Real-Time Stochastic Assessment of Dynamic N-1 Grid Contingencies
Ayrton Almada 1 2 , Laurent Pagnier 1 , Igal Goldshtein 3 , Saif R. Kazi 2 , and Michael (Misha) Chertkov 1 · 2025
Read Real-Time Stochastic Assessment of Dynamic N-1 Grid Contingencies by Ayrton Almada 1 2 , Laurent Pagnier 1 , Igal Goldshtein 3 , Saif R. Kazi 2 , and Michael (Misha) Chertkov 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-18007.txt.
Introduction
Ayrton Almada, with five co-authors, builds a live dashboard that judges power grid safety after sudden line faults. Standard checks look at steady states and miss the frequency swings that follow common faults. Their screen assumes a balanced grid, drops a random line for about a second, and flags cases where flows pass safe limits afterward.
A system-wide overload score sums the risk, and a linear-scaling method rates faults without brute-force simulation. Adaptive sampling focuses effort on the dangerous cases. Grid operators get a faster way to ask what-if questions while conditions change through the day.
Catalog record
| CREATOR | Ayrton Almada 1 2 , Laurent Pagnier 1 , Igal Goldshtein 3 , Saif R. Kazi 2 , and Michael (Misha) Chertkov 1 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-18007
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Real-Time Stochastic Assessment of Dynamic N-1 Grid Contingencies Ayrton Almada 1 2 , Laurent Pagnier 1 , Igal Goldshtein 3 , Saif R. Kazi 2 , and Michael (Misha) Chertkov 1 Abstract Power system operators need tools for rapid, real-time counterfactual assessments of grid security under fast-changing conditions. Traditional N-1 contingency analysis lacks dynamic evaluation, especially of frequency swings from common faults. This paper introduces a real-time dashboard framework to screen dynamic contingencies. It assumes: (a) the grid starts in a balanced state; (b) faults can occur randomly on any transmission line, temporarily de-energizing and then reconnecting it within about one second; and (c) contingencies are flagged if post-fault transients cause line flows to exceed safety thresholds. The key contributions are: (1) Overload Indicator – a system-wide metric quantifying integrated N-1 dynamic risk from a given state; (2) Scalable Fault Evaluation Algorithm – a linear-scaling method to assess dynamic fault impacts without brute-force simulations; and (3) Risk Estimation – a Cross Entropy Adaptive Importance Sampling method estimating the likelihood of low probability by high risk events, e.g. associated with potential transformer over-current. We demonstrate the framework on the Israeli transmission power grid (IG). \old@ps@headings \thanksto The work at UA was supported by (a) NSF DMS-2229012: "Collaborative Research: AMPS: Rare Events in Power Systems: Novel Mathematics, Statistics and Algorithms"; (b) subcontract from NOGA on "Uncertainty-Aware Toolbox for Simulation, Optimization, Control and Planning of Inter Connected Gas and Power Systems". AA was supported by LANL as a graduate research assistant. We are thankful to Prof. J. Blanchet (Stanford) and his team for multiple discussions and consultations on the application of the adaptive importance sampling (cross-entropy) method; and Dr. Bent (LANL) for multiple comments and suggestions. I Introduction I-A Motivation Power-system stability is the grid’s ability to regain an acceptable operating state after a disturbance without loss of system integrity. It ensures that perturbations – from routine load/generation fluctuations to severe faults – do not trigger unnecessary tripping or cascading outages. Stability margins depend on network topology, the pre-disturbance operating point, disturbance characteristics, and protection/control settings; sustained synchronous operation of electromechanical modes is central to this resilience [ 1 ] . Transient stability focuses on maintaining synchronism in the milliseconds-to-tens-of-seconds window immediately after a disturbance and is governed by the initial operating state, the disturbance severity, and the calibration/coordination of automatic protection. With growing system heterogeneity and uncertainty (e.g., variable renewables, demand response, advanced power electronics), real-time dynamic-security assessment is increasingly necessary, complementing – rather than replacing – traditional offline studies [ 2 ] . Recent continent-scale and U.S. incidents illustrate how fast voltage/frequency and protection interactions can tip the grid into low-probability, high-impact outcomes (Iberia, 28 April 2025; Texas, February 2021) [ 3 , 4 ] . In this manuscript we examine the possibility of creating such a new on-line tool – a dashboard to inform the system operator about counterfactual (possible) contingencies . Specifically, we are concerned about resolution of faults, e.g. single- and three- phase faults at any (of many) lines in the system, and thus estimate probability of a secondary effect – rare but potentially extremely damaging – caused by such a fault – with one of a critical elements of the system, e.g. a transformer, becoming overload during the primary post-fault seconds-long transient. I-B Main Contributions This project introduces • An efficient methodology for N-1 dynamic screening – analyze and simulate the dynamics of an ensemble of transients associated with counterfactual faults of any line for a given state. Fig. ( 1 ) shows an example of a undesirable transient behavior that results from triggering a fault in the system that lasts 0.5s . • A probabilistic framework to estimate how system overloads evolve over time based on angular phase behavior, supporting assessment of grid reliability under stochastic disturbances. We obtain a three–orders-of-magnitude (i.e., \sim 10^{3} ) efficiency gain over direct Monte Carlo of counterfactual contingencies via three cumulative improvements: 1. Analytic reduction: replace direct time–domain simulations with linear-algebraic solutions of the governing models; 2. Iterative reuse: approximate the analytic solution with fast iterations that reuse factorizations/pre-conditioners across many contingencies, thereby avoiding a full large matrix inversion per case; 3. Targeted sampling: employ cross-entropy–based importance sampling to concentrate trials on high-impact contingency candidates. Fig. 1: Single-Phase Fault Example (IG) – Unstable Case. If the faulty line isn’t re-energized quickly, loss of synchronization occurs (system splits). Fault, Clearance, Finish occurs at time: 0.0, 0.5 and 20s respectively. I-C Outline The rest of this paper is divided into the following sections III - VIII : • System and Fault Modeling [Section III ]: Develop a scalable, linearized second-order model of power grid dynamics based on swing equations. Model both single- and three-phase faults to simulate disturbances over three intervals: pre-fault, faulted, and post-clearing. • Overload Indicator [Section IV ]: Introduce a metric that captures fault severity using thermal limits, fault duration, and phase oscillations—providing a continuous assessment of system response. • Scalable Dynamic Evaluation [Section V ]: Apply analytics to re-solve the system’s dynamics efficiently. • Probabilistic Risk Estimation [Section VI ]: Estimate rare-event probabilities using Monte Carlo and advanced importance sampling method – efficient Cross-Entropy (validated on brute force MCMC) – through our custom simulator, N1Plus . • Case Study: Israeli Grid [Section VII ]: Demonstrate the framework’s accuracy and efficiency on the IG network (32 buses, 36 lines, 11 generators) under various fault scenarios. • Conclusion and Outlook [Section VIII ]: Extend the framework to more complex grids, e.g. utilizing AI tools to scale it up. We illustrate and summarize the constitution of this work on Fig. ( 2 ): Fig. 2: Conceptual Dashboard for Dynamic Contingency Screening. II Related Work and Background II-A N-1 Contingency Analysis The N-1 contingency analysis 1 1 1 Our methodology also generalizes to broader classes of contingencies, including those defined from historical data rather than strict N-1 rules. is a fundamental reliability tool in power system planning and operation. It ensures that the grid remains stable and within operational limits after the failure of any single critical component, such as a line, generator, or transformer. This principle underlies the N-1 security standard adopted by organizations such as NERC and ENTSO-E [ 5 ] . Traditionally, N-1 analysis simulates individual outages from a base-case power flow to identify violations (e.g., thermal or voltage limits) and flag critical contingencies for corrective actions such as re-dispatch or network reconfiguration [ 6 ] 2 2 2 Voltage regulation and dispatch corrections are not addressed in this manuscript. . While effective, these static methods are computationally demanding and often overlook dynamic effects, especially in large, renewable-rich systems with low inertia and high variability. To address these limitations, we introduce a framework for real-time screening of dynamic N-1 contingencies, focusing on short-term transient behavior (sub-second to tens of seconds; see Table I ) following typical disturbances such as single-phase short circuits. Our method estimates the probability of secondary risks – seconds-long but severe overloads on critical lines – that may trigger further instabilities in an already stressed grid. The ultimate goal is to provide operators with a clear, actionable dashboard that enhances situational awareness of rare but potentially severe post-fault events, such as line tripping induced by transient