power system modeling

Under energy transition, high renewable penetration and cross-border interconnections intensify grid uncertainty, rendering conventional power flow solvers inadequate for real-time operation, while purely data-driven models lack physical consistency. This paper proposes a physics-informed neural network (PINN) framework for power flow simulation, integrating Kirchhoff’s laws and other physical priors via hybrid modeling—combining MLP/GNN architectures with physics-constrained regularization and unsupervised loss. We introduce a novel four-dimensional evaluation framework (accuracy, physical consistency, industrial deployability, out-of-distribution generalization) and the LIPS benchmark platform. Systematic ablation studies demonstrate that explicit graph-structured modeling and direct optimization of physical equations are critical to reliability. The resulting model achieves high accuracy, strong physical consistency, robust out-of-distribution generalization, and practical industrial deployment potential. Code is fully open-sourced.

Existing power grid resilience research remains largely conceptual or focuses on isolated components, lacking a system-level, quantifiable definition and assessment framework. Method: Leveraging 15-minute-resolution customer outage time-series data and high-resolution meteorological records, we develop a spatiotemporal statistical model incorporating resilience sensitivity simulation and outage propagation dynamics inference. Contribution/Results: We propose the first system-level, empirically measurable definition of grid resilience. The model uncovers cumulative outage effects under extreme weather, inter-regional outage propagation mechanisms, and systemic response patterns. It identifies critical reinforcement nodes that reduce customer outage magnitude by nearly 50%. Validated across three major U.S. East Coast utility service territories, the model achieves high accuracy in forecasting outage progression—enabling actionable support for real-time dispatch decisions and emergency response.

Conventional power flow computation struggles with real-time performance under high renewable energy penetration. Method: This study proposes an AI-driven, physics-informed simulation framework. It introduces the novel LIPS evaluation framework—assessing machine learning performance, physical consistency, industrial applicability, and out-of-distribution generalization—and integrates graph neural networks, explicit physics-constrained embedding, uncertainty quantification, and lightweight deployment techniques. The approach is end-to-end validated on a realistic regional power grid model with 30% renewable penetration. Contribution/Results: Multiple configurations achieve 10–100× speedup while meeting engineering accuracy requirements, significantly enhancing timeliness in fault anticipation and stability assessment. Crucially, this work presents the first systematic validation of AI-based simulation reliability and robustness on a real-world power system. It establishes both theoretical foundations and practical implementation pathways for sustainable, trustworthy intelligent-grid simulation paradigms.

This study addresses the degradation of accuracy observed when physics-informed surrogate models—despite performing well in isolated evaluations—are integrated into dynamic power system simulators, particularly under high-stress operating conditions. Framing this challenge as a verification and validation (V&V) problem, the work proposes a theoretical framework that combines algebraic coupling sensitivity, dynamic error amplification mechanisms, and finite-time simulation horizons to bound prediction errors. The approach integrates model-based verification with data-driven conformal calibration. Using a physics-informed neural network as a surrogate for synchronous machine dynamics, the authors conduct residual analysis and generate conformal predictions within a differential-algebraic equation simulation framework. Their findings reveal that small residuals in governing equations do not necessarily guarantee small trajectory errors, thereby underscoring the critical need—and a new paradigm—for validating surrogate models in coupled dynamical environments.

Public power system datasets often lack fine-grained topology and branch power flow direction information, hindering node-level carbon emission assessment, demand response modeling, and renewable integration. To address this, we propose a constraint-optimization-based method for reconstructing power flow directions, integrating multi-source open data (system load, generator outputs, network connectivity), GIS geospatial information, and multimodal alignment techniques. This is the first approach to achieve high-accuracy branch-wise flow direction recovery without prior topological knowledge. Validated on the Alberta power grid, our method successfully reconstructs nodal generation mix, real-time power flow directions, and spatial carbon intensity distributions—enabling supply-demand interaction modeling, synthetic data generation, and clean energy policy support. The implementation is fully open-source and demonstrates strong cross-regional generalizability.

This study addresses the growing risk of climate-induced power outages by proposing a novel resilience assessment framework that integrates empirical analysis with cascading failure simulations in coupled power-communication networks. Leveraging large-scale empirical outage data from EAGLE-I and a multi-layer interdependent infrastructure model (MIIM), the work develops a data-driven approach on the IEEE 118-bus system to uncover climate-outage relationships and geographically heterogeneous cross-layer risk amplification. Results reveal that climate-related outages have increased by approximately 9,100 incidents annually on average, with coastal regions exhibiting significantly higher vulnerability. Under extreme weather conditions, system operational capacity drops sharply to 17.6%, demonstrating that interdependencies between infrastructure layers exacerbate regional resilience disparities and overcome the underestimation of coastal risks inherent in conventional aggregate statistical methods.

This work proposes an automated power system simulation framework based on multi-agent artificial intelligence and the Model Context Protocol (MCP) to address the lack of intelligent coordination and human–machine collaboration in traditional simulation approaches. By introducing the first pypowsybl-MCP interface, the framework enables large language models to invoke power system simulation tools through a standardized protocol, facilitating an interactive, auditable, and scalable multi-agent workflow under human supervision. The platform supports end-to-end automation of simulation configuration, execution, and analysis, integrating quantitative technical metrics with expert feedback for comprehensive evaluation. This approach significantly enhances the intelligence and collaborative efficiency of transmission system operators in power grid studies.

This study addresses the challenge of heterogeneity in devices, contracts, and operational data that hinders the development of digital twins for Renewable Energy Communities (RECs). To overcome this, the authors propose a model-driven systems engineering (MBSE) approach integrating SysML with the SAREF4ENER ontology. Building upon a validated REC domain model, they employ the open-source tool Modelio to develop SysML models for a representative residential subset, producing two module definition diagrams: one for device classification and another for community organization. This work pioneers the incorporation of the semantics-rich SAREF ontology for smart energy into the SysML modeling workflow, identifying four semantic gaps and bridging them through the integration of the SAREF4ENER reference package. The resulting framework offers a unified modeling foundation for REC digital twins, combining structural modeling capabilities with expressive semantic representation.

This work addresses the projected shortage of 1.5 million distribution engineering professionals by 2030 by proposing a natural language interaction framework based on the Model Context Protocol (MCP), which enables the first deep integration of large language models (LLMs) with OpenDSS. The framework supports a range of engineering tasks—including power flow analysis, voltage profiling, quasi-static time-series simulation, and automated optimization—through 36 domain-specific tools and three core optimization capabilities, such as capacitor placement and overvoltage mitigation. Designed for both cloud-edge协同 and offline deployment, the system reduces execution time for typical workflows like distributed energy resource interconnection screening from several hours to under two minutes, while maintaining full fidelity with native OpenDSS scripts. An interactive web platform provides real-time visualization and immediate user feedback.

This study addresses the computational inefficiency of branch-and-bound methods in solving large-scale security-constrained unit commitment problems, which arises from the vast number of decision variables. The authors propose a structure-aware dimensionality reduction framework that leverages large language models to identify and pre-fix a small subset of structurally stable on/off binary variables. This approach substantially reduces problem size while preserving the full modeling capabilities of mixed-integer linear programming solvers for handling network, ramping, reserve, and security constraints. Under the guarantee of solution feasibility and solver-certified optimality, the method significantly decreases both the number of branch-and-bound nodes and solution time across multiple IEEE benchmark systems and large-scale security-constrained scenarios, achieving up to an order-of-magnitude speedup on high-complexity instances with near-optimal objective values.