Los puntos clave no están disponibles para este artículo en este momento.
Abstract With the continuous development of digital economy, the number and scale of data centers are increasing year by year, and the energy consumption of data centers is also increasing. At present, most of the data center heat system operation regulation depends on artificial experience, to meet the needs of the safe operation of the server is considered, not from the perspective of energy saving, heat system of indoor side heat environment control, and the cold source system side is not according to the end demand and the demand difference between the end of the water supply temperature control and water flow optimization distribution, not only makes the natural cold source is not effectively use, and reduce the active cold source. For the multi-terminal data center with the centralized water cooling source system with natural cooling source, this study analyzed the heat transfer and thermodynamic process of the system, established the corresponding numerical model, and optimized the operation strategy of the system of centralized water-cooled air conditioning system in the data center. In view of the energy consumption waste of air conditioning system caused by supply and demand mismatch and excessive cooling in most data center current operation, this study proposes an personalized optimization strategy for centralized water-cooled air-conditioning systems in multi-terminal data centers based on supply-demand matching. Using this optimization method, a typical case was simulated and optimized with the outdoor environment of Beijing. Compared with the non-optimized working condition, the annual average chilled water supply temperature was raised from 12°C to 14.27°C, and the upper operating temperature of the natural cooling source was raised from 10°C to 15.897°C. Moreover, the NCM operation hours of the data center cooling source system were extended by 32.57% and the energy consumption of the cooling source system reduced by 28.70%.
Ge et al. (Mon,) studied this question.