Control-oriented modeling and optimization for the temperature and airflow management in an air-cooled data-center

This paper presents a method for minimizing the power consumption of a data-center cooling system by optimizing the airflow pattern and the supplied cold air temperature simultaneously. To discover the potential benefits of reorganizing the rack flow rates, a gray-box fast-temperature evaluation model is proposed for the first time, which reflects the thermal relationship among the components of the data center and has low computational complexity. Next, a model-based constrained nonlinear optimization problem is formulated with the aim of minimizing the power consumption of both cooling fans and air conditioners. Meanwhile, the safety thermal guidelines are considered as the main constraints. At last, the optimal settings of rack airflow rates and supplied cold air temperature are obtained by solving the optimization problems. The simulation results show that the proposed method can yield an effective thermal management tool and reduce significant cooling power for air-cooled data centers.

[1]  Mengxuan Song,et al.  Thermal-Aware Energy Management of an HPC Data Center via Two-Time-Scale Control , 2017, IEEE Transactions on Industrial Informatics.

[2]  T. Scherer.  Energy Efficient Thermal Management of Air-Cooled Data Centers , 2017 .

[3]  Van P. Carey,et al.  Exploration of a Potential-Flow-Based Compact Model of Air-Flow Transport in Data Centers , 2009 .

[4]  Qi Gong,et al.  QoS-Driven Power Management of Data Centers via Model Predictive Control , 2016, IEEE Transactions on Automation Science and Engineering.

[5]  Mengxuan Song,et al.  Numerical Study on the Optimized Control of CRACs in a Data Center Based on a Fast Temperature-Predicting Model , 2017 .

[6]  Ayan Banerjee,et al.  Integrating cooling awareness with thermal aware workload placement for HPC data centers , 2011, Sustain. Comput. Informatics Syst..

[7]  Yan Bai,et al.  Chip Temperature-Based Workload Allocation for Holistic Power Minimization in Air-Cooled Data Center , 2017 .

[8]  Cullen E. Bash,et al.  DIMENSIONLESS PARAMETERS FOR EVALUATION OF THERMAL DESIGN AND PERFORMANCE OF LARGE-SCALE DATA CENTERS , 2002 .

[9]  Mehmet Demirci,et al.  A Survey of Machine Learning Applications for Energy-Efficient Resource Management in Cloud Computing Environments , 2015, 2015 IEEE 14th International Conference on Machine Learning and Applications (ICMLA).

[10]  Xin Zhou,et al.  Efficient Compute-Intensive Job Allocation in Data Centers via Deep Reinforcement Learning , 2020, IEEE Transactions on Parallel and Distributed Systems.

[11]  Ayan Banerjee,et al.  Model-driven coordinated management of data centers , 2010, Comput. Networks.

[12]  Riccardo Lucchese,et al.  ColdSpot: A thermal supervisor aimed at server rooms implementing a raised plenum cooling setup , 2019, 2019 American Control Conference (ACC).

[13]  Xue Liu,et al.  Adaptive Power Management through Thermal Aware Workload Balancing in Internet Data Centers , 2015, IEEE Transactions on Parallel and Distributed Systems.

[14]  Bruno Sinopoli,et al.  Reducing data center energy consumption via coordinated cooling and load management , 2008, CLUSTER 2008.

[15]  Xing Zhang,et al.  Study of hot air recirculation and thermal management in data centers by using temperature rise distribution , 2016 .

[16]  Canbing Li,et al.  Optimizing energy consumption for data centers , 2016 .

[17]  Qinghui Tang,et al.  Sensor-Based Fast Thermal Evaluation Model For Energy Efficient High-Performance Datacenters , 2006, 2006 Fourth International Conference on Intelligent Sensing and Information Processing.

[18]  Douglas G. Down,et al.  ALTM: Adaptive learning-based thermal model for temperature predictions in data centers , 2019, 2019 IEEE Sustainability through ICT Summit (StICT).

[19]  Manish Marwah,et al.  Optimal Fan Speed Control for Thermal Management of Servers , 2009 .

[20]  Bruno Sinopoli,et al.  A Cyber–Physical Systems Approach to Data Center Modeling and Control for Energy Efficiency , 2012, Proceedings of the IEEE.