Backend
Profile Screening and Frequency-Secure Operating Limit Derivation for Data Center Microgrids
arXiv cs.SY0 views
arXiv:2611.23504v1 Announce Type: new Abstract: Rapid power variations from artificial intelligence (AI) workloads pose frequency-security challenges for islanded data-center microgrids. This paper proposes a generator step- response-based framework to rapidly screen candidate load pro- files and derive frequency-secure operating limits for IT workload scheduling. Identified generator responses are superimposed to assess the cumulative frequency effects of successive load changes without repeated dynamic simulations. The framework identifies potentially unsafe intervals and determines allowable load magni- tudes, ramp rates, and minimum event spacing, while accounting for prediction uncertainty and governor limits. Validation on a 2.5 MVA islanded microgrid with eight generator configurations and 14 synthetic AI workload profiles shows that 94.75% of 40,208 assessment windows receive decisive screening labels, with no frequency-violating windows classified as secure.
Source:https://arxiv.org/abs/2611.23504PreviousAdaptive Line Switching for Cascade Mitigation in Power Systems via Graph Reinforcement LearningNextReinforcement Learning Optimized Flight Control for Interception of High-Altitude Maneuvering Targets
Related posts
Build a RAG Pipeline With Access Control: Ingestion, Retrieval and Evaluation
0dev.to Security
Apache Roller XML-RPC flaw allows unauthenticated code execution in older versions
0dev.to Security
Why Pydantic V2 is a Game Changer for Python Backends (Migration & Key Patterns)
0dev.to Python

How to Extract Contract Obligations With Source Citations
0dev.to Python
Your guardrail regex didn't protect your agent. It froze your entire Python event loop.
0dev.to Python

APIC Desktop v2: A Free Offline Photo Editor With Layers, RAW, Curves & a Healing Brush
0dev.to Python