Method for prediction of a surface event

• US 11,555,942 B2
• Filed: 04/24/2018
• Issued: 01/17/2023
• Est. Priority Date: 04/24/2017
• Status: Active Grant

• Current Assignee: University of Melbourne
• Original Assignee: University of Melbourne
• Inventors: Tordesillas, Antoinette, Batterham, Robin
• Primary Examiner(s): Le, Toan M
• Application Number:
  US16/606,453
• Publication Number:
  US 20200132881A1
• Time in Patent Office: 1,729 Days
• CPC Class Codes: G01V 20/00 Geomodelling in general
• CPC Class Codes: G01V 99/00 Subject matter not provided for in other groups of this subclass
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G08B 31/00 Predictive alarm systems characterised by extrapolation or other computation using updated historic data

Abstract

Methods and systems for predicting surface failure of a surface, for example a method comprising the steps of: obtaining a group of measured datasets, each including: a measurement value of at least a first type for each of a plurality of grid elements, each grid element associated with a location on the surface; and a time value, such that the group of datasets includes datasets associated with a plurality of unique time values, identifying an interface set of grid elements for each measured dataset, each interface set comprising grid elements of the associated measured dataset meeting a connection threshold according to a connection rule in dependence on the measurement values of the grid elements, determining a risk of surface failure in accordance with identification of a pattern of grid elements of the interface set which has a persistent location with respect to the surface of interface sets over a plurality of measured datasets.

Claims

1. A method for predicting surface failure of a surface, the method comprising the steps of:

obtaining a group of measured datasets, each including;

a measurement value of at least a first type for each of a plurality of grid elements, each grid element associated with a location on the surface; and

a time value, such that the group of datasets includes datasets associated with a plurality of unique time values,

for each measured dataset;

constructing a network of nodes in a state space of the measurement values, wherein each node is uniquely associated with a grid element, wherein the network of nodes enables, for a selected node, a determination of a connectedness of said selected node within said state space to all other nodes of the network of nodes; and

identifying an interface set of grid elements comprising a subset of the grid elements of the measured dataset corresponding to the grid elements meeting a connection threshold according to a connection rule in dependence on the measurement values of the grid elements, wherein the connection threshold defines a minimum connectedness to all other grid elements, and

determining a risk of surface failure in accordance with identification of a pattern defined by the locations of the grid elements of the interface sets which has a persistent location with respect to the surface over a plurality of measured datasets, such that the pattern has the persistent location over a period of time.

2. A method as claimed in claim 1, further comprising the steps of:

assigning a connection to every pair of nodes meeting a connection criterion, wherein the connection criterion defines a required closeness of the measurement values of the pair of nodes in order for the pair to be assigned the connection,

wherein the grid elements of the interface set of grid elements are determined in accordance with identification of a set of nodes having a closest connection to all other nodes.

3. A method as claimed in claim 2, wherein the step of constructing a network of nodes includes:

connecting each node to k other nodes in accordance with a node connection rule on the basis of a closeness in measurement value between the node to the k other nodes,

wherein k is selected in accordance with a minimum number of connections for each node that provides a connected network.

4. A method as claimed in claim 3, wherein the step of constructing a network of nodes further includes:

determining, for each node, a closeness centrality for the node; and

identifying a predetermined number of nodes for the set of nodes having largest closeness centralities.

5. A method as claimed in claim 4, wherein the predetermined number is between 10% and 30% of the total number of nodes.

6. A method as claimed in claim 4, wherein the step of constructing a network of nodes further includes:

generating a plurality of candidate interface sets,

wherein the interface set is selected from one of the plurality of candidate interface sets.

7. A method as claimed in claim 4, wherein, for each nodes i, the closeness centrality C_c is determined according to the following equation:

8. A method as claimed in claim 1, wherein the measurement values comprise movement values.

9. A method as claimed in claim 1, wherein the measured datasets are generated periodically or at least substantially periodically.

10. A method as claimed in claim 1, further comprising the step of:

generating an alert in response to determining the risk exceeds a predefined risk threshold.

11. A method as claimed in claim 1, further comprising the step of generating a plurality of spatial synthetic datasets, wherein the step of identifying an interface set of grid elements for each measured dataset comprises analyzing the one or more spatial synthetic datasets in addition to the measured datasets.

12. A method as claimed in claim 1, further comprising the step of generating a plurality of temporal synthetic datasets, wherein the step of identifying an interface set of grid elements for each measured dataset comprises analyzing the one or more temporal synthetic datasets in addition to the measured datasets.

13. A method for predicting surface failure of a surface, the method comprising the steps of:

obtaining a group of measured datasets, each including;

a measurement value of at least a first type for each of a plurality of grid elements, each grid element associated with a location on the surface; and

a time value, such that the group of datasets includes datasets associated with a plurality of unique time values,

for each measured dataset, analysing the measurement values to identify one or more clusters of measured values and at least one outlier measured value not associated with a cluster, wherein the, or each, cluster represents a plurality of grid elements having similar measurement values according to a predefined criterion to one another and wherein the remaining grid elements not associated with a cluster are identified to be outliers,

identifying an interface set of grid elements for each measured dataset, each interface set comprising a subset of the grid elements of the associated measured dataset corresponding to the grid elements identified as outliers, and

determining a risk of surface failure in accordance with identification of a pattern defined by the locations of the grid elements of the interface sets which has a persistent location with respect to the surface over a plurality of measured datasets, such that the pattern has the persistent location over a period of time.

14. A method as claimed in claim 13, wherein the measurement values comprise movement values.

15. A method as claimed in claim 13, wherein the measured datasets are generated periodically or at least substantially periodically.

16. A method as claimed in claim 13, further comprising the step of:

generating an alert in response to determining the risk exceeds a predefined risk threshold.

17. A method as claimed in claim 13, further comprising the step of generating a plurality of spatial synthetic datasets, wherein the step of identifying an interface set of grid elements for each measured dataset comprises analyzing the one or more spatial synthetic datasets in addition to the measured datasets.

18. A method as claimed in claim 13, further comprising the step of generating a plurality of temporal synthetic datasets, wherein the step of identifying an interface set of grid elements for each measured dataset comprises analyzing the one or more temporal synthetic datasets in addition to the measured datasets.

19. A surface failure prediction system comprising:

a measurement apparatus configured to;

obtain a group of measured datasets, each including;

a measurement value of at least a first type for each of a plurality of grid elements, each grid element associated with a location on the surface; and

a time value, such that the group of datasets includes datasets associated with a plurality of unique time values, and

a computing device configured to;

for each measured dataset;

construct a network of nodes in a state space of the measurement values, wherein each node is uniquely associated with a grid element, wherein the network of nodes enables, for a selected node, a determination of a connectedness of said selected node within said state space to all other nodes of the network of nodes; and

identify an interface set of grid elements comprising a subset of the grid elements of the associated measured dataset corresponding to the grid elements meeting a connection threshold according to a connection rule in dependence on the measurement values of the grid elements, wherein the connection threshold defines a minimum connectedness to all other grid elements, and

determine a risk of surface failure in accordance with identification of a pattern defined by the locations of the grid elements of the interface sets which has a persistent location with respect to the surface over a plurality of measured datasets, such that the pattern has the persistent location over a period of time.

20. A surface failure prediction system comprising:

a measurement apparatus configured to;

obtain a group of measured datasets, each including;

a measurement value of at least a first type for each of a plurality of grid elements, each grid element associated with a location on the surface; and

a time value, such that the group of datasets includes datasets associated with a plurality of unique time values,

for each measured dataset, analyse the measured values to identify one or more clusters of grid elements based on a similarity between the respective measurement values and at least one outlier grid element not associated with a cluster, wherein the, or each, cluster represents grid element having similar measurement values according to a predefined criterion and wherein the remaining grid elements are identified to be outliers,

identify an interface set of grid elements for each measured dataset, each interface set comprising grid elements of the associated measured dataset corresponding to the grid elements identified as outliers, and

determine a risk of surface failure in accordance with identification of a pattern defined by the locations of the of grid elements of the interface sets which has a persistent location with respect to the surface over a plurality of measured datasets, such that the pattern has the persistent location over a period of time.