real-time power state estimation

To address computational inefficiency and algorithmic coupling challenges in state estimation (SE) for large-scale power systems, this paper introduces JuliaGrid—a full-stack, open-source SE framework implemented in Julia. JuliaGrid unifies key SE functionalities—including observability analysis, weighted least squares (WLS), least absolute value (LAV) estimation, bad data detection, and phasor measurement unit (PMU) data integration—while deeply coupling Newton–Raphson power flow and interior-point optimal power flow solvers to enable closed-loop simulation. Evaluated on realistic test systems with 10,000–70,000 buses, JuliaGrid demonstrates superior convergence robustness and runtime performance compared to leading open-source SE tools. It achieves high accuracy, real-time capability, and cross-platform scalability without compromising numerical fidelity. By providing a high-performance, modular, and reusable SE infrastructure, JuliaGrid advances situational awareness for modern large-scale power grids.

This study addresses the gap between simulation-based research and real-world deployment in smart grid state estimation by presenting an experimental validation over a commercial 5G network. The authors develop a multi-node testbed integrating Raspberry Pi edge nodes with Typhoon hardware-in-the-loop (HIL) simulation, implementing end-to-end real-time state estimation and fault detection using an IEEE 4-bus feeder model, a phasor data concentrator (PDC), and key performance indicators (KPIs). Experimental results demonstrate that 5G achieves an average end-to-end latency approximately 6.5 times lower than LTE Cat-M, maintains high estimation accuracy under both steady-state and dynamic conditions, and enables fault detection with a latency as low as 0.80 seconds. This work provides the first empirical evidence of the real-time capability and reliability of smart grid state awareness in an operational 5G environment, effectively bridging the gap between simulation and practical implementation.

To address latency, scalability, and reliability bottlenecks arising from high-frequency phasor measurement unit (PMU) data in real-time power grid monitoring, this paper proposes a cloud-native intelligent PMU analytics framework. The framework introduces a novel, security-by-design architecture that natively integrates edge-cloud collaborative stream processing, elastic resource orchestration, and time-series machine learning. It leverages Apache Flink and Kafka for distributed stream processing, Kubernetes-based containerized microservices for deployment agility, and privacy-enhancing computation for data confidentiality. The system enables real-time anomaly detection and predictive analytics under dynamic grid conditions. Experimental evaluation demonstrates sub-second end-to-end latency, linear throughput scaling to over 10,000 PMU nodes, and significantly improved grid observability and fault prediction accuracy. This work establishes a scalable, highly reliable, and security-assured paradigm for real-time situational awareness in smart grids.

This work addresses the pressing need for highly accurate, scalable, distributed, and near real-time state estimation in complex power systems by proposing a factor graph–based vectorized Gaussian Belief Propagation (GBP) framework tailored for PMU-driven applications. The core innovations include multivariate GBP, which jointly models electrically coupled state variables and measurement relationships, and fused GBP, which simplifies the graph structure by integrating multiple measurements associated with the same group of variables. The approach enables fully distributed computation at the bus level and demonstrates rapid convergence—typically within a single iteration—while maintaining high accuracy and strong scalability, as validated on IEEE 1354- and 13,659-bus test systems.

Modern power grids with high inverter penetration face challenges including weak stability and delayed fault response. To address these, this paper proposes a physics-informed temporal foundation model framework—first introducing the Wiener–Kallianpur–Rosenblatt innovation process into power system modeling to intrinsically embed electromagnetic transient dynamics and sinusoidal waveform characteristics, thereby replacing generic large language models. Methodologically, the framework integrates high-resolution synchronized waveform measurements, physics-guided generative pretraining, causal time-series modeling, streaming data compression, and probabilistic anomaly detection. Evaluated on real-world grid data, it achieves a 23.6% improvement in fault detection accuracy and reduces average response latency by 41%. The approach significantly enhances real-time situational awareness, rapid fault localization, and robust protection capabilities. It establishes an interpretable, deployable intelligent monitoring paradigm tailored for inverter-dominated, grid-forming power systems.

本文评估了多种机器学习方法在电力系统中故障检测和线路识别的有效性，旨在解决传统保护系统面对复杂电网挑战时的不足。

This study addresses the challenges of heterogeneous, sparse data and cross-network generalization in power distribution grids by proposing Mycelium, a heterogeneous graph Transformer. Methodologically, the authors construct a unified grid ontology and a physics-based simulation pipeline to generate high-quality training data. The model introduces structure-aware communication edges and electrical reference feature encoding, integrated with task-specific temporal readout mechanisms, to enable physics-driven, cross-grid universal representation learning. Experimental results demonstrate that Mycelium surpasses specialized baselines on unseen benchmark networks, substantially improving multi-task inference performance and cross-domain generalization capabilities.

This study addresses the limitations of conventional computing reliant on specialized hardware by introducing, for the first time, a “power-grid-native computing” paradigm that leverages existing electrical grids as fixed physical computational operators. Methodologically, computation is executed through voltage perturbations and current responses, integrating time-domain simulation, Kirchhoff’s laws, power electronic interfaces, and surrogate models to construct a comprehensive encoding–decoding framework. This approach substantially reduces trainable parameters while supporting spatial concurrency and temporal multiplexing. Experimental evaluations demonstrate classification accuracies of 91.5% and 82.25% on MNIST and Fashion-MNIST, respectively. These results validate the critical influence of grid topology and signal representation on computational utility, establishing a promising foundation for physics-based analog computing.

本文提出了一种高效的联合仿真策略，用于智能电网的研究，通过改进的事件驱动同步方法和NS-3事件处理优化，显著提高了仿真效率。

为解决高压输电网络中故障检测与分类问题，提出了一种两阶段混合机器学习流水线，通过结合异常检测和监督分类提高准确性。