냉각수온도 최적화 알고리즘을 적용한 냉동기 시스템의 운전 효율 실측 분석

초록

This study applies a deep neural network (DNN)-based condenser water temperature optimization algorithm to an actual central cooling system and analyzes its impact on chiller and system level performance. The proposed algorithm determines optimal condenser water temperature setpoints by minimizing predicted power consumption under given operating conditions, enabling adaptive control compared to conventional fixed setpoint operation. Field measurement data were used to compare baseline and algorithm-applied periods under identical cooling load conditions. The results show that system coefficient of performance (COP) improved noticeably during the algorithm-applied period, particularly within the cooling load range of 1200~2000 kW, where cooling operation was most frequent. Within this range, the system COP increased by an average of 10.82%, with the largest improvement observed in the 1200~1400 kW interval. Analysis of individual chiller operation indicates that load distribution and chiller staging significantly influence COP behavior under part-load conditions. These findings demonstrate that DNN-based condenser water temperature optimization can effectively enhance system efficiency under practical operating conditions and highlight the importance of integrated control strategies considering both setpoint optimization and chiller operation characteristics.

키워드