Experimental Analysis of Different Measurement Techniques of Server-Rack Airflow Predictions Towards Proper DC Airflow Management

Data center cooling energy efficiency is critical to the successful operation of modern large data centers. In 2014, data centers in the U.S. consumed an estimated 70 billion kWh of electricity, representing about 1.8% of total U.S. electricity consumption. Given that the cooling infrastructure can average 40% of the total data center energy consumption, suggests the data center cooling energy consumed in 2014 can be approximated at 28 billion kWh. These numbers indicate that improving airflow management in order to improve the efficiency of cooling in data centers can significantly affect operating costs and allow for increased IT capacity, thereby extending the life of the data center. Some of the methods used to improve airflow include, but are not limited to, hot aisle and cold aisle containment, IT equipment alignment and configuration changes, bypass air management (e.g. cable penetrations), recirculation management (e.g. blanking panels). Other methods that can be deployed to improve cooling energy efficiency include air and/or waterside economization, variable frequency drives (VFD), and increased IT equipment inlet (supply air) temperatures, etc.Most of the above-mentioned thermal management technologies concentrate on managing airflow to achieve the desired server inlet temperature (supply air operating set point) and not to manage the amount of cool air (CFM) that each IT server should receive in order to remove the produced heat. However, airflow is equally important for quantifying adequate cooling to IT equipment, but it is more challenging to measure the airflow per server and hence per rack. Therefore, as a potential option for measuring this airflow an experimental based airflow measurement was performed in this study to quantify and compare between different devices including commercial flow hood, vane anemometer, and Mobile Temperature/Velocity Mesh (MTVM). Furthermore, the effect of measurement location (rack front/rear), type of IT equipment/rack, rack location and depth were investigated. On one hand, the results revealed that the rack airflow rate prediction using average inlet/outlet temperature across the rack was the most accurate and practical technique when compared to airflow reference value which was based on IT equipment pressure-flowrate curve. On the other hand, the measured flow rate using the flow hood at rack inlet face reported a 10% off from the reference value for rack C 1-8. Using flow hood for rack airflow is impractical to be used in real data centers. Therefore and based on the conducted comparison in this study, measuring air temperature across the rack inlet and outlet could be the easiest method to predict the actual rack airflow rate (i.e. supply at rack intake) and hence manage the airflow by compromising the supply to the IT equipment demand based on their flow curves.