A Machine Learning Surrogate for Component Criticality Ranking in Interdependent Power-Communication Networks

2026-07-09 · Source: Machine Learning · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Power Systems Engineering · Depth: Advanced, quick

Summary

A machine learning surrogate model has been developed to rapidly rank component criticality in interdependent power-communication networks, addressing the computational burden of high-fidelity simulators for resilience planning. Using the Modified Implicative Interdependency Model (MIIM) as ground truth, a Gradient Boosting surrogate predicts contingency severity from leakage-free structural features. Tested on the IEEE 118-bus system, the model achieved Spearman correlations of 0.849 for per-contingency severity prediction and 0.853 for per-component criticality ranking. This performance matches MIIM's own reproducibility ceiling of approximately 0.85 and significantly outperforms topological centrality baselines (0.60-0.69 Spearman). The surrogate's effectiveness is primarily attributed to its use of inter-layer dependency information, supporting a two-stage workflow where the ML model quickly ranks components for hardening analysis, reserving MIIM for selective verification.

Key takeaway

For power systems engineers and resilience planners evaluating interdependent power-communication networks, you should integrate machine learning surrogates into your workflow. This Gradient Boosting model offers rapid, accurate component criticality ranking, achieving Spearman correlations of 0.853 on the IEEE 118-bus system. You can use this to quickly identify critical components for hardening, reserving computationally intensive simulations like MIIM for final verification. This approach significantly streamlines resilience planning and resource allocation.

Key insights

A Gradient Boosting surrogate accurately predicts power-communication network component criticality, matching high-fidelity simulator reproducibility.

Principles

• Inter-layer dependency information is crucial for criticality prediction.
• ML surrogates can achieve empirical ceiling performance of complex simulators.
• Computational burden of N-k contingency analysis can be mitigated.

Method

A Gradient Boosting model uses leakage-free structural features to predict contingency severity and derive component criticality, validated against MIIM on the IEEE 118-bus system.

In practice

• Use Gradient Boosting for rapid component criticality ranking.
• Prioritize hardening analysis based on surrogate predictions.
• Employ a two-stage workflow for resilience planning.

Topics

• Machine Learning
• Power Systems
• Cyber-Physical Systems
• Component Criticality
• Network Interdependencies
• Gradient Boosting

Best for: AI Scientist, Machine Learning Engineer, Research Scientist

Related on AIssential

• Sparse Inter-Layer Dependencies of Transformer FFN Neurons — Transformer FFN neurons exhibit sparse, structured inter-layer dependencies, discoverable via a training-free attribution method.
• Cyber-Physical Anomaly Detection in IoT-Enabled Smart Grids Using Machine Learning and Metaheuristic Feature Optimization — Optimized feature selection significantly enhances cyber-physical anomaly detection in smart grids, reducing data redundancy while improving accuracy.
• Surrogate-Assisted Framework for SI-Compliant Interconnect Design Optimization Using the Earth Mover's Distance — The framework optimizes SI-compliant PCB designs deterministically using neural surrogates, decision trees, and Earth Mover's Distance for efficient, explainable evaluation.
• MxGPS: Multiplex Graph Transformers for a Power Grid Foundation Model — Multi-task joint training prevents topology overfitting in power grid GNNs, enabling robust generalization across unseen grid structures.
• Neptuna: A Comprehensive Machine Learning Framework for Benchmarking Complex Multiphase Flows — Neptuna benchmarks shock-driven multiphase flows, showing composite losses and SoftAdapt improve ML surrogate fidelity, despite no single model excelling.
• Correcting Prompt Dependence in LLM Benchmarks: A Bayesian Hierarchical Model with Embedding-Space Clustering — Reliable LLM vulnerability evaluation requires a Bayesian hierarchical model with embedding-space clustering to quantify uncertainty and mitigate prompt bias.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine Learning.