[CCoE Notice] Dissertation Announcement: Thuan Pham, "Faster Optimal Power Flow Using Graph Neural Network-Assisted Methods"

[Greenwell, Stephen J]

[cid:image001.png at 01DB3A7A.26A377F0]
Faster Optimal Power Flow Using
Graph Neural Network-Assisted Methods

Thuan Pham
December 3rd, 2024; 9:00 AM - 11:00 AM
Location: Zoom
https://urldefense.com/v3/__https://uh-edu-cougarnet.zoom.us/j/97669241989__;!!LkSTlj0I!CxjldWXkq3FdIIImbiRg0mrWDXU4HUUuTUt-w5LYHzaRWSZUAWdbfOTP1b64n4cf1PMl9y93ugVcdhmUr-3STCu9vGc$ 
Committee Chair:
Xingpeng Li, Ph.D.
Committee Members:
Kaushik Rajashekara, Ph.D. | David Jackson, Ph.D. | Harish Krishnamoorthy, Ph.D. | Lei Fan, Ph.D.
Abstract
Optimal Power Flow (OPF) is a critical tool for power system operations, analysis, and scheduling, including real-time economic dispatch. It involves optimizing an objective function, such as minimizing generation costs, while adhering to physical and operational constraints like generation limits and line thermal ratings. The applications of OPF have broadened significantly, addressing challenges in grid management and resource allocation.
Traditional power system optimization models often fall short, either due to insufficient predictive capabilities or slow computation times. Advanced machine learning approaches, such as Graph Neural Networks (GNNs), have emerged as a promising solution. By leveraging the topology of the power network, GNNs enable the flow of information between adjacent nodes and edges across multiple layers, capturing both local and global relationships within the graph.
This dissertation investigates the application of GNNs to optimize OPF problems. GNNs are utilized to simplify the complexities of OPF by predicting line congestion and the maximum generation capacity of resources. Furthermore, the adoption of GNNs for N-1 contingency prediction has demonstrated significant reductions in computational time. GNNs have also been applied to network-reconfigured OPF scenarios, where they optimize network topology to improve solution efficiency and operational performance.
[cid:image002.png at 01DB3A7A.26A377F0]

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://Bug.EGR.UH.EDU/pipermail/engi-dist/attachments/20241119/d4e9b64c/attachment-0001.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 22744 bytes
Desc: image001.png
Url : http://Bug.EGR.UH.EDU/pipermail/engi-dist/attachments/20241119/d4e9b64c/attachment-0002.png 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.png
Type: image/png
Size: 6744 bytes
Desc: image002.png
Url : http://Bug.EGR.UH.EDU/pipermail/engi-dist/attachments/20241119/d4e9b64c/attachment-0003.png