Bad-Data Sequence Detection for Power System State Estimation via ICA-GAN

A deep learning approach to the detection of bad-data sequences in power systems is proposed. The bad-data model is nonparametric that includes arbitrary natural and adversarial data anomalies. No historical samples of data anomaly are assumed. The probability distribution of data in anomaly-free system operations is also non-parametric, unknown, but with historical training samples. A uniformity test is proposed based on a generative adversarial network (GAN) that extracts independent components of the measurement sequence via independent component analysis (ICA). Referred to as ICA-GAN, the developed approach to bad-data sequence detection can be applied at the individual sensor level or jointly at the system level. Numerical results demonstrate significant improvement over the state-of-the-art solutions for a variety of bad-data cases using PMU measurements from the EPFL smart grid testbed and that from the synthetic Northern Texas grid.

[1]  Two combinatorial tests of whether a sample has come from a given population. , 1950 .

[2]  Oliver Kosut,et al.  Vulnerability Analysis and Consequences of False Data Injection Attack on Power System State Estimation , 2015, IEEE Transactions on Power Systems.

[3]  Sean P. Meyn,et al.  Generalized Error Exponents for Small Sample Universal Hypothesis Testing , 2012, IEEE Transactions on Information Theory.

[4]  H. Salehfar,et al.  A neural network preestimation filter for bad-data detection and identification in power system state estimation , 1995 .

[5]  Aaron C. Courville,et al.  Improved Training of Wasserstein GANs , 2017, NIPS.

[6]  Bernhard Schölkopf,et al.  Support Vector Method for Novelty Detection , 1999, NIPS.

[7]  Aapo Hyvärinen,et al.  Nonlinear ICA Using Auxiliary Variables and Generalized Contrastive Learning , 2018, AISTATS.

[8]  Lang Tong,et al.  Impact of Data Quality on Real-Time Locational Marginal Price , 2012, IEEE Transactions on Power Systems.

[9]  Lang Tong,et al.  Universal Data Anomaly Detection via Inverse Generative Adversary Network , 2020, IEEE Signal Processing Letters.

[10]  Ying Chen,et al.  Evaluation of Reinforcement Learning-Based False Data Injection Attack to Automatic Voltage Control , 2019, IEEE Transactions on Smart Grid.

[11]  A. Abur,et al.  Autoregressive Filters for the Identification and Replacement of Bad Data in Power System State Estimation , 1987, IEEE Transactions on Power Systems.

[12]  Thomas J. Overbye,et al.  Grid Structural Characteristics as Validation Criteria for Synthetic Networks , 2017, IEEE Transactions on Power Systems.

[13]  Anurag K. Srivastava,et al.  Ensemble-Based Algorithm for Synchrophasor Data Anomaly Detection , 2019, IEEE Transactions on Smart Grid.

[14]  Léon Bottou,et al.  Wasserstein GAN , 2017, ArXiv.

[15]  Xiangyu Niu,et al.  Dynamic Detection of False Data Injection Attack in Smart Grid using Deep Learning , 2018, 2019 IEEE Power & Energy Society Innovative Smart Grid Technologies Conference (ISGT).

[16]  Lang Tong,et al.  Data Framing Attack on State Estimation , 2013, IEEE Journal on Selected Areas in Communications.

[17]  Mario Paolone,et al.  Real-time state estimation of the EPFL-campus medium-voltage grid by using PMUs , 2015, 2015 IEEE Power & Energy Society Innovative Smart Grid Technologies Conference (ISGT).

[18]  H. Vincent Poor,et al.  Machine Learning Methods for Attack Detection in the Smart Grid , 2015, IEEE Transactions on Neural Networks and Learning Systems.

[19]  Lalitha Sankar,et al.  Physical System Consequences of Unobservable State-and-Topology Cyber-Physical Attacks , 2016, IEEE Transactions on Smart Grid.

[20]  Two combinatorial tests of whether a sample has come from a given population. , 1950, Biometrika.

[21]  Lang Tong,et al.  Malicious Data Attacks on the Smart Grid , 2011, IEEE Transactions on Smart Grid.

[22]  Fred C. Schweppe,et al.  Bad Data Suppression in Power System Static State Estimation , 1971 .

[23]  H. Vincent Poor,et al.  Smarter security in the smart grid , 2012, 2012 IEEE Third International Conference on Smart Grid Communications (SmartGridComm).

[24]  Chong Li,et al.  Online Cyber-Attack Detection in Smart Grid: A Reinforcement Learning Approach , 2018, IEEE Transactions on Smart Grid.

[25]  A. G. Expósito,et al.  Power system state estimation : theory and implementation , 2004 .

[26]  Zhigang Chu,et al.  Can Attackers With Limited Information Exploit Historical Data to Mount Successful False Data Injection Attacks on Power Systems? , 2017, IEEE Transactions on Power Systems.

[27]  V. P. Chistyakov.  On the Calculation of the Power of the Test of Empty Boxes , 1964 .

[28]  Erkki Oja,et al.  Independent component analysis: algorithms and applications , 2000, Neural Networks.

[29]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[30]  Lang Tong,et al.  Subspace Methods for Data Attack on State Estimation: A Data Driven Approach , 2014, IEEE Transactions on Signal Processing.

[31]  Lang Tong,et al.  Dynamic attacks on power systems economic dispatch , 2014, 2014 48th Asilomar Conference on Signals, Systems and Computers.

[32]  Aapo Hyvärinen,et al.  Variational Autoencoders and Nonlinear ICA: A Unifying Framework , 2019, AISTATS.

[33]  Georg Langs,et al.  f‐AnoGAN: Fast unsupervised anomaly detection with generative adversarial networks , 2019, Medical Image Anal..

[34]  Liam Paninski,et al.  A Coincidence-Based Test for Uniformity Given Very Sparsely Sampled Discrete Data , 2008, IEEE Transactions on Information Theory.

[35]  A. Brameller,et al.  Power System Tracking State Estimation and Bad Data Processing , 1982, IEEE Transactions on Power Apparatus and Systems.

[36]  S. Brahma,et al.  Online Analytical Characterization of Outliers in Synchrophasor Measurements: A Singular Value Perturbation Viewpoint , 2018, IEEE Transactions on Power Systems.

[37]  Artur Czumaj,et al.  Testing monotone continuous distributions on high-dimensional real cubes , 2010, SODA '10.