The challenge
Every chilled water plant is different, and finding its best setpoints by hand takes engineering time that is in short supply. CSIRO's Synergy Building is an office and laboratory building in Acton, Canberra. Its plant has three air-cooled York screw chillers of 830 kW each, three primary chilled water pumps, three secondary chilled water loops and a thermal storage (buffer) tank.
Condenser water reset doesn't apply to an air-cooled plant, so the opportunity was in chiller staging. The project also tested whether Exergenics could onboard a building directly from CSIRO's Data Clearing House (DCH), a building data platform.
On an air-cooled plant, staging is the main lever. When one chiller can meet the whole load, there is little left to optimize.
Site profile
Verified results
Measured over the summer reporting period
Savings were measured about one month after implementation against the pre-optimization baseline. Figures are as published in the IEA EBC Annex 81 case study. At the published ACT grid intensity of 0.79 kgCO2e/kWh, the measured saving is about 1.6 tCO2e. At the published generic tariff of 15 c/kWh, it is about $307 over the reporting period.
| Outcome | Result | Status |
|---|---|---|
| Energy saving | 4.3% of plant energy over the summer reporting period | Verified |
| Energy avoided | 2,049 kWh over the summer reporting period | Verified |
| Carbon reduced | About 1.6 tCO2e over the reporting period | Verified |
| Chiller run time | Balanced across the three chillers. One chiller had been carrying a disproportionate share of the load. | Observed |
| Annual energy saving | 1.7%, simulated across a full year | Modeled |
| Peak demand saving | Not yet verified, because no peak demand event occurred during the M&V period | Pending |
The annual figure is lower than the summer result because, in cooler months, one chiller often meets the whole load and there is little scope to improve the chiller schedule.
The Exergenics approach
Setpoints for the incumbent BMS contractor
Exergenics doesn't write to the BMS. The optimized setpoints went to the building's incumbent BMS contractor in a simple format, and the contractor integrated them into the existing controller within days.
Connect to the Data Clearing House
Exergenics integrated with CSIRO's Data Clearing House by API to extract historical chilled water plant data. Any building already on the platform can now be onboarded the same way.
Model the plant
Chiller efficiency was modeled as a function of lift and part load using bi-quadratic regression, checked with k-fold cross-validation. Site knowledge set the constraints. Optimizing equipment loading at every cooling load and ambient condition gave the stage-up and stage-down demand setpoints.
Implementation by the incumbent contractor
The incumbent BMS contractor integrated the new staging setpoints into the existing controller. No hardware was added.
Measure and revisit
Savings were measured one month after implementation against the pre-optimization baseline. Measurement and verification will be revisited to confirm the long-term result.
What the project showed
Data in a building data platform needs interpretation before it's useful. Equipment metadata, such as location, hierarchy and nameplate data, wasn't stored with the trend data, so the team used BMS screenshots, drawings and site documents to work out what each point meant.
The data set came from a period of low occupancy, so it didn't fully represent normal operation. Storing occupancy, public holidays and normal operating hours alongside the trend data makes the next project faster.