Every mixed-use tower we design hits the same fork in the road: VRF or central chillers? The honest answer is "it depends" — but on more variables than most owners realise. Below is the comparison we walked a Colombo developer through last quarter for a 12-storey building.

The building: 12 floors, ~9,500 m² conditioned area, mixed retail (ground), serviced offices (floors 1–4), apartments (floors 5–12). Total cooling load: roughly 480 TR (1,690 kW) at peak.

Capex — chillers usually win, but not by as much as you'd think

For this building, the central chiller plant came in at LKR 235 million all-in (chillers, primary/secondary pumps, AHUs, FCUs, cooling towers, plant room civil work, BMS, controls, piping, insulation). The Hitachi VRF option came in at LKR 280 million for equivalent zoning.

That's a 19% capex premium for VRF — meaningful but smaller than it used to be. Three things have closed the gap: VRF outdoor unit costs have come down on per-TR basis, chiller plant rooms now require more BMS sophistication than they did 10 years ago, and the labour cost of installing a chilled-water system has risen faster than the labour cost of installing refrigerant piping.

Energy — VRF wins on partial load, chillers win on peak

This is where Sri Lankan climate data matters. Colombo sits at an effective full-load equivalent (EFLH) of around 2,800 hours/year for cooling — but only ~600 of those hours are at >80% peak load. The other 2,200 hours are at part load.

VRF inverter compressors track part-load demand efficiently — a Hitachi RAS-series VRF runs at IPLV ratings 30–40% better than its full-load EER. A central chiller without variable-speed drives on its compressors and pumps can lose 15–20% of its rated efficiency at part load. With VSDs across the chiller plant, the gap closes — but the VSD package adds about LKR 8 million to the chiller capex.

Net: for this building, modelled annual energy consumption was 1,180,000 kWh for VRF vs. 1,275,000 kWh for the VSD chiller plant — VRF roughly 7.5% lower on the year.

Maintenance — chillers are simpler at scale, VRF is simpler at first

Central chiller plants need fewer things serviced but each thing is bigger and more specialised. Annual service for the chiller plant: ~340 engineer-hours, mostly water treatment, BMS calibration, motor bearings, expansion tank checks. VRF: ~280 engineer-hours, but spread across many more outdoor units, each requiring routine refrigerant pressure checks and electronic expansion valve diagnostics.

Lifetime: chillers are typically rated for 20 years, VRF for 15. After 12–15 years, you're starting to look at outdoor-unit replacements on VRF, which is a non-trivial logistics exercise on a tall tower with limited crane access.

The surprise: cooling tower water

One factor most lifecycle calculations skip is municipal water cost. Colombo's water tariff for commercial buildings has roughly doubled in the last 8 years and continues to climb. A central chiller's cooling tower for a building this size makes up about 12 m³/day on a hot day — a non-trivial line item on the OPEX side. VRF, being air-cooled, has no equivalent cost.

Over the 15-year horizon, the water cost differential alone shifted total OPEX by about 4% in VRF's favour — enough to overturn what looked like a chiller-favourable lifecycle calculation in the original spec.

The recommendation

For this specific building, with this specific mix of tenants, VRF won on 15-year total cost of ownership by approximately 6%. But more importantly, it won on tenant flexibility — the ability to bill electricity by floor and to retrofit individual zones for new tenants without affecting the rest of the building.

If the building had been single-tenant — say, a corporate HQ with uniform operating hours — the chiller plant would have come out ahead. The mixed-use brief tipped it.

The takeaway

Don't take a generic answer on VRF vs. chiller. The right choice is a function of building geometry, tenant mix, climate hours and water tariffs — all of which are specific to your project. We'll model both options against your actual brief if you'd like a real comparison rather than a vendor pitch. Talk to an HVAC engineer.