US 12,494,972 B2
Network management based on modeling of cascading effect of failure
Aanchal Aggarwal, White Plains, NY (US); Younghun Kim, Pleasantville, NY (US); Tarun Kumar, Chappaqua, NY (US); Mark A. Lavin, Katonah, NY (US); Giuliano Andrea Pagani, Groningen (NL); and Abhishek Raman, Mahopac, NY (US)
Assigned to Utopus Insights, Inc., Valhalla, NY (US)
Filed by Utopus Insights, Inc., Valhalla, NY (US)
Filed on May 9, 2023, as Appl. No. 18/314,628.
Application 18/314,628 is a continuation of application No. 16/546,276, filed on Aug. 20, 2019, granted, now 11,683,238.
Application 16/546,276 is a continuation of application No. 14/665,316, filed on Mar. 23, 2015, granted, now 10,389,600, issued on Aug. 20, 2019.
Prior Publication US 2024/0007357 A1, Jan. 4, 2024
Int. Cl. G06N 7/01 (2023.01); H04L 41/0631 (2022.01); H04L 41/14 (2022.01); H04L 41/142 (2022.01); G06N 20/00 (2019.01)
CPC H04L 41/145 (2013.01) [G06N 7/01 (2023.01); H04L 41/0636 (2013.01); H04L 41/065 (2013.01); H04L 41/142 (2013.01); G06N 20/00 (2019.01); Y04S 40/00 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method of managing a network with utility assets of an electrical grid, the method comprising:
generating, using a processor, a directed graph with each of the utility assets represented as a node in the directed graph, the utility assets including electrical assets of the electrical grid;
determining a historical failure rate for each node of at least a portion of the utility assets over a predetermined period of time;
determining a relationship of the historical failure rates with one or more physical attributes of one or more of the utility assets;
generating a training model based on the historical failure rates and the relationships of the historical failure rates with the one or more physical attributes of the one or more utility assets;
determining, using the training model, an individual failure probability of each node;
computing, using the processor, an upstream failure probability of each of at least a subset of the nodes of the directed graph according to an arrangement of the nodes in the directed graph, the computing the upstream failure probability including:
determining a path between a selected node of the at least a subset of nodes and an initial failure node of the directed graph, the initial failure node representing a source node from which the directed graph to affect downstream nodes; and
computing the upstream failure probability of the selected node based on aggregated failure probabilities of nodes along the determined path between the selected node and the initial failure node;
computing a network failure probability for each node of the at least the subset of nodes based on the corresponding individual failure probability and the corresponding upstream failure probability; and
increasing frequency of inspection of one or more nodes based on the network failure probability of the nodes probability.