity control is the point. Process chillers serving moulds, reactors, plating baths or batching plants face a different version of the same climate problem.

There, high ambient humidity mostly shows up as condensation: on chilled water pipework, on cold surfaces, on product. Insulation specifications drawn from temperate practice routinely prove inadequate here, because the condition that matters is the ambient dew point, and at 25°C dew point a great many surfaces that would stay dry in Europe will sweat continuously in Shah Alam. Vapour barrier integrity matters more than insulation thickness, and a poorly sealed joint will wet the insulation, destroy its performance and corrode the pipe underneath.

The sizing logic is different, but the underlying lesson is identical: the imported default assumes a drier world than the one your plant is standing in.

What to check on a system you already have

If you suspect an existing building has this problem, four measurements will tell you.

1. Measure space relative humidity, not just temperature, at several points and across a full day including after rain. If it sits above 70%, you have confirmation.
2. Measure supply air condition — dry bulb and RH, giving you supply dew point. If supply dew point is above room dew point, the system is not dehumidifying at all, whatever the temperatures say.
3. Check the actual thermostat setpoints against the design values. Widespread manual override toward 20°C is a symptom, not a user problem.
4. Log chiller loading. A machine spending most of its life below 40% load in a building that is uncomfortable is the signature of this failure.

The fix is not always a new chiller. Frequently it is coil replacement, chilled water temperature correction, a dedicated outdoor air unit, or control changes. Establish which before anyone quotes you for plant.

The short version

Malaysian air is only a few degrees hotter than comfortable. It is enormously wetter. Software, textbooks and design intuition developed in temperate climates are built around the opposite balance, and every default in the chain reflects that.

Check the outdoor design conditions your software loaded. Check the dehumidification design day, not just the cooling one. Check the SHR. Calculate the supply dew point you actually need and work backwards to a chilled water temperature. And when in doubt, size accurately rather than generously — because in this climate, a chiller that is too big is not a safe choice. It is a wet building with a bigger electricity bill.