Reducing Power Consumption in Data Center by Predicting Temperature Distribution and Air Conditioner Efficiency with Machine Learning

Authors:

Yuya Tarutani , Kazuyuki Hashimoto, Go Hasegawa , Yutaka Nakamura, Takumi Tamura, Kazuhiro Matsuda, and Morito Matsuoka

Conference:

IEEE International Conference on Cloud Engineering 2016

Abstract:

To reduce the power consumption in data centers, the coordinated control of the air conditioner and the servers is required. It takes tens of minutes for changes of operational parameters of air conditioners including outlet air temperature and volume to be reflected in the temperature distribution in the whole data center. So, the proactive control of the air conditioners is required according to the prediction temperature distribution corresponding to the load on the servers. In addition, since the power efficiency of the air conditioner is strongly dependent on the operational parameters, the air conditioner have to be operated at the optimum point in real time with reflecting the predicted return air temperature. In this paper, the temperature distribution and the power efficiency of air conditioner were predicted by using a machine-learning technique, and also we propose a method to follow-up proactive control of the air conditioner under the predicted optimum condition. The temperature distribution after load application were predicted by using the regression models, and the power efficiency of the air conditioner was derived by support vector regression model. The temperature distribution was predicted with an accuracy of 0.2◦C at 10 minute in the future. Eventually, the optimum point of the air conditioner reflecting the predicted temperature distribution was performed. Consequently, by the follow-up proactive control of the air conditioner and the load of servers, power consumption reduction of 30% at maximum was demonstrated.