An official website of the United States government
A machine learning-based detection framework is proposed to detect a class of cyber-attacks that redistribute loads by modifying measurements. The detection framework consists of a multi-output support vector regression (SVR) load predictor and a subsequent support vector machine (SVM) attack detector to determine the existence of load redistribution (LR) attacks utilizing loads predicted by the SVR predictor. Historical load data for training the SVR are obtained from the publicly available PJM zonal loads and are mapped to the IEEE 30-bus system. The features to predict loads are carefully extracted from the historical load data capturing both temporal and spatial correlations. The SVM attack detector is trained using normal data and randomly created LR attacks, so that it can maximally explore the attack space. An algorithm to create random LR attacks is introduced. The results show that the SVM detector trained merely using random attacks can effectively detect not only random attacks, but also intelligently designed attacks. Moreover, using the SVR predicted loads to re-dispatch generation when attacks are detected can significantly mitigate the attack consequences.
more »
« less
Data-Driven Generation of Synthetic Load Datasets Preserving Spatio-Temporal Features
A generative model for the creation of realistic historical bus-level load data for transmission grid models is presented. A data-driven approach based on principal component analysis is used to learn the spatio-temporal correlation between the loads in a system and build a generative model. Given a system topology and a set of base case loads, individual, realistic time-series data for each load can be generated. This technique is demonstrated by learning from a large proprietary dataset and generating historical data for the 2383-bus Polish test case.
more »
« less
- PAR ID:
- 10185959
- Date Published:
- Journal Name:
- 2019 IEEE Power & Energy Society General Meeting (PESGM)
- Page Range / eLocation ID:
- 1 to 5
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract The availability of large datasets is crucial for the development of new power system applications and tools; unfortunately, very few are publicly and freely available. The authors designed an end‐to‐end generative framework for the creation of synthetic bus‐level time‐series load data for transmission networks. The model is trained on a real dataset of over 70 Terabytes of synchrophasor measurements spanning multiple years. Leveraging a combination of principal component analysis and conditional generative adversarial network models, the developed scheme allows for the generation of data at varying sampling rates (up to a maximum of 30 samples per second) and ranging in length from seconds to years. The generative models are tested extensively to verify that they correctly capture the diverse characteristics of real loads. Finally, an opensource tool called LoadGAN is developed which gives researchers access to the fully trained generative models via a graphical interface.more » « less
-
null (Ed.)A framework for the generation of synthetic time-series transmission-level load data is presented. Conditional generative adversarial networks are used to learn the patterns of a real dataset of hourly-sampled week-long load profiles and generate unique synthetic profiles on demand, based on the season and type of load required. Extensive testing of the generative model is performed to verify that the synthetic data fully captures the characteristics of real loads and that it can be used for downstream power system and/or machine learning applications.more » « less
-
null (Ed.)In prior work, the Distributed Gradient Projection (DGP) algorithm was proposed to allow loads or load aggregators to provide contingency service to the grid using local frequency measurements. The DGP algorithm was shown to perform well in linear simulations. The goal of this work is to evaluate the performance of the DGP algorithm in more realistic scenarios and its robustness to issues of practical implementation, such as time delay, model mismatch, measurement noise, and stochastic disturbance. Simulation results from the IEEE 39-bus system indicate that the DGP algorithm performs well in mitigating the effects of contingencies and that it is robust to issues of practical implementation.more » « less
-
Topology optimization problems typically consider a single load case or a small, discrete number of load cases; however, practical structures are often subjected to infinitely many load cases that may vary in intensity, location and/or direction (e.g. moving/rotating loads or uncertain fixed loads). The variability of these loads significantly influences the stress distribution in a structure and should be considered during the design. We propose a locally stress-constrained topology optimization formulation that considers loads with continuously varying direction to ensure structural integrity under more realistic loading conditions. The problem is solved using an Augmented Lagrangian method, and the continuous range of load directions is incorporated through a series of analytic expressions that enables the computation of the worst-case maximum stress over all possible load directions. Variable load intensity is also handled by controlling the magnitude of load basis vectors used to derive the worst-case load. Several two- and three-dimensional examples demonstrate that topology-optimized designs are extremely sensitive to loads that vary in direction. The designs generated by this formulation are safer, more reliable, and more suitable for real applications, because they consider realistic loading conditions.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/PESGM40551.2019.8973532
