This study proposes a method for determining summer outdoor design parameters and design days using the current code. Based on the method from the current code in China, four typical cities are selected according to the adaptation zone of evaporative cooling technology, forming outdoor design parameters for evaporative cooling air conditioning with wet bulb temperature as the primary factor and dry bulb temperature as the secondary factor. Based on the historical long-term meteorological data, a new method for determining the design days is proposed. The characteristics and trends of outdoor climate change in typical cities in summer are extracted, and the data of the summer design days are obtained using the hourly coefficient method. The results show that the new method reduces the temperature difference between dry and wet bulbs and the evaporative cooling potential. Guiding the design according to the method in the current code results in an insufficient design. The design days in the current code cannot accurately reflect the variation trend of the outdoor climate in summer in different regions. However, the design days obtained by the hourly coefficient method proposed in this study are more consistent with the measured data of typical cities.

To address the heat dissipation challenges associated with high power density in intelligent computing centers, this study proposes an air-liquid dual source cooling system. The system utilizes a unified cold source with a stepwise water flow sequence of "air-cooling first, liquid-cooling second" to reduce exergy loss and match the temperature-grade of the respective cold source. The internal flow channel of the cooling unit was optimized through computational fluid dynamics simulations, and its performance was experimentally tested under high-temperature with high-humidity, moderate-temperature with moderate-humidity, and low-temperature conditions.Results indicate that the optimized unit achieves an outlet water temperature as low as 32.17 ℃ with a maximum coefficient of performance (COP) of 28.03 under high-temperature conditions. The unit also enables complete natural cooling under moderate conditions and reaches a peak COP of 33.4 under low-temperature conditions while maintaining stable operation. The system's overall approach degree ranges from 0.1 ℃ to 1.8 ℃, and its cooling capacity reaches 105.3%, outperforming the national standard. This study provides an efficient and practical cooling solution for high-density intelligent computing centers.