BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260202T201248Z
LOCATION:Second Floor Atrium
DTSTART;TZID=America/Chicago:20251120T080000
DTEND;TZID=America/Chicago:20251120T170000
UID:submissions.supercomputing.org_SC25_sess533_post288@linklings.com
SUMMARY:Distributed Modular Digital Twin Network for High-Performance and 
 Reliable Data Centers
DESCRIPTION:Yan Chen, Xing Lu, Cary Faulkner, Alex Vlachokostas, Hanlong W
 an, and Jeremy Lerond (Pacific Northwest National Laboratory (PNNL))\n\nHi
 gh performance computing (HPC) workloads are driving rack power densities 
 beyond 100 kW, creating unprecedented stress on data center cooling and po
 wer systems. Conventional CFD-based digital twins provide high-fidelity de
 sign optimization but are too computationally intensive and rigid for oper
 ational use. We present the first physics-constrained Distributed Modular 
 Digital Twin Network (DMDTN), designed for real-time performance evaluatio
 n, load prediction, and fault detection. Each subsystem (e.g., cooling, po
 wer, IT load) is represented by an AI-driven surrogate model, interconnect
 ed through conservation laws and coordinated via a distributed message bus
 . This modular design preserves physical consistency while enabling scalab
 ility and rapid adaptability. Using synthetic datasets, DMDTN achieved ~60
 % lower prediction error (RMSE 172 vs. 450) and more than 2× faster traini
 ng (201 vs. 442 seconds) than a monolithic model, while maintaining robust
 ness under stress. DMDTN complements CFD by enabling accurate, real-time o
 perational management of HPC data centers.\n\nTag: Research & ACM SRC Post
 ers\n\nRegistration Category: Technical Program Reg Pass\n\nSession Chairs
 : Kento Sato (RIKEN Center for Computational Science (R-CCS)); Chris Schli
 palius (Pawsey Supercomputing Research Centre; Commonwealth Scientific and
  Industrial Research Organisation (CSIRO), Australia); and Anja Gerbes (Ge
 org-August-Universität Göttingen)\n\n
END:VEVENT
END:VCALENDAR