Cyber-physical Design of Data Centers Cooling Systems Automation

Modern data centers in many aspects are akin to industrial plants that generate a lot of heat by consuming substantial amount of energy and require powerful cooling and ventilation. Cooling system contributes with 30 to 50% of the total energy consumption of data centers. An effective way to address energy efficiency in such cooling systems is to apply advanced automation solutions, similar to that of industrial and building automation systems. However, existing automation solutions are not flexible enough to meet requirements of cooling systems in modern data centres. This paper is an endeavour to utilize distributed adaptive automation architecture in order to improve energy efficiency of cooling. The proposed automation algorithms are validated in a simulation environment which models the thermal behaviour of a server room and helps to find the most energy efficient control strategy for controlling the cooling devices. This paper describes the simulation tool comprising of thermal behaviour modelling in MATLAB/SIMULINK connected in closed-loop with the distributed control environment of IEC 61499 standard. Simulation of a typical server room under certain constraints using the proposed tool is described and the results are presented. The results demonstrate the potential of improving higher energy efficiency, flexibility and better decision-making ability for controlling the cooling systems.