Technical knowledge hub
Technical papers, research notes, reliability methodology, industry developments, commentary, and learning resources — organized so visitors can find depth without navigating a software catalog.
Common questions about GR services, software, training, secure AI, and how engagements begin.
A. Papers & research
GR technical papers, research notes and ongoing reliability work. Numerical results in illustrative studies are not field-validated and do not represent any specific hyperscaler.
Why N-1 capacity alone does not establish AI-compute availability: generation, fuel, cooling, protection, restoration, and qualified probabilistic interruption risk.
Engineering Research Note · October 2026Illustrative Framework and 100-MW Reference Study
A technical working paper exploring how probabilistic reliability methods can be extended to estimate usable electrical capacity in AI data centers while maintaining defined reliability targets.
Technical Working Paper — Illustrative StudyResearch note on the next phase of GR’s work: moving from an illustrative 100-MW reference framework toward topology-aware chronological reliability analysis.
Research in ProgressB. Individual publications
This section lists selected publications authored or co-authored by Sudhir Agarwal. These are the individual’s publications and are not necessarily GR corporate publications.
Selected publications are being added. Entries will include title, authors, publication or conference, year, a short description where available, and a PDF or external link when legally available. Publication details will not be invented.
C. Methodology
Methods, concepts and systems perspectives across industries — including intended function, operating conditions, availability, maintainability, resilience, probabilistic methods, outage-data interpretation and common-mode / dependent failures.
What GR means by reliability, repairable-system concepts, and a cross-industry mapping of intended function to example measures.
How data, models, methods, assumptions, interpretation and engineering judgment turn analysis into a defensible decision.
Framework under development Vendor-neutral methodology for assessing whether proposed capacity and flexibility are feasible, deliverable, evidence-supported and appropriately qualified for decisions.
Framework under development Lifecycle map from infrastructure need to dependable service, connecting capacity, InfraRel and the established SUBREL, DISREL and TRANSREL foundations.
Engineering assessment framework How large computational loads behave at the POI across operating states and time scales—and what models fit the claim and study.
How operating experience becomes reliability statistics and planning decisions.
Published examples of SUBREL and related planning applications — prior experience, not current endorsement.
D. External developments
Selected NERC, FERC, EPRI and related reliability developments. Inclusion is not affiliation or endorsement. Where GR provides engineering interpretation, that is labeled as GR commentary.
Source-verified selected developments affecting transmission planning, large loads, protection, operations and cybersecurity — with GR engineering interpretation clearly distinguished from primary sources.
Primary-source gateway to selected NERC, FERC, EPRI, research and large-load materials.
Independent GR discussion of EPRI’s operational AI-risk initiative for the electric sector. Not an EPRI endorsement of GR.
E. Commentary
GR articles, adapted perspectives and engineering commentary. This material is differentiated from formal technical papers and from primary-source industry documents.
Practical screening and governance perspectives for AI used in business and engineering workflows.
Commentary on productivity with AI-assisted workflows that remain human-reviewed and traceable.
F. Learning
Seminars, training programs, workshops and educational materials. Full training details live on the Training page.
Three-day power-system reliability seminar, workshops and hands-on instruction.
Industry-informed collaboration with university faculty and doctoral researchers.
