Back to results
US 2026/0119755 A1Published application
Wildfire Risk Mitigation Modelling with Wildfire Spread as a Directed Graph
AssigneeXylo Risk, Inc.
InventorsScott Farley, David Winnacker
Life of this patent
Estimating the term…
Abstract
Machine-learning-based methods and systems for vegetation treatment project design. The approach autonomously creates, evaluates, and optimizes wildfire risk mitigation activities to intelligently reduce wildfire risk to communities in a resource-constrained environment. The methods and/or systems are accompanied by a cloud-enabled user interface that provides decision support to vegetation management specialists charged with implementing wildfire risk reduction measures.
Claims21
A computer-implemented system for wildfire risk planning, where simulated wildfire spread is represented as a directed graph, the system comprising:
- a landscape database, configured for storing spatial descriptions of a landscape, including one or more of slope, land elevation, vegetation type, size, volume, arrangement, continuity, and other values relevant to prediction of fire behavior;
- an asset database, configured for storing geospatial locations of one or more community assets at risk;
- a graph generation module, configured to construct the directed graph to represent a temporal evolution of a wildfire front originating from an ignition location, under a weather scenario, wherein:
- nodes of the directed graph represent combustible fuel sources within a simulation domain;
- and
- edges of the directed graph represent directional fire spread between the nodes over time;
- a graph solver module, configured for selecting one or more solution pathways of a wildfire front applied to the directed graph, based at least in part on physical properties of the combustible fuel sources, including one or more of rate of spread, fuel load, and slope;
- a back-propagation module, configured for operating on each solution pathway to determine, for each node along each solution pathway:
- which community assets are exposed by that node;
- backpropagate the exposed community assets upstream along that solution pathway;
- and
- aggregate resulting exposed community assets along that solution pathway;
- a node importance rank module, configured to compute a node importance measure for each combustible fuel source node and each solution pathway based on one or more of:
- its topological position within the directed graph,
- its role within one or more fastest evolution pathways, or
- its cumulative asset exposure contribution as determined by the back-propagation module.
13 more elements
A computer-implemented system for wildfire risk planning, comprising:
- one or more processors;
- and
- a non-transitory memory storing instructions that, when executed by the one or more processors, cause the one or more processors to:
- (a) ingest multi-source geospatial and temporal data for a geographic region, the data including one or more of weather observations, weather forecasts, remotely-sensed imagery, topographic data, vegetation data, fuel condition data, and historical fire activity data;
- (b) normalize and georeference the ingested data to a common spatiotemporal index;
- (c) generate, for each cell of the geographic region and for each time interval within a selectable forecast horizon, a feature vector derived from the ingested data;
- (d) apply a trained machine-learned model to each feature vector to generate, for the corresponding cell and time interval, a wildfire behavior model representing fire arrival times and fire spread pathways across the geographic region, and including risk values indicative of a likelihood of ignition within the fire spread pathways;
- (e) construct a directed graph from the feature vectors and the fire spread pathways, with nodes of the directed graph representing geographically-located points within the geographic region, and edges of the directed graph representing the fire spread pathways between the points;
- (f) process the directed graph to produce calibrated risk values and corresponding uncertainty values;
- (g) select mitigation actions for at least a subset of the cells based on the calibrated risk values and the corresponding uncertainty values;
- (h) prioritize and schedule the mitigation actions subject to resource and operational constraints to form an actionable task plan;
- and
- (i) output the calibrated risk values, the uncertainty values, and the actionable task plan to an interface for display.
10 more elements
A computer-implemented method for wildfire risk mitigation planning, the method comprising:
- (a) receiving, by one or more processors, multi-source geospatial and temporal data for a geographic region, the data comprising at least weather fields, topography, vegetation or fuel condition data, and asset exposure data;
- (b) generating, from the data, a time-indexed spatial mesh of the geographic region comprising a plurality of cells at a temporal resolution and identifying a set of candidate ignition locations for a selectable forecast horizon;
- (c) simulating, from the candidate ignition locations and over the time-indexed spatial mesh, a plurality of fire spread pathways using a spread model to produce, for each pathway, an ordered sequence of cell transitions with associated propagation likelihoods and intensity estimates;
- (d) computing, for each fire spread pathway, an impact score based at least on the associated propagation likelihoods, the intensity estimates, and exposure of assets intersected by the pathway;
- (e) filtering the plurality of fire spread pathways by removing (i) pathways having a propagation likelihood below a threshold, (ii) dominated pathways based on a multi-criteria dominance test using probability and impact, and (iii) duplicate pathways based on path-overlap similarity;
- (f) constructing, from the filtered fire spread pathways, a directed graph comprising nodes representing respective (cell, time) states and edges representing transitions between the states, each edge annotated with a propagation likelihood and an incremental loss value, the directed graph including source nodes representing the candidate ignition locations and sink nodes representing termination states;
- (g) defining a library of candidate treatment actions, each treatment action comprising a geographic footprint mapped to a subset of the nodes or edges of the directed graph, an effect model specifying at least a reduction in propagation likelihood and/or intensity on affected edges, and an implementation cost;
- (h) evaluating, for each candidate treatment action and at least one combination of candidate treatment actions, a risk-reduction value computed as a change in expected loss on the directed graph when applying the effect model to the subset of nodes or edges associated with the candidate treatment action(s);
- (i) selecting, from the candidate treatment actions, a treatment plan comprising a subset of the candidate treatment actions that maximizes a risk-reduction objective subject to one or more constraints comprising a budget constraint and at least one resource, geographic, or timing constraint;
- and
- (j) outputting the treatment plan together with geospatial identifiers and scheduling information to a planning interface for execution.
8 more elements
Dependencies are inferred from the claim text.