The challenge
Queensland Children's Hospital in South Brisbane is a working tertiary pediatric hospital. Its central energy plant supplies chilled water to the main hospital building and the Centre for Children's Health Research. The plant has one 1,100 kWr low-load variable-speed centrifugal chiller and five 3,315 kWr variable-speed centrifugal chillers, configured for N+1 redundancy.
In a hospital, cooling is not optional, so any change to the plant has to cut energy and carbon without affecting patient comfort or causing disruption. With five large identical chillers and one smaller chiller, how the plant stages and shares load across them decides how efficiently it runs.
The aim was to decarbonize the plant's operation without touching patient comfort or adding hardware.
Site profile
Verified results
Measured over 12 months after implementation
Measurement and verification followed IPMVP Option B over 12 months after the site team deployed the recommendations in March 2022. Emissions and operating cost figures are as reported in the Exergenics M&V summary for the site.
| Outcome | Result | Status |
|---|---|---|
| Efficiency improvement | 3.3% improvement in chilled water plant efficiency per year | Verified |
| Emissions reduced | 41 tCO2 per year | Verified |
| Operating cost saving | $28,000 per year | Verified |
| Patient comfort | No disruption reported during or after implementation | Observed |
| Pre-implementation study | 7.1% energy and 5.4% peak demand, simulated by the digital twin across a year of historical data | Modeled |
The modeled figure comes from the 2021 QUT technology evaluation for i-Hub, which simulated all three recommendations together before any change was made. The verified figures measure the plant as it ran after the site team's deployment.
The Exergenics approach
A digital twin, then recommendations for the site team
Exergenics doesn't write to the BMS. The recommendations went to the hospital's site team, who deployed them in the existing controls.
Data and documentation
A year of plant data came from the hospital's data warehousing platform: chiller energy, cooling load, temperatures and flow rates. The facilities management team flagged plant details that the data alone didn't show.
Build the digital twin
Machine learning models of each chiller, pump and cooling tower were combined into a model of the whole plant and checked against a randomized 20% holdout of the data. The model was assessed independently by QUT as part of i-Hub's Healthcare Living Laboratories program, funded by ARENA.
Develop the strategy
Exergenics adjusted the model's constraints with feedback from the site, then optimized the plant at every load and weather condition. The recommendations covered chiller staging, a dynamic condenser water temperature reset and load balancing between the small and large chillers.
Implement and verify
The site team deployed the recommendations in March 2022. Savings were measured to IPMVP Option B over the following 12 months.