Industrial monitoring system device dislodgement detection
An industrial monitoring system comprises a monitoring device attached to an industrial device by a bond. Sensor data collected by the monitoring device during a commissioning period is received and used to train a machine learning model. Subsequent to the commissioning period, additional sensor data is collected by the monitoring device. An abnormal state of the bond between the monitoring device and industrial device is determined based on the additional sensor data and a characteristic inferred by the trained machine learning model. A notification of the abnormal state is generated.
1. A system, comprising:
at least one processor; and
at least one memory comprising instructions that, in response to execution by the at least one processor, cause the device to at least:
receive first data indicative of a metric obtained by a monitoring device, the monitoring device attached by a bond to an industrial device, the first data obtained by the monitoring device during a commissioning period of the monitoring device;
train a neural network, based at least in part on the first data, to infer a characteristic of the metric;
receive second data indicative of the metric, the second data obtained by the monitoring device subsequent to the commissioning period;
determine, based at least in part on the second data and the characteristic of the metric inferred by the neural network, that the bond is in an abnormal state; and
generate a notification of the abnormal state of the bond.
2. The system of claim 1 , wherein the abnormal state of the bond comprises at least one of weakening of the bond between the monitoring device and the industrial device or separation of the monitoring device from the industrial device.
3. The system of claim 1 , the at least one memory comprising further instructions that, in response to execution by the at least one processor, cause the system to at least train the neural network to infer a state of the bond.
4. The system of claim 1 , wherein the inferred characteristic is a rate of temperature change and the abnormal state of the bond is determined by comparison of the inferred rate of temperature change to an observed rate of temperature change indicated by the second data.
5. The system of claim 1 , wherein the inferred characteristic is a frequency spectrum of the metric and the abnormal state of the bond is determined by comparison of the inferred frequency spectrum with an observed frequency spectrum indicated by the second data.
6. A method, comprising:
receiving first data indicative of a metric obtained by a monitoring device, the monitoring device attached by a bond to an industrial device, the first data obtained by the monitoring device during a first time period;
training a machine learning model, based at least in part on the first data, to infer a characteristic of the metric;
receiving second data indicative of the metric, the second data obtained by the monitoring device during a second time period;
determining, based at least in part on the second data and the characteristic of the metric inferred by the machine learning model, that a state of the bond has changed; and
generating a notification of the changed state of the bond.
7. The method of claim 6 , wherein the changed state of the bond comprises at least one of weakening of the bond between the monitoring device and the industrial device or separation of the monitoring device from the industrial device.
8. The method of claim 6 , wherein the first data exhibits periodic fluctuations in values of the metric.
9. The method of claim 6 , wherein the inferred characteristic is a predicted rate of temperature change after a shutoff of the industrial device.
10. The method of claim 9 , further comprising:
identifying the changed state of the bond by at least comparing the inferred rate of temperature change to an observed rate of temperature change indicated by the second data.
11. The method of claim 6 , further comprising:
identifying the changed state of the bond by at least comparing a signal characteristic predicted by the machine learning model with an observed signal characteristic indicated by the second data.
12. The method of claim 6 , further comprising:
training the machine learning model to infer the state of the bond.
13. The method of claim 6 , further comprising:
sending instructions to reattach the monitoring device to the industrial device; and
causing the monitoring device to enter a commissioning period subsequent to reattachment.
14. A non-transitory computer-readable storage medium comprising instructions that, in response to execution by at least one processor of a computing device, cause the computing device to at least:
obtain first data indicative of a metric obtained by a monitoring device, the monitoring device attached by a bond to an industrial device, the first data obtained by the monitoring device during a first time period;
obtain second data indicative of the metric, the second data obtained by the monitoring device during a second time period;
determine, based at least in part on the second data and an inferred characteristic of the metric, that a state of the bond has changed, wherein the inferred characteristic is inferred using a machine learning model trained using the first data; and
send a notification of the changed state of the bond.
15. The non-transitory computer-readable storage medium of claim 14 , wherein the changed state of the bond comprises at least one of weakening of the bond between the monitoring device and the industrial device or separation of the monitoring device from the industrial device.
16. The non-transitory computer-readable storage medium of claim 14 , wherein the inferred characteristic is a rate of temperature change.
17. The non-transitory computer-readable storage medium of claim 16 , comprising further instructions that, in response to execution by the at least one processor, cause the computing device to at least:
identify the changed state of the bond by at least comparing the inferred rate of temperature change with an observed rate of temperature change indicated by the second data.
18. The non-transitory computer-readable storage medium of claim 14 , wherein the inferred characteristic is a frequency spectrum of the metric.
19. The non-transitory computer-readable storage medium of claim 18 , comprising further instructions that, in response to execution by the at least one processor, cause the computing device to at least:
identify the changed state of the bond by at least comparing the inferred frequency spectrum with an observed frequency spectrum indicated by the second data.
20. The non-transitory computer-readable storage medium of claim 14 , comprising further instructions that, in response to execution by the at least one processor, cause the computing device to at least:
send instructions to reattach the monitoring device to the industrial device, wherein the monitoring device is placed into a commissioning period subsequent to reattachment.