Filters








58 Hits in 7.7 sec

Graph Neural Network and Koopman Models for Learning Networked Dynamics: A Comparative Study on Power Grid Transients Prediction [article]

Sai Pushpak Nandanoori, Sheng Guan, Soumya Kundu, Seemita Pal, Khushbu Agarwal, Yinghui Wu, Sutanay Choudhury
2022 arXiv   pre-print
The other class of models is based on the graph convolutional neural networks which are adept at capturing the inherent spatio-temporal correlations within the power network.  ...  In particular, accurate and timely prediction of the (electro-mechanical) transient dynamic trajectories of the power grid is necessary for early detection of any instability and prevention of catastrophic  ...  Conclusion and Future Work Using the example of power systems transient dynamics, we presented a comparative study of different data-driven models, based on the Koopman operator theory and the graph convolutional  ... 
arXiv:2202.08065v1 fatcat:iszdezbr5rhjjepuhc5vvxgufm

Graph Neural Network and Koopman Models for Learning Networked Dynamics: A Comparative Study on Power Grid Transients Prediction

Sai Pushpak Nandanoori, Sheng Guan, Soumya Kundu, Seemita Pal, Khushbu Agarwal, Yinghui Wu, Sutanay Choudhury
2022 IEEE Access  
INDEX TERMS Data-driven techniques, graph neural network, Koopman operator, transient dynamics.  ...  In particular, accurate and timely prediction of the (electro-mechanical) transient dynamic trajectories of the power grid is necessary for early detection of any instability and prevention of catastrophic  ...  CONCLUSION AND FUTURE WORK Using the example of power systems transient dynamics, we presented a comparative study of different data-driven models, based on the Koopman operator theory and the graph convolutional  ... 
doi:10.1109/access.2022.3160710 fatcat:b7e7mzojljbrjbcbvgzbsrrl4i

A critical review of data-driven transient stability assessment of power systems: principles, prospects and challenges [article]

Shitu Zhang, Zhixun Zhu, Yang Li
2021 arXiv   pre-print
Transient stability assessment (TSA) has always been a fundamental means for ensuring the secure and stable operation of power systems.  ...  This paper makes a comprehensive review from the following four aspects: feature extraction and selection, model construction, online learning and rule extraction; and then, summarizes the challenges and  ...  Based on a message transfer graph neural network, a fast transient stability assessment method based on steady-state data is proposed for power systems in reference [70] .  ... 
arXiv:2111.00978v1 fatcat:byrrmsopbfdnxjghsw4vn7p4im

A Critical Review of Data-Driven Transient Stability Assessment of Power Systems: Principles, Prospects and Challenges

Shitu Zhang, Zhixun Zhu, Yang Li
2021 Energies  
Transient stability assessment (TSA) has always been a fundamental means for ensuring the secure and stable operation of power systems.  ...  This paper makes a comprehensive review from the following four aspects: feature extraction and selection, model construction, online learning and rule extraction; and then, summarizes the challenges and  ...  Reference [29] proposes a TSA and instability mode prediction model based on convolutional neural network (CNN).  ... 
doi:10.3390/en14217238 fatcat:ioui7lgvlvb6ne33peldzmmdla

Modern Koopman Theory for Dynamical Systems [article]

Steven L. Brunton, Marko Budišić, Eurika Kaiser, J. Nathan Kutz
2021 arXiv   pre-print
in terms of measurements, making it ideal for leveraging big-data and machine learning techniques, and 3) simple, yet powerful numerical algorithms, such as the dynamic mode decomposition (DMD), have  ...  Koopman spectral theory has emerged as a dominant perspective over the past decade, in which nonlinear dynamics are represented in terms of an infinite-dimensional linear operator acting on the space of  ...  We also thank Shervin Bagheri, Bing Brunton, Bethany Lusch, Ryan Mohr, Frank Noe, Josh Proctor, Clancy Rowley, and Peter Schmid for many fruitful discussions on DMD, Koopman theory, and control.  ... 
arXiv:2102.12086v2 fatcat:2oylyx25dbctvkjfnirfcgjuqu

Forecasting Sequential Data using Consistent Koopman Autoencoders [article]

Omri Azencot and N. Benjamin Erichson and Vanessa Lin and Michael W. Mahoney
2020 arXiv   pre-print
Recurrent neural networks are widely used on time series data, yet such models often ignore the underlying physical structures in such sequences.  ...  We evaluate our method on a wide range of high-dimensional and short-term dependent problems, and it achieves accurate estimates for significant prediction horizons, while also being robust to noise.  ...  Brunton, Lionel Mathelin and Alejandro Queiruga for valu-  ... 
arXiv:2003.02236v2 fatcat:3nz7rko5ofakrbkqcqglfgudde

Tractable Dendritic RNNs for Reconstructing Nonlinear Dynamical Systems [article]

Manuel Brenner, Florian Hess, Jonas M. Mikhaeil, Leonard Bereska, Zahra Monfared, Po-Chen Kuo, Daniel Durstewitz
2022 arXiv   pre-print
Motivated by the emerging principles of dendritic computation, we augment a dynamically interpretable and mathematically tractable piecewise-linear (PL) recurrent neural network (RNN) by a linear spline  ...  We show that the dendritically expanded PLRNN achieves better reconstructions with fewer parameters and dimensions on various dynamical systems benchmarks and compares favorably to other methods, while  ...  Acknowledgements This work was funded by the German Research Foundation (DFG) within Germany's Excellence Strategy -EXC-2181 -390900948 ('Structures'), by DFG grant Du354/10-1 to DD, and the European Union  ... 
arXiv:2207.02542v1 fatcat:y5rnvigbyzdqrn67apk4cza45q

Data-driven approximations of dynamical systems operators for control [article]

Eurika Kaiser, J. Nathan Kutz, Steven L. Brunton
2019 arXiv   pre-print
The Koopman and Perron Frobenius transport operators are fundamentally changing how we approach dynamical systems, providing linear representations for even strongly nonlinear dynamics.  ...  Obtaining low-dimensional matrix approximations of these operators is paramount for applications, and the dynamic mode decomposition has quickly become a standard numerical algorithm to approximate the  ...  Acknowledgements EK gratefully acknowledges support by the "Washington Research Foundation Fund for Innovation in Data-Intensive Discovery" and a Data Science Environments project award from the Gordon  ... 
arXiv:1902.10239v1 fatcat:23bw5qdalbenddnelslxtmes7m

Deep Learning for Reduced Order Modelling and Efficient Temporal Evolution of Fluid Simulations [article]

Pranshu Pant, Ruchit Doshi, Pranav Bahl, Amir Barati Farimani
2021 arXiv   pre-print
In this work, we develop a novel deep learning framework DL-ROM (Deep Learning - Reduced Order Modelling) to create a neural network capable of non-linear projections to reduced order states.  ...  Our model DL-ROM is able to create highly accurate reconstructions from the learned ROM and is thus able to efficiently predict future time steps by temporally traversing in the learned reduced state.  ...  To make this DNS dataset amenable for use with our deep learning model we sample the dataset on a uniform grid of 512 x 128 at 2500 timesteps.  ... 
arXiv:2107.04556v1 fatcat:sqgherbi45fh7fg2kq42ry3w5m

Integrating Scientific Knowledge with Machine Learning for Engineering and Environmental Systems [article]

Jared Willard, Xiaowei Jia, Shaoming Xu, Michael Steinbach, Vipin Kumar
2022 arXiv   pre-print
Application-centric objective areas for which these approaches have been applied are summarized, and then classes of methodologies used to construct physics-guided ML models and hybrid physics-ML frameworks  ...  We then provide a taxonomy of these existing techniques, which uncovers knowledge gaps and potential crossovers of methods between disciplines that can serve as ideas for future research.  ...  in a graph neural network.  ... 
arXiv:2003.04919v6 fatcat:r4gbmcxqjzhhdfx65tmj7fz2de

Parameter Identification of Electromechanical Oscillation Mode in Power Systems Driven by Data: A Quasi-Real-Time Method Based on Randomized-DMD-Multilayer Artificial Neural Networks

Guowei Cai, Shujia Guo, Cheng Liu
2022 Frontiers in Energy Research  
Therefore, a method based on machine learning (multilayer artificial neural networks) is proposed to identify the electromechanical oscillation mode parameters of the power system.  ...  This method can take the monitorable variables of the WAMS as the input of the model and the key characteristic information such as frequency and damping ratio as the output.  ...  search, data acquisition, data analysis, case studies, and manuscript preparation and editing. CL performed manuscript review.  ... 
doi:10.3389/fenrg.2022.908937 fatcat:2morux7jyvgzlkebixpnc5ne2y

CD-ROM: Complementary Deep-Reduced Order Model [article]

Emmanuel Menier, Michele Alessandro Bucci, Mouadh Yagoubi, Lionel Mathelin, Marc Schoenauer
2022 arXiv   pre-print
We use multi layer perceptrons (MLP) to learn a continuous in time closure model through the recently proposed Neural ODE method.  ...  This paper proposes a closure modeling approach for classical POD-Galerkin reduced order models (ROM).  ...  At the same time, the deep learning community has developed powerful tools by adapting physical modeling concepts to deep learning, for instance the Neural ODE approach was proposed for the modeling of  ... 
arXiv:2202.10746v2 fatcat:kw5hmypctjhjfg2tdy7jvqff3m

Data-Driven Modeling and Prediction of Non-Linearizable Dynamics via Spectral Submanifolds [article]

Mattia Cenedese, Joar Axås, Bastian Bäuerlein, Kerstin Avila, George Haller
2022 arXiv   pre-print
We illustrate the power of data-driven SSM reduction on high-dimensional numerical data sets and experimental measurements involving beam oscillations, vortex shedding and sloshing in a water tank.  ...  We develop a methodology to construct low-dimensional predictive models from data sets representing essentially nonlinear (or non-linearizable) dynamical systems with a hyperbolic linear part that are  ...  Another recent approach is cluster-based network modeling, which uses the toolkit of network science and statistical physics for modeling nonlinear dynamics [11] .  ... 
arXiv:2201.04976v1 fatcat:eoooz3bdujfs7c4mqr5w3smz2a

Table of Contents

2022 IEEE Robotics and Automation Letters  
Beetz, and U. Frese Multi-Robot Collaborative Perception With Graph Neural Networks . . . . . . ...Y. Zhou, J. Xiao, Y. Zhou, and G.  ...  Kim Comparing Single Touch to Dynamic Exploratory Procedures for Robotic Tactile Object Recognition .  ... 
doi:10.1109/lra.2022.3165102 fatcat:enjzebowe5hn7hsfwklc7nieuy

Physics-based Deep Learning [article]

Nils Thuerey and Philipp Holl and Maximilian Mueller and Patrick Schnell and Felix Trost and Kiwon Um
2022 arXiv   pre-print
This digital book contains a practical and comprehensive introduction of everything related to deep learning in the context of physical simulations.  ...  modeling.  ...  Interestingly, unstructured meshes share many properties with graph neural networks (GNNs), which extend the classic ideas of DL on Cartesian grids to graph structures.  ... 
arXiv:2109.05237v3 fatcat:pz7ot63dlbdkriihkwloefk3im
« Previous Showing results 1 — 15 out of 58 results