skip to main content


This content will become publicly available on October 15, 2024

Title: Time-Synchronized State Estimation Using Graph Neural Networks in Presence of Topology Changes
Recently, there has been a major emphasis on developing data-driven approaches involving machine learning (ML) for high-speed static state estimation (SE) in power systems. The emphasis stems from the ability of ML to overcome difficulties associated with model-based approaches, such as handling of non-Gaussian measurement noise. However, topology changes pose a stiff challenge for performing ML-based SE because the training and test environments become different when such changes occur. This paper circumvents this challenge by formulating a graph neural network (GNN)-based time-synchronized state estimator that considers the physical connections of the power system during the training itself. The results obtained using the IEEE 118-bus system indicate that the GNN-based state estimator outperforms both the model-based linear state estimator and a data-driven deep neural network-based state estimator in the presence of non-Gaussian measurement noise and topology changes, respectively.  more » « less
Award ID(s):
2145063
NSF-PAR ID:
10486850
Author(s) / Creator(s):
; ; ;
Publisher / Repository:
IEEE
Date Published:
Page Range / eLocation ID:
1 to 6
Format(s):
Medium: X
Location:
Asheville, NC, USA
Sponsoring Org:
National Science Foundation
More Like this
  1. null (Ed.)
    Traditional state estimation (SE) methods that are based on nonlinear minimization of the sum of localized measurement error functionals are known to suffer from nonconvergence and large residual errors. In this paper we propose an equivalent circuit formulation (ECF)-based SE approach that inherently considers the complete network topology and associated physical constraints. We analyze the mathematical differences between the two approaches and show that our approach produces a linear state-estimator that is mathematically a quadratic programming (QP) problem with closed-form solution. Furthermore, this formulation imposes additional topology-based constraints that provably shrink the feasible region and promote convergence to a more physically meaningful solution. From a probabilistic viewpoint, we show that our method applies prior knowledge into the estimate, thus converging to a more physics-based estimate than the traditional observation-driven maximum likelihood estimator (MLE). Importantly, incorporation of the entire system topology and underlying physics, while being linear, makes ECF-based SE advantageous for large-scale systems. 
    more » « less
  2. State estimation (SE) is an important energy management system application for power system operations. Linear state estimation (LSE) is a variant of SE based on linear relationships between state variables and measurements. LSE estimates system state variables, including bus voltage magnitudes and angles in an electric power transmission network, using a network model derived from the topology processor and measurements. Phasor measurement units (PMUs) enable the implementation of LSE by providing synchronized high-speed measurements. However, as the size of the power system increases, the computational overhead of the state-of-the-art (SOTA) LSE grows exponentially, where the practical implementation of LSE is challenged. This paper presents a distributed linear state estimation (D-LSE) at the substation and area levels using a hierarchical transmission network topology processor (H-TNTP). The proposed substation-level and area-level D-LSE can efficiently and accurately estimate system state variables at the PMU rate, thus enhancing the estimation reliability and efficiency of modern power systems. Network-level LSE has been integrated with H-TNTP based on PMU measurements, thus enhancing the SOTA LSE and providing redundancy to substation-level and area-level D-LSE. The implementations of D-LSE and enhanced LSE have been investigated for two benchmark power systems, a modified two-area four-machine power system and the IEEE 68 bus power system, on a real-time digital simulator. The typical results indicate that the proposed multilevel D-LSE is efficient, resilient, and robust for topology changes, bad data, and noisy measurements compared to the SOTA LSE.

     
    more » « less
  3. State estimation (SE) is an important energy management system application for power system operations. Linear state estimation (LSE) is a variant of SE based on linear relationships between state variables and measurements. LSE estimates system state variables, including bus voltage magnitudes and angles in an electric power transmission network, using a network model derived from the topology processor and measurements. Phasor measurement units (PMUs) enable the implementation of LSE by providing synchronized high-speed measurements. However, as the size of the power system increases, the computational overhead of the state-of-the-art (SOTA) LSE grows exponentially, where the practical implementation of LSE is challenged. This paper presents a distributed linear state estimation (D-LSE) at the substation and area levels using a hierarchical transmission network topology processor (H-TNTP). The proposed substation-level and area-level D-LSE can efficiently and accurately estimate system state variables at the PMU rate, thus enhancing the estimation reliability and efficiency of modern power systems. Network-level LSE has been integrated with H-TNTP based on PMU measurements, thus enhancing the SOTA LSE and providing redundancy to substation-level and area-level D-LSE. The implementations of D-LSE and enhanced LSE have been investigated for two benchmark power systems, a modified two-area four-machine power system and the IEEE 68 bus power system, on a real-time digital simulator. The typical results indicate that the proposed multilevel D-LSE is efficient, resilient, and robust for topology changes, bad data, and noisy measurements compared to the SOTA LSE. 
    more » « less
  4. Parameter estimation with non-Gaussian stochastic fields is a common challenge in astrophysics and cosmology. In this paper, we advocate performing this task using the scattering transform, a statistical tool sharing ideas with convolutional neural networks (CNNs) but requiring neither training nor tuning. It generates a compact set of coefficients, which can be used as robust summary statistics for non-Gaussian information. It is especially suited for fields presenting localized structures and hierarchical clustering, such as the cosmological density field. To demonstrate its power, we apply this estimator to a cosmological parameter inference problem in the context of weak lensing. On simulated convergence maps with realistic noise, the scattering transform outperforms classic estimators and is on a par with the state-of-the-art CNN. It retains advantages of traditional statistical descriptors, has provable stability properties, allows to check for systematics, and importantly, the scattering coefficients are interpretable. It is a powerful and attractive estimator for observational cosmology and the study of physical fields in general. 
    more » « less
  5. The increasing uncertainty of distributed energy resources promotes the risks of transient events for power systems. To capture event dynamics, Phasor Measurement Unit (PMU) data is widely utilized due to its high resolutions. Notably, Machine Learning (ML) methods can process PMU data with feature learning techniques to identify events. However, existing ML-based methods face the following challenges due to salient characteristics from both the measurement and the label sides: (1) PMU streams have a large size with redundancy and correlations across temporal, spatial, and measurement type dimensions. Nevertheless, existing work cannot effectively uncover the structural correlations to remove redundancy and learn useful features. (2) The number of event labels is limited, but most models focus on learning with labeled data, suffering risks of non-robustness to different system conditions. To overcome the above issues, we propose an approach called Kernelized Tensor Decomposition and Classification with Semi-supervision (KTDC-Se). Firstly, we show that the key is to tensorize data storage, information filtering via decomposition, and discriminative feature learning via classification. This leads to an efficient exploration of structural correlations via high-dimensional tensors. Secondly, the proposed KTDC-Se can incorporate rich unlabeled data to seek decomposed tensors invariant to varying operational conditions. Thirdly, we make KTDC-Se a joint model of decomposition and classification so that there are no biased selections of the two steps. Finally, to boost the model accuracy, we add kernels for non-linear feature learning. We demonstrate the KTDC-Se superiority over the state-of-the-art methods for event identification using PMU data. 
    more » « less