Abstract
A large-scale chiller system operates with many components and their different combinations influence the system coefficient of performance (COP)—cooling capacity output divided by the total electric power input of components. This study aims to analyse how different component combinations and their transitions influenced the system COP. A total of 12,541 sets of operating conditions were collected at 15-min intervals from an existing system with five chillers of two different capacities. A total of 51 types of combinations of operating system components were identified. The development of general linear models ascertained that the two highest significant variables were the system part load ratio and component combination. The pair-up combinations with the same number of operating chillers, pumps and cooling towers accounted for 58.23% of the total time with a higher average system COP of 3.34. Results of network analysis indicated that the system COP tended to be higher when the transition was outside clusters with pair-up combinations. Survival analysis identified specific ranges for the temperature of return chilled water and the temperature of water leaving the condensers to prolong the pair-up combination in order to increase the system COP.
| Original language | English |
|---|---|
| Pages (from-to) | 699-710 |
| Number of pages | 12 |
| Journal | Applied Thermal Engineering |
| Volume | 154 |
| DOIs | |
| Publication status | Published - 25 May 2019 |
Keywords
- Coefficient of performance
- Cooling tower
- General linear model
- Network analysis
- Survival analysis
- Water-cooled chiller