Chilled water and HVAC operational optimisation

Overview

Central cooling plant serving a large government facility, with water-cooled chillers, primary and secondary pumping, cooling towers, and an air side of AHUs and VAV terminals.

The review asked what the plant costs to run, and how much of that cost sits in operating strategy rather than equipment condition.

Why review operating strategy

Plant of this type is commonly commissioned once and then operated to the settings established at handover. Those settings can outlast changes in occupancy, load profile and equipment condition, because nothing in normal operation forces them to be revisited.

That makes control strategy worth reviewing periodically on its own merits, independently of whether any fault is suspected.

Engineering challenge

Cooling is continuous and the facility is occupied, so nothing could be trialled by taking plant offline. Changes had to be assessed against normal operation and implemented in a way that could be reversed.

Load across a facility of this size is not uniform. Where areas with different occupancy patterns share a distribution system, a plant-level change that suits the bulk of the building can leave a specific zone under-served. Any strategy change therefore has to be assessed against the worst-served terminal, not the average one.

Technical analysis

The review followed the water from the chiller to the coil and back.

Chilled water supply temperature. A supply temperature lower than the air side actually requires costs chiller efficiency directly. The review considered what supply temperature the air side needs to meet space conditions and dehumidification, and whether a reset strategy could be applied against it.

Differential pressure control. Secondary pumping held at a fixed differential pressure setpoint delivers more head than the system needs for much of its operating range. The review considered whether differential-pressure settings could be optimised against actual system demand, including sensor location and the scope for demand-following control.

Chiller staging. Running two machines each well below their efficient range consumes materially more energy than running one machine properly loaded, which makes part-load staging a standing optimisation question in any multi-chiller plant. Staging logic and thresholds were reviewed as an operational optimisation opportunity against the load profile.

Air-side operation. AHU scheduling, fresh-air proportion and coil valve behaviour were reviewed, along with VAV terminal operation. Terminals sitting at minimum or hunting distort the load the plant sees, and are a common reason a plant appears to be the problem when it is not.

Return temperature. Low return water temperature — low ΔT across the system — was used as a diagnostic route. It typically points to valve, coil or terminal behaviour rather than to the plant, and it can make a plant appear overloaded when it is not.

Findings and recommendations

The work identified operational optimisation opportunities rather than capital proposals: chilled water supply temperature reset against actual air-side requirement; differential-pressure setpoint review including sensor placement; chiller staging thresholds assessed against the real load profile; AHU scheduling matched to occupancy; and investigation of terminals contributing to low system ΔT.

Each was framed so it could be trialled, monitored and reversed.

Results

Performance improvement opportunities were identified. Quantifying them honestly requires before-and-after measurement at comparable load and weather, which had not been completed. No saving percentage is claimed.

Lessons learned

On plant of this type the available improvement usually sits in operating strategy rather than equipment — in assumptions that have outlived the conditions they were set for, rather than in anything that has failed.

Low system ΔT is worth treating as the first diagnostic on any chilled water complaint. It sends the investigation to the air side, where the cause usually is, instead of to the chillers, where it usually is not.

And a strategy change that cannot be measured cannot be defended. The instrumentation to prove a saving is part of the scope of proposing one, not a follow-up task.