An official website of the United States government
DC microgrids incorporate several converters for distributed energy resources connected to different passive and active loads. The complex interactions between the converters and components and their potential failures can significantly affect the grids' resilience and health; hence, they must be continually assessed and monitored. This paper presents a machine learning-assisted prognostic health monitoring (PHM) and diagnosis approach, enabling progressive interactions between the converters at multiple nodes to dynamically examine the grid's (or micro-grid's) health in real time. By measuring the resulting impedance at the power converters' terminals at various grid nodes, a neural network-based classifier helps detect the grid's health condition and identify the potential fault-prone zones, along with the type and location of the fault type in the grid topology. For a faulty grid, a Naive Bayes and a support vector machine (SVM)-based classifiers are used to locate and identify the faulty type, respectively. A separate neural network-based regression model predicts the source power delivered and the loads at different terminals in a healthy grid network. The proposed concepts are supported by detailed analysis and simulation results in a simple four-terminal DC microgrid topology and a standard IEEE 5 Bus system.
more »
« less
Power Grid Anomaly Detection via Hybrid LSTM-GIN Model (Student Abstract)
Cyberattacks on power grids pose significant risks to national security. Power grid attacks typically lead to abnormal readings in power output, frequency, current, and voltage. Due to the interconnected structure of power grids, abnormalities can spread throughout the system and cause widespread power outages if not detected and dealt with promptly. Our research proposes a novel anomaly detection system for power grids that prevents overfitting. We created a network graph to represent the structure of the power grid, where nodes represent power grid components like generators and edges represent connections between nodes such as overhead power lines. We combine the capabilities of Long Short-Term Memory (LSTM) models with a Graph Isomorphism Network (GIN) in a hybrid model to pinpoint anomalies in the grid. We train our model on each category of nodes that serves a similar structural purpose to prevent overfitting of the model. We then assign each node in the graph a unique signature using a GIN. Our model achieved a 99.92% accuracy rate, which is significantly higher than a version of our model without structural encoding, which had an accuracy level of 97.30%. Our model allows us to capture structural and temporal components of power grids and develop an attack detection system with high accuracy without overfitting.
more »
« less
- Award ID(s):
- 1950599
- PAR ID:
- 10570324
- Publisher / Repository:
- AAAI 2024
- Date Published:
- Journal Name:
- Proceedings of the AAAI Conference on Artificial Intelligence
- Volume:
- 38
- Issue:
- 21
- ISSN:
- 2159-5399
- Page Range / eLocation ID:
- 23525 to 23527
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
We present a model for recognizing typeset math formula images from connected components or symbols. In our approach, connected components are used to construct a line-of-sight (LOS) graph. The graph is used both to reduce the search space for formula structure interpretations, and to guide a classification attention model using separate channels for inputs and their local visual context. For classification, we used visual densities with Random Forests for initial development, and then converted this to a Convolutional Neural Network (CNN) with a second branch to capture context for each input image. Formula structure is extracted as a directed spanning tree from a weighted LOS graph using Edmonds’ algorithm. We obtain strong results for formulas without grids or matrices in the InftyCDB-2 dataset (90.89% from components, 93.5% from symbols). Using tools from the CROHME handwritten formula recognition competitions, we were able to compile all symbol and structure recognition errors for analysis. Our data and source code are publicly available.more » « less
-
Phasor Measurement Units (PMU), due to their capability for providing highly precise and time-synchronized measurements of synchrophasors, have now become indispensable in wide area monitoring of power-grid systems. Successful and reliable delivery of synchrophasor packets from the PMUs to the Phasor Data Concentrators (PDCs) and beyond, requires a backbone communication network that is robust and resilient to failures. These networks are vulnerable to a range of failures that include cyber-attacks, system or device level outages and link failures. In this paper, we present a framework to evaluate the resilience of a PMU network in the context of link failures. We model the PMU network as a connected graph and link failures as edges being removed from the graph. Our approach, inspired by model checking methods, involves exhaustively checking the reachability of PMU nodes to PDC nodes, for all possible combinations of link failures, given an expected number of links fail simultaneously. Using the IEEE 14-bus system, we illustrate the construction of the graph model and the solution design. Finally, a comparative evaluation on how adding redundant links to the network improves the Power System Observability, is performed on the IEEE 118 bus-system.more » « less
-
Complex systems such as smart cities and smart power grids rely heavily on their interdependent components. The failure of a component in one network may lead to the failure of the supported component in another network. Components which support a large number of interdependent components may be more vulnerable to attacks and failures. In this paper, we study the robustness of two interdependent networks under node failures. By modeling each network using a random geometric graph (RGG), we study conditions for the percolation of two interdependent RGGs after in-homogeneous node failures. We derive analytical bounds on the interdependent degree thresholds (k 1 ,k 2 ), such that the interdependent RGGs percolate after removing nodes in G i that support more than k j nodes in G j (∀i, j ∈ {1, 2}, i ≠ j). We verify the bounds using numerical simulation, and show that there is a tradeoff between k 1 and k 2 for maintaining percolation after the failures.more » « less
-
Inertia from rotating masses of generators in power systems influence the instantaneous frequency change when an imbalance between electrical and mechanical power occurs. Renewable energy sources (RES), such as solar and wind power, are connected to the grid via electronic converters. RES connected through converters affect the system's inertia by decreasing it and making it time-varying. This new setting challenges the ability of current control schemes to maintain frequency stability. Proposing adequate controllers for this new paradigm is key for the performance and stability of future power grids. The contribution of this paper is a framework to learn sparse time-invariant frequency controllers in a power system network with a time-varying evolution of rotational inertia. We model power dynamics using a Switched-Affine hybrid system to consider different modes corresponding to different inertia coefficients. We design a controller that uses as features, i.e. input, the systems states. In other words, we design a control proportional to the angles and frequencies. We include virtual inertia in the controllers to ensure stability. One of our findings is that it is possible to restrict communication between the nodes by reducing the number of features in the controller (from 22 to 10 in our case study) without disrupting performance and stability. Furthermore, once communication between nodes has reached a threshold, increasing it beyond this threshold does not improve performance or stability. We find a correlation between optimal feature selection in sparse controllers and the topology of the network.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1609/AAAI.V38I21.30457
