System and methods for failure occurrence prediction and failure duration estimation
Aspects of the present disclosure are directed to systems and methods for determining execution of failure prediction models and duration prediction models for a sensor system. Systems and methods can involve receiving streaming data from one or more sensors and for a failure prediction model processing the streaming data indicating a predicted failure with a probability higher than a threshold, obtaining a duration of the predicted failure from a duration prediction model configured to predict durations of detected failures based on the streaming data; deactivating the failure prediction model when the predicted failure occurs; and determining a time to reactivate the failure prediction model based on the obtained duration of the predicted failure.
1. A method, comprising:
receiving streaming data from one or more sensors;
for a failure prediction model processing the streaming data indicating a predicted failure with a probability higher than a threshold:
obtaining a duration of the predicted failure from a duration prediction model configured to predict durations of detected failures based on the streaming data;
deactivating the failure prediction model when the predicted failure occurs; and
determining a time to reactivate the failure prediction model based on the obtained duration of the predicted failure.
2. The method of claim 1 , wherein the duration prediction model is executed on the streaming data continuously.
3. The method of claim 2 , wherein the failure prediction model is continuously executed on the streaming data, wherein the deactivating the failure prediction model when the failure prediction occurs comprises discarding output from the failure prediction model until the failure prediction model is reactivated.
4. The method of claim 1 , wherein the determining the time to reactivate the failure prediction model comprises:
determining an end time of the predicted failure; and
determining the time to reactivate the failure prediction model within an alert window time period before the end time of the predicted failure.
5. The method of claim 1 , wherein the determining the time to reactivate the failure prediction model comprises:
executing a modified failure prediction model on the streaming data configured to detect a successive failure after the predicted failure; and
determining the time to reactivate the failure prediction model as when the predicted failure ends.
6. The method of claim 1 , further comprising reactivating the failure prediction model for processing the stream data at the determined time.
7. The method of claim 1 , wherein the duration prediction model is trained to predict durations based on start points and end points of historical failure data.
8. A non-transitory computer readable medium, storing instructions for executing a process, the instructions comprising:
receiving streaming data from one or more sensors;
for a failure prediction model processing the streaming data indicating a predicted failure with a probability higher than a threshold:
obtaining a duration of the predicted failure from a duration prediction model configured to predict durations of detected failures based on the streaming data;
deactivating the failure prediction model when the predicted failure occurs; and
determining a time to reactivate the failure prediction model based on the obtained duration of the predicted failure.
9. The non-transitory computer readable medium of claim 8 , wherein the duration prediction model is executed on the streaming data continuously.
10. The non-transitory computer readable medium of claim 9 , wherein the failure prediction model is continuously executed on the streaming data, wherein the deactivating the failure prediction model when the failure prediction occurs comprises discarding output from the failure prediction model until the failure prediction model is reactivated.
11. The non-transitory computer readable medium of claim 8 , wherein the determining the time to reactivate the failure prediction model comprises:
determining an end time of the predicted failure; and
determining the time to reactivate the failure prediction model within an alert window time period before the end time of the predicted failure.
12. The non-transitory computer readable medium of claim 8 , wherein the determining the time to reactivate the failure prediction model comprises:
executing a modified failure prediction model on the streaming data configured to detect a successive failure after the predicted failure; and
determining the time to reactivate the failure prediction model as when the predicted failure ends.
13. The non-transitory computer readable medium of claim 8 , further comprising reactivating the failure prediction model for processing the stream data at the determined time.
14. The non-transitory computer readable medium of claim 8 , wherein the duration prediction model is trained to predict durations based on start points and end points of historical failure data.
15. An apparatus, comprising:
a processor, configured to:
receive streaming data from one or more sensors;
for a failure prediction model processing the streaming data indicating a predicted failure with a probability higher than a threshold:
obtain a duration of the predicted failure from a duration prediction model configured to predict durations of detected failures based on the streaming data;
deactivate the failure prediction model when the predicted failure occurs; and
determine a time to reactivate the failure prediction model based on the obtained duration of the predicted failure.
16. The apparatus of claim 15 , wherein the duration prediction model is executed on the streaming data continuously.
17. The apparatus of claim 16 , wherein the failure prediction model is continuously executed on the streaming data, wherein the processor is configured to deactivate the failure prediction model when the failure prediction occurs by discarding output from the failure prediction model until the failure prediction model is reactivated.
18. The apparatus of claim 15 , wherein the processor is configured to determine the time to reactivate the failure prediction model by:
determining an end time of the predicted failure; and
determining the time to reactivate the failure prediction model within an alert window time period before the end time of the predicted failure.
19. The apparatus of claim 15 , wherein the processor is configured to determine the time to reactivate the failure prediction model by:
executing a modified failure prediction model on the streaming data configured to detect a successive failure after the predicted failure; and
determining the time to reactivate the failure prediction model as when the predicted failure ends.
20. The apparatus of claim 15 , further comprising reactivating the failure prediction model for processing the stream data at the determined time.