Co-Design Optimization for Data Center Cooling System via Digital Twin

Liquid-cooled exascale supercomputers dissipate heat through cooling plants organized as multiple parallel subloops, but how to allocate coolant distribution units across subloops and distribute flow among them has not been systematically addressed at this scale. This paper presents a three-layer optimization framework that jointly determines the integer partition of CDUs across subloops, continuous flow fraction allocation, and per-timestep co-design optimization of total flow rate and supply temperature subject to per-subloop thermal safety constraints. Using a Modelica digital twin of the Frontier exascale supercomputer calibrated against one year of operational data, all 611 feasible partitions of 25 CDUs are evaluated. The globally optimal two-subloop design achieves 35.48% annual cooling energy savings, only 0.18% above the current three-subloop Frontier design, and flow fraction optimization compensates for any feasible CDU assignment, reducing design sensitivity by 93%.