Cooling plant case
Hospital cooling plant
A mixed centrifugal and screw-chiller plant used API integration to optimize chiller leaving-water temperature, pump frequencies, and cooling-tower frequency.
- Energy reduction
- 32.18%
- Baseline daily energy
- 3,470.24 kWh
- AI daily energy
- 2,353.38 kWh
Calculated from comparable daily average energy.
Baseline period average, Oct 5-Oct 10.
AI full-day period average, Oct 12-Oct 19.
Plant profile
Four chillers with large and small pump trains
The plant used two centrifugal chillers and two screw chillers. Large pumps and towers served the centrifugal chillers; smaller pumps and towers served the screw chillers. Pumps and towers were variable-speed capable.
Equipment
Operating equipment
The equipment inventory included parallel large and small trains across chillers, chilled-water pumps, condenser-water pumps, and cooling towers.
| Equipment | Quantity | Specification | Status |
|---|---|---|---|
| Centrifugal chillers | 2 | 2461.9 kW cooling capacity / Nominal efficiency 0.56 kW/ton | Running |
| Screw chillers | 2 | 843 kW cooling capacity / Nominal efficiency 0.60 kW/ton | Running |
| Large chilled-water pumps | 3 | 75 kW / 36 m head / 466 m³/h flow | Running |
| Small chilled-water pumps | 3 | 30 kW / 34 m head / 160 m³/h flow | Running |
| Large condenser-water pumps | 3 | 45 kW / 22 m head / 556 m³/h flow | Running |
| Small condenser-water pumps | 3 | 18.5 kW / 22 m head / 190 m³/h flow | Running |
| Large cooling-tower groups | 2 | 7.5 kW x 3 / 615 m³/h flow | Running |
| Small cooling-tower groups | 2 | 7.5 kW / 205 m³/h flow | Running |
Control scope
- Integration
- AI optimization connected to the air-conditioning platform through API.
- Variables optimized
- Chiller leaving-water temperature, chilled-water pump frequency, condenser-water pump frequency, and cooling-tower frequency.
- Weather comparability
- Baseline average outdoor temperature 21.72 C; AI period average 22.07 C.
Calculation detail
- Baseline window
- Oct 5-Oct 10; average total energy 3,470.24 kWh/day; average outdoor temperature 21.72 C.
- AI window
- Oct 12-Oct 19; average total energy 2,353.38 kWh/day; average outdoor temperature 22.07 C.
- Transition day
- Oct 11 was excluded because AI was connected during the morning.
Formula
Energy reduction = (Baseline daily average energy - AI daily average energy) / Baseline daily average energy x 100%
Case result
(3,470.24 - 2,353.38) / 3,470.24 x 100% = 32.18%
Daily operating data
Daily operating data
The scatter plot uses the measured daily total energy and outdoor average temperature. The transition day is shown in the table but excluded from the comparison and fitted lines.
| Mode | Date | Chiller kWh | Chilled-water pump kWh | Condenser-water pump kWh | Cooling tower kWh | Total energy kWh | Outdoor average temp C |
|---|---|---|---|---|---|---|---|
| Baseline control | Oct 5 | 4422.43 | 667.23 | 409.36 | 393.00 | 5892 | 23.68 |
| Baseline control | Oct 6 | 3866.14 | 485.04 | 277.08 | 209.00 | 4837 | 21.20 |
| Baseline control | Oct 7 | 3354.48 | 486.27 | 286.75 | 224.24 | 4352 | 20.99 |
| Baseline control | Oct 8 | 3861.93 | 491.30 | 312.49 | 219.62 | 4885 | 21.28 |
| Baseline control | Oct 9 | 2742.01 | 447.16 | 238.87 | 214.30 | 3642 | 21.56 |
| Baseline control | Oct 10 | 2574.44 | 393.44 | 224.76 | 218.23 | 3411 | 21.59 |
| Transition day | Oct 11 | 1653.96 | 327.09 | 132.25 | 204.40 | 2318 | 21.42 |
| AI control | Oct 12 | 3080.84 | 504.61 | 186.13 | 213.33 | 3985 | 21.78 |
| AI control | Oct 13 | 2408.67 | 402.85 | 135.16 | 214.97 | 3162 | 23.21 |
| AI control | Oct 14 | 2959.52 | 415.65 | 169.42 | 207.22 | 3752 | 22.33 |
| AI control | Oct 15 | 2082.68 | 382.04 | 139.65 | 198.54 | 2803 | 22.35 |
| AI control | Oct 16 | 2730.15 | 460.97 | 181.67 | 210.36 | 3583 | 22.06 |
| AI control | Oct 17 | 1351.56 | 335.12 | 116.99 | 196.66 | 2000 | 21.54 |
| AI control | Oct 18 | 2151.34 | 340.96 | 155.24 | 220.79 | 2868 | 21.59 |
| AI control | Oct 19 | 2062.27 | 328.67 | 129.29 | 210.81 | 2731 | 21.66 |
Compare your plant against these operating patterns.
Bring plant topology, BMS trend logs, and utility data. We will map what can be measured before proposing an optimization scope.