Article
Engineering, Electrical & Electronic
Debottam Mukherjee
Summary: This letter proposes a novel attack vector formulation policy for the linear state estimation algorithm by exploring the low-rank subspace of the mapping matrix. An extensive analysis on the IEEE 14 bus test bench supports the aforementioned propositions.
IEEE TRANSACTIONS ON SMART GRID
(2022)
Article
Engineering, Electrical & Electronic
Yanli Liu, Junyi Wang, Peng Wang
Summary: This paper proposes a hybrid data-driven method for joint estimation of distribution network topology and line parameters based on small data sets. The method utilizes the sparse characteristics of distribution network to determine the initial topology using a neighbor node selecting mechanism and estimate the initial values of line parameters and phase angle using linear regression. The initial topology, line parameters, and phase angle are then iteratively optimized using the Newtonian method with a decoupled linear power flow model. The test results show that the proposed method accurately and stably estimates the topology and parameters with only small data sets, and is adaptable to different network scales and types.
INTERNATIONAL JOURNAL OF ELECTRICAL POWER & ENERGY SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Di Cao, Junbo Zhao, Weihao Hu, Qishu Liao, Qi Huang, Zhe Chen
Summary: This paper investigates the distribution system state estimation (DSSE) with unknown topology change. A specific kernel that can transfer across different topologies is used to find relevant patterns and induce knowledge transfer. Bayesian inference is employed to quantify the uncertainties of the DSSE results. Comparative results with other methods demonstrate the improved accuracy and robustness against topology change.
IEEE TRANSACTIONS ON SMART GRID
(2023)
Article
Mathematics, Applied
Difeng Cai, Edmond Chow, Yuanzhe Xi
Summary: This paper investigates how to approximate a kernel matrix using low-rank approximation when the point sets X and Y are large and arbitrarily distributed. Such matrices often arise in Gaussian process regression with high-dimensional data. To linearly or nearly linearly handle large data, this paper proposes a method of geometrically selecting subsets of points to construct a low-rank approximation, with an analysis on how the selection should be performed.
NUMERICAL LINEAR ALGEBRA WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Qin Jiang, Xi-Le Zhao, Jie Lin, Ya-Ru Fan, Jiangtao Peng, Guo-Cheng Wu
Summary: Recently, low-rank matrix/tensor-based methods have gained attention for recovering multi-dimensional multimedia data. However, the assumption of low rank is often violated due to diverse local similarity. In this article, a size-adaptive super-tensor is proposed to flexibly exploit local similarity at different scales. The proposed method outperforms other competing methods in recovering multimedia data.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Interdisciplinary Applications
Ahmet Sahin, Ismail Sevim, Erinc Albey, Mehmet Guray Guler
Summary: This paper proposes a data-driven matching algorithm for ride pooling, which improves the accuracy of predicting passengers' choices by learning feature weights and candidate rankings. Experimental results show a significant improvement in prediction accuracy compared to the existing methods.
COMPUTERS & OPERATIONS RESEARCH
(2022)
Article
Engineering, Electrical & Electronic
Yazhou Jiang, Thomas H. Ortmeyer, Miao Fan, Xiaomeng Ai
Summary: This paper proposes a computational approach based on low-rank tensor approximation (LRA) to evaluate the impact of commercial fast EV charging stations on electric power distribution systems. By implementing a stochastic transportation model to predict the charging demand and constructing polynomial bases using the predicted results, the proposed approach optimizes the determination of coefficients and configuration of the LRA model, significantly reducing the simulation time of distribution feeder circuits while maintaining high accuracy.
IEEE TRANSACTIONS ON SMART GRID
(2023)
Article
Geochemistry & Geophysics
Peng Lin, Suping Peng, Yang Xiang, Chuangjian Li, Xiaoqin Cui, Wenkai Zhang
Summary: Random noise attenuation is crucial in seismic data processing for accurate structural imaging and data inversion. A novel low-rank approximation method using CUR matrix decomposition is proposed to address this issue. The CUR decomposition decomposes a matrix into three matrices, C, U, and R, to obtain a low-rank approximation.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Mathematics, Applied
Jeffry Chavarria-Molina, Juan Jose Fallas-Monge, Pablo Soto-Quiros
Summary: This paper introduces a fast-GLRMA method for computing generalized low-rank matrix approximation, which utilizes tensor product and Tikhonov's regularization to improve computational efficiency while maintaining accuracy.
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS
(2022)
Article
Geochemistry & Geophysics
Chong Wang, Zhiyuan Gu, Zhihui Zhu
Summary: This article discusses the importance of seismic data reconstruction and denoising in seismic data processing algorithms. It proposes an iterative algorithm based on the Hankel low-rank reconstruction problem, utilizing the Hankel and low-rank structures of clear and complete seismic data to improve reconstruction and denoising performance.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Mathematics, Interdisciplinary Applications
Vahid Keshavarzzadeh, Robert M. Kirby, Akil Narayan
Summary: The study introduces a low rank approximation approach for topology optimization of parametrized linear elastic structures. By using an Artificial Neural Network to map low resolution design iterates, the computational cost of high resolution topology optimization is significantly reduced.
COMPUTATIONAL MECHANICS
(2021)
Article
Engineering, Electrical & Electronic
Adnan Anwar, Abdun Naser Mahmood, Zahir Tari, Akhtar Kalam
Summary: This study highlights the importance of smart grid cyber-security and the construction of sparse False Data Injection attacks using data-driven methods. By revealing grid topology through measurement signals and utilizing the ADMM method to solve the complex problem, the accuracy of revealing grid topology is evaluated using graph-theoretic measures. The research findings demonstrate that manipulating a few sensor devices can construct sparse attacks, significantly impacting operational performance.
ELECTRIC POWER SYSTEMS RESEARCH
(2022)
Article
Energy & Fuels
Nan-Ying Lan, Fan-Chang Zhang, Xing-Yao Yin
Summary: Seismic data reconstruction is a crucial step in improving the quality of seismic data processing. This paper proposes a novel regularization method based on a low dimensional manifold model, which achieves excellent reconstruction results by enforcing low dimensionality during the reconstruction process.
Article
Energy & Fuels
Enrico Dalla Maria, Mattia Dallapiccola, Davide Aloisio, Giovanni Brunaccini, Francesco Sergi, David Moser, Grazia Barchi
Summary: This paper presents a novel data-driven parameter estimation procedure for high-current battery modelling, which has been applied to experimental measurement data conducted on Lithium Titanate batteries. The proposed methodology encompasses the entire process of parameter estimation, starting from raw experimental data. The validation results confirmed that the proposed approach had good modelling performance.
JOURNAL OF ENERGY STORAGE
(2023)
Article
Mathematics, Applied
Dominik Mokris, Bert Juettler
Summary: This paper focuses on finding bivariate tensor-product spline functions that approximate gridded data using least-squares approximation. The proposed method involves low rank approximation of matrices and solving a sequence of univariate fitting problems efficiently. It is an extension of the method proposed by Georgieva and Hofreither (2017) that combines cross approximation with spline interpolation. The algorithm provides the best least-squares approximation after r steps, but can also terminate earlier to obtain a low-rank approximation with sufficient accuracy. A stopping criterion based on a lower error estimate is also presented for efficient usage when the required number of degrees of freedom is unknown in advance.
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS
(2024)