GNN-based physics solver for time-independent PDEs
Authors:
Rini Jasmine Gladstone,
Helia Rahmani,
Vishvas Suryakumar,
Hadi Meidani,
Marta D'Elia,
Ahmad Zareei
Abstract:
Physics-based deep learning frameworks have shown to be effective in accurately modeling the dynamics of complex physical systems with generalization capability across problem inputs. However, time-independent problems pose the challenge of requiring long-range exchange of information across the computational domain for obtaining accurate predictions. In the context of graph neural networks (GNNs)…
▽ More
Physics-based deep learning frameworks have shown to be effective in accurately modeling the dynamics of complex physical systems with generalization capability across problem inputs. However, time-independent problems pose the challenge of requiring long-range exchange of information across the computational domain for obtaining accurate predictions. In the context of graph neural networks (GNNs), this calls for deeper networks, which, in turn, may compromise or slow down the training process. In this work, we present two GNN architectures to overcome this challenge - the Edge Augmented GNN and the Multi-GNN. We show that both these networks perform significantly better (by a factor of 1.5 to 2) than baseline methods when applied to time-independent solid mechanics problems. Furthermore, the proposed architectures generalize well to unseen domains, boundary conditions, and materials. Here, the treatment of variable domains is facilitated by a novel coordinate transformation that enables rotation and translation invariance. By broadening the range of problems that neural operators based on graph neural networks can tackle, this paper provides the groundwork for their application to complex scientific and industrial settings.
△ Less
Submitted 27 March, 2023;
originally announced March 2023.
SuperCell: A Wide-Area Coverage Solution Using High-Gain, High-Order Sectorized Antennas on Tall Towers
Authors:
Pratheep Bondalapati,
Abhishek Tiwari,
Mustafa Emin Sahin,
Qi Tang,
Srishti Saraswat,
Vishvas Suryakumar,
Ali Yazdan,
Julius Kusuma,
Amit Dubey
Abstract:
In this article we introduce a novel solution called SuperCell, which can improve the return on investment (ROI) for rural area network coverage. SuperCell offers two key technical features: it uses tall towers with high-gain antennas for wide coverage and high-order sectorization for high capacity. We show that a solution encompassing a high-elevation platform in excess of 200 meters increases co…
▽ More
In this article we introduce a novel solution called SuperCell, which can improve the return on investment (ROI) for rural area network coverage. SuperCell offers two key technical features: it uses tall towers with high-gain antennas for wide coverage and high-order sectorization for high capacity. We show that a solution encompassing a high-elevation platform in excess of 200 meters increases coverage by 5x. Combined with dense frequency reuse by using as many as 36 azimuthal sectors from a single location, our solution can adequately serve the rural coverage and capacity demands. We validate this through propagation analysis, modeling, and experiments. The article gives a design perspective using different classes of antennas: Luneburg lens, active/passive phased array, and spatial multiplexing solutions. For each class, the corresponding analytical model of the resulting signal-to-interference plus noise ratio (SINR) based range and capacity prediction is presented. The spatial multiplexing solution is also validated through field measurements and additional 3D ray-tracing simulation. Finally, in this article we also shed light on two recent SuperCell field trials performed using a Luneburg lens antenna system. The trials took place in rural New Mexico and Mississippi. In the trials, we quantified the coverage and capacity of SuperCell in barren land and in a densely forested location, respectively. In the article, we demonstrate the results obtained in the trials and share the lessons learned regarding green-field and brown-field deployments.
△ Less
Submitted 21 December, 2020; v1 submitted 30 November, 2020;
originally announced December 2020.
