大多数中小型数据中心使用的冷却单元由冷却单元返回时测得的温度控制，而不是服务器入口。这种间接且不协调的控制模式无法确保数据中心自身的高效冷却。现有的协调控制旨在根据当前条件控制冷却装置的供应温度。然而，由于回风温度控制的冷却装置无法明确控制其供应温度，因此此类控制的范围有限。此外，这种控制系统通常需要一个广泛的传感器网络来进行温度监测。我们在这里展示一个协调的、真实的- 时间监控和控制算法，确保配备回风温度控制冷却装置的数据中心高效冷却。该算法使用本文作者开发的影响指数度量概念来预测服务器入口温度，从而消除了广泛传感器网络的需要。该算法根据预测的服务器入口温度，通过其智能温度分析仪评估数据中心的热健康状况，并根据冷却装置的设定点，通过协调建议自动做出响应。优化技术旨在以渐进的方式将数据中心带到热安全、高效和稳定的状态。控制算法在一个状态下进行了测试- 最先进的热优化生产数据中心。观察到冷却装置的设定值从24°C（75°F）系统性地增加到高达28°C（82°F），从而使已经优化的数据中心的冷却能耗额外节省9%。据作者所知，这是第一个满足控制配备有回风温度控制冷却装置的数据中心要求的控制算法。引文:伊利诺伊州芝加哥ASHRAE交易——第121卷第1部分

The majority of the small- to medium-size data centers employ cooling units that are controlled by the temperatures measured at the returns of the cooling units instead of the server inlets. This indirect and uncoordinated mode of control fails to ensure the efficient cooling of the data center on its own. Existing coordinated controls aim at controlling the supply temperatures of cooling units in accordance with the prevailing conditions. However, such controls find limited scope due to the inability of the return-airtemperature- controlled cooling units to explicitly govern their supply temperature. Moreover, such control systems typically need an extensive network of sensors for temperature monitoring. We here present a coordinated, real-time monitoring and control algorithm that ensures efficient cooling for data centers equipped with return-airtemperature- controlled cooling units. The algorithm predicts the server inlet temperatures using the concept of influence index metrics developed by the authors of this paper, eliminating the need of an extensive sensor network.The algorithm evaluates the thermal health of the data center through its intelligent temperature analyzer, based on the predicted server inlet temperatures, and responds automatically through coordinated recommendations in terms of setpoints of cooling units. The optimization technique aims to take the data center to a thermally safe, efficient, and stable state in a gradual fashion. The control algorithm was tested in a state-of-the-art, thermally optimized production data center. A systematic increase in the setpoints of the cooling units from 24°C (75°F) to as high as 28°C (82°F) was observed, resulting in the saving of an additional 9% in the cooling energy consumption of an already optimized data center. To the best of the authors’ knowledge, this is the first control algorithm catering to the requirements of controlling a data center equipped with return-air-temperature-controlled cooling units.