The smart electricity grids have been evolving to a more complex cyber-physical ecosystem of infrastructures with integrated communication networks, new carbon-free sources of power generation, advanced monitoring and control systems, and a myriad of emerging modern physical hardware technologies. With the unprecedented complexity and heterogeneity in dynamic smart grid networks comes additional vulnerability to emerging threats such as cyber attacks. Rapid development and deployment of advanced network monitoring and communication systems on one hand, and the growing interdependence of the electric power grids to a multitude of lifeline critical infrastructures on the other, calls for holistic defense strategies to safeguard the power grids against cyber adversaries. In order to improve the resilience of the power grid against adversarial attacks and cyber intrusions, advancements should be sought on detection techniques, protection plans, and mitigation practices in all electricity generation, transmission, and distribution sectors. This survey discusses such major directions and recent advancements from a lens of different detection techniques, equipment protection plans, and mitigation strategies to enhance the energy delivery infrastructure resilience and operational endurance against cyber attacks. This undertaking is essential since even modest improvements in resilience of the power grid against cyber threats could lead to sizeable monetary savings and an enriched overall social welfare.

/pdf/electric-power-grid-resilience-to-cyber-adversaries-state-of-190drb8ihn.pdf

Electric Power Grid Resilience to Cyber Adversaries: State of the Art

Mobile power sources (MPSs), including mobile emergency generators, truck-mounted mobile energy storage systems, and electric vehicles, have great potentials to be employed as grid-support resources during power grid emergency operating conditions to supply the critical loads and enhance the resilience of distribution system (DS) via a swift disaster restoration. We here investigate the MPS dispatch (i.e., routing and scheduling) in coordination with DS dynamic network reconfiguration. We propose a two-stage restoration scheme to facilitate the DS restoration following the high-impact low-probability (HILP) seismic disasters. In the first stage, a seismic hazard is simulated through a Monte Carlo simulation engine to estimate the unavailability of power distribution branches under a suite of seismic force scenarios. In the second stage, a mixed-integer nonlinear programming (MINLP) optimization model is formulated for DS restoration that cooptimizes the routing and scheduling of MPSs and DS dynamic network reconfiguration. The MINLP model is then linearized to a mixed-integer linear programming model to reduce the computation complexity, where the seismic-resilience recovery strategies are generated at different timescales. The efficacy of the proposed method is evaluated on the IEEE 33-node test system and the results verify a significant reduction in the load outages and an improved power system resilience to HILP earthquakes.

Seismic-Resilient Electric Power Distribution Systems: Harnessing the Mobility of Power Sources

Classifying breast cancer tissue by Raman spectroscopy with one-dimensional convolutional neural network.

The recent advances in computing technologies and the increasing availability of large amounts of data in smart grids and smart cities are generating new research opportunities in the application of Machine Learning (ML) for improving the observability and efficiency of modern power grids. However, as the number and diversity of ML techniques increase, questions arise about their performance and applicability, and on the most suitable ML method depending on the specific application. Trying to answer these questions, this manuscript presents a systematic review of the state-of-the-art studies implementing ML techniques in the context of power systems, with a specific focus on the analysis of power flows, power quality, photovoltaic systems, intelligent transportation, and load forecasting. The survey investigates, for each of the selected topics, the most recent and promising ML techniques proposed by the literature, by highlighting their main characteristics and relevant results. The review revealed that, when compared to traditional approaches, ML algorithms can handle massive quantities of data with high dimensionality, by allowing the identification of hidden characteristics of (even) complex systems. In particular, even though very different techniques can be used for each application, hybrid models generally show better performances when compared to single ML-based models.

https://www.mdpi.com/1996-1073/14/16/4776/pdf

Advances in the Application of Machine Learning Techniques for Power System Analytics: A Survey

Power grid operation continuously undergoes state transitions caused by internal and external uncertainties, e.g., equipment failures and weather-driven faults, among others. This prompts an observation of different types of waveforms at the measurement points (substations) in power systems. Modern power systems utilize phasor measurement units (PMUs) and intelligent electronic devices embedded with PMU functionality to capture the corresponding peculiarities through synchrophasor measurements. However, existing PMU devices are equipped with only one synchrophasor estimation algorithm (SEA) and are, thus, not always robust to handle different types of signals across the network. This article proposes a PMU-embedded framework that ensures real-time grid surveillance and potentially enables adaptive selection of preinstalled SEAs in the PMU. Therefore, it ensures high-fidelity measurements at all times and irrespective of the input signals. Our proposed framework consists of: 1) a pseudocontinuous quadrature wavelet transform which generates the featured scalograms and 2) a convolutional neural network for event classification based on the extracted features in the scalograms. Our experiments demonstrate that the proposed framework achieves high classification accuracy on multiple types of prevailing events in power grids, through which an enhanced grid-scale situational awareness in real time can be realized.

Power Grid Online Surveillance Through PMU-Embedded Convolutional Neural Networks

One consensus in the machine learning community is that obtaining good representations of the data is crucial for the classification tasks. But establishing a clear objective for representation learning is an open question and difficult. In this paper, we propose the Discriminant Analysis Loss Function (DALF) for the representation learning in Deep Neural Networks (DNNs). The gradients of DALF explicitly minimize the within-class variances (scatter) and maximize the between-class variances. We use DALF to drive the training of DNNs and call them Discriminant Analysis Deep Neural Networks (DisAnDNNs). Compared to other Linear Discriminant Analysis (LDA)-based cost functions, the computational cost of DALF is drastically reduced by avoiding eigen-decomposition and matrix inversion. We used simple datasets to illustrate the geometric meaning of DALF and compared it with LDA, then experimented with DALF-driven Residual Learning Nets (ResNets) on the pediatric pneumonia (chest X-ray image) dataset. The experimental results show that the DisAnDNNs achieve state-of the-art accuracy in the binary classification task. Particularly, in the pediatric pneumonia dataset, we achieved the accuracy of 96.63%, with a sensitivity of 99.23% and a specificity of 92.30%, all of which are better than the results in the literature that published the dataset.

Discriminant Analysis Deep Neural Networks

A loss function has two crucial roles in training a conventional discriminant deep neural network (DNN): (i) it measures the goodness of classification and (ii) generates the gradients that drive the training of the network. In this paper, we approximate the gradients of cross-entropy loss which is the most often used loss function in the classification DNNs. The proposed approximations are noise-free, which means they depend only on the labels of the training set. They have a fixed length to avoid the vanishing gradient problem of the cross-entropy loss. By skipping the forward pass, the computational complexities of the proposed approximations are reduced to O(n) where n is the batch size. Two claims are established based on the experiments of training DNNs using the proposed approximations: (i) It is possible to train a discriminant network without explicitly defining a loss function. (ii) The success of training does not imply the convergence of network parameters to fixed values. The experiments show that the proposed gradient approximations achieve comparable classification accuracy to the conventional loss functions and can accelerate the training process on multiple datasets.

/pdf/approximating-the-gradient-of-cross-entropy-loss-function-3cb768m4k9.pdf

Approximating the Gradient of Cross-Entropy Loss Function

Power grid operation continuously experiences state transitions caused by the internal and external uncertainties, e.g., equipment failures and weather-driven faults. This prompts an observation of different types of waveforms at the measurement points (substations) in power systems captured by the phasor measurement units (PMUs) and intelligent electronic devices (IEDs) embedded with PMU functionality, e.g., digital relays and fault recorders. The PMU should be, hence, equipped with either one synchrophasor estimation algorithm (SEA) that is accurate and robust to many different types of signals any time across the network, or should adaptively select the promising SEA, among an embedded suite of algorithms. This paper proposes a PMU-embedded framework that can ensure real-time grid surveillance and potentially enables adaptive selection of SEA for more accurate synchrophasor estimation. Our proposed framework is consisted of two components: (i) a pseudo continuous quadrature wavelet transform (PCQ-WT) algorithm using a modified Gabor wavelet transform, which generates the featured-scalograms; and (ii) a convolutional neural network (CNN), that classifies the events based on the extracted features in the scalograms. Our experiments demonstrate that the proposed framework achieves state-of-the-art classification accuracy on multiple types of prevailing events in power grids, through which an enhanced grid-scale situational awareness in real-time can be realized.

Power Grid Online Surveillance through PMU-Embedded Convolutional Neural Networks

Geomagnetically induced currents (GICs) in power grids are mainly caused by geomagnetic disturbances especially during solar storms. Such currents can potentially cause negative impacts on power grid equipment and even damage the power transformers resulting in a significant risk of blackouts. Therefore, monitoring GICs in power systems and developing solutions to mitigate their impacts before rising to a certain threatening level is urgently in need. Monitoring GICs is, however, quite a challenge and costly, as they usually appear in forms of DC components in the high voltage transmission lines, which are barely accessible through transformers. By examining the measured currents from the current transformers (CTs), this paper proposes a framework to detect GICs in power transmission systems through a hybrid time-frequency analysis combined with machining learning technology. Simulated results verify that the proposed approach can promisingly estimate GICs in power systems during a variety of grid operating conditions.

Li Li

Papers

Power Grid Online Surveillance Through PMU-Embedded Convolutional Neural Networks

Discriminant Analysis Deep Neural Networks

Approximating the Gradient of Cross-Entropy Loss Function

Power Grid Online Surveillance through PMU-Embedded Convolutional Neural Networks

A Machine Learning Approach to Detection of Geomagnetically Induced Currents in Power Grids