Data model update for structural-damage predictor after an earthquake

1. A method comprising:

identifying training data for a machine-learning program, the training data including values of features associated with structures, the features comprising:

characteristics of the structures;

fragility functions for predicting structural damage to the structures, each fragility function expressing a probability function for damage to one structure caused by a range of spectral accelerations caused by an earthquake, the fragility functions being stored in a database; and

values of damage to the structures caused by one or more earthquakes;

generating a model by training the machine-learning program with the training data:

estimating, using the model, a first damage to a first structure after a first earthquake utilizing the fragility functions for the first structure;

updating, using one or more processors, one or more of the fragility functions in the database for the first structure based on the first damage to the first structure when the first damage to the first structure is above a predetermined damage threshold;

accessing, using the one or more processors, shaking data for a second earthquake; and

estimating, using the model, a second damage to the first structure after the second earthquake utilizing the updated fragility functions for the first structure.

2. The method as recited in claim 1 , wherein the fragility function for an element of the structure expresses a probability that the element of the structure will suffer damage after the first earthquake based on a shaking suffered by the structure caused by the first earthquake.

3. The method as recited in claim 2 , wherein the fragility function has a mean and a standard deviation, wherein updating the one or more of the fragility functions due to structural damage during an earthquake further comprises:

decreasing the mean of the fragility function based on the first damage.

4. The method as recited in claim 1 , wherein updating the one or more of the fragility functions after the structure has suffered the first damage causes an increase of a probability of damage to the structure for the second earthquake.

5. The method as recited in claim 1 , wherein the predetermined damage threshold is defined by the structure having structural damage.

6. The method as recited in claim 1 , wherein updating the one or more of the fragility functions further comprises:

assigning a new fragility function that is a combination of two or more predefined fragility functions, each predefined fragility function of the combination having a respective weight.

7. The method as recited in claim 1 , further comprising:

receiving a notification that the structure has been repaired; and

updating the one or more fragility functions in response to the notification of the repair.

8. The method as recited in claim 1 , wherein the features are classified into built environment, natural environment, or instantaneous line, wherein the built environment is for structures that have been built, wherein the natural environment is for structures occurring in nature, wherein the instantaneous line comprises shaking data from one or more sensors in one or more locations.

9. The method as recited in claim 8 , wherein the built environment features comprise a structure location, a structure price, a year in which the structure was built, a number of stories of the structure, whether the structure is commercial or residential, and a building material of the structure.

10. A system comprising:

a memory comprising instructions;

a database; and

one or more computer processors, wherein the instructions, when executed by the one or more computer processors, cause the one or more computer processors to perform operations comprising:

identifying training data for a machine-learning program, the training data including values of features associated with structures, the features comprising:

characteristics of the structures;

fragility functions for predicting structural damage to the structures, each fragility function expressing a probability function for damage to one structure caused by a range of spectral accelerations caused by an earthquake, the fragility functions being stored in a database; and

values of damage to the structures caused by one or more earthquakes;

generating a model by training the machine-learning program with the training data;

estimating, using the model, a first damage to a first structure after a first earthquake utilizing the fragility functions for the first structure;

updating one or more of the fragility functions in the database for the first structure based on the first damage to the first structure when the first damage to the first structure is above a predetermined damage threshold;

accessing shaking data for a second earthquake; and

estimating, using the model, a second damage to the first structure after the second earthquake utilizing the updated fragility functions for the first structure.

11. The system as recited in claim 10 , wherein the fragility function for an element of the structure expresses a probability that the element of the structure will suffer damage after the first earthquake based on a shaking suffered by the structure caused by the first earthquake.

12. The system as recited in claim 11 , wherein the fragility function has a mean and a standard deviation, wherein updating the one or more of the fragility functions due to structural damage during an earthquake further comprises:

decreasing the mean of the fragility function based on the first damage.

13. The system as recited in claim 10 , wherein updating the one or more of the fragility functions after the structure has suffered the first damage causes an increase of a probability of damage to the structure for the second earthquake.

14. The system as recited in claim 10 , wherein updating the one or more of the fragility functions further comprises:

assigning a new fragility function that is a combination of two or more predefined fragility functions, each predefined fragility function of the combination having a respective weight.

15. A non-transitory machine-readable storage medium including instructions that, when executed by a machine, cause the machine to perform operations comprising:

identifying training data for a machine-learning program, the training data including values of features associated with structures, the features comprising:

characteristics of the structures;

fragility functions for predicting structural damage to the structures, each fragility function expressing a probability function for damage to one structure caused by a range of spectral accelerations caused by an earthquake the fragility functions being stored in a database; and

values of damage to the structures caused by one or more earthquakes;

generating a model by training the machine-learning program with the training data;

estimating, using the model, a first damage to a first structure after a first earthquake utilizing the fragility functions for the first structure;

updating one or more of the fragility functions in the database for the first structure based on the first damage to the first structure when the first damage to the first structure is above a predetermined damage threshold;

accessing shaking data for a second earthquake; and

estimating, using the model, a second damage to the first structure after the second earthquake utilizing the updated fragility functions for the first structure.

16. The machine-readable storage medium as recited in claim 15 , wherein the fragility function for an element of the structure expresses a probability that the element of the structure will suffer damage after the first earthquake based on a shaking suffered by the structure caused by the first earthquake.

17. The machine-readable storage medium as recited in claim 16 , wherein the fragility function has a mean and a standard deviation, wherein updating the one or more of the fragility functions due to structural damage during an earthquake further comprises:

decreasing the mean of the fragility function based on the first damage.

18. The machine-readable storage medium as recited in claim 15 ,

wherein updating the one or more of the fragility functions after the structure has suffered the first damage causes an increase of a probability of damage to the structure for the second earthquake.

19. The machine-readable storage medium as recited in claim 15 , wherein the machine further performs operations comprising:

receiving a notification that the structure has been repaired; and

updating the one or more fragility functions in response to the notification of the repair.

20. The machine-readable storage medium as recited in claim 15 , wherein updating the one or more of the fragility functions further comprises:

assigning a new fragility function that is a combination of two or more predefined fragility functions, each fragility function of the combination having a respective weight.