Control of data center cooling using data driven heuristic approach
Baxendale, Mitchell David
MetadataShow full item record
This thesis proposes a method of heuristic temperature prediction with real-time data to control the cooling infrastructure of a data center. This thesis addresses thermal management in data centers through the control scheme in one case study, in which dynamic control of the air conditioner temperature set point with heuristic temperature prediction of the server rack inlet temperature results in reaching a target server rack temperature in ~30 min. This thesis addresses energy savings through another case study, in which dynamic controls utilize outside air economization, resulting in energy reduction of ~1-2 kW while reducing space temperature variation to under 2°C during normal operation.