Machine health monitoring, failure detection and prediction using non-parametric data
According to some embodiments, system and methods are provided, comprising receiving, at a machine health module, non-parametric data associated with operation of an installed product; generating, via the machine health module, a health status for at least one of a failure type and a remaining useful life of the installed product, based on the received non-parametric data; and generating an operating response of the installed product based on the generated health status. Numerous other aspects are provided.
1. A computer implemented method comprising:
receiving, at a machine health module, non-parametric data associated with operation of an installed product;
generating, via the machine health module, at least one of a diagnosis for a failure type and a prognosis of a remaining useful life of the installed product, based on the received non-parametric data; and
operating the installed product based on the generated at least one of the diagnosis and the prognosis.
2. The method of claim 1 , wherein the non-parametric data is at least one of an operator log, a machine calibration log, a machine software log, an event history data log, and a maintenance and repair log.
3. The method of claim 1 , wherein the at least one of the diagnosis and the prognosis for the installed product is generated based on non-parametric data associated with two or more installed products.
4. The method of claim 1 , wherein the non-parametric data is any data beyond machine sensor data which cannot be stored in relational databases.
5. The method of claim 1 , further comprising:
comparing the generated at least one diagnosis and prognosis to a second at least one diagnosis and prognosis based on parametric data; and
generating a confidence level for the generated at least one diagnosis and prognosis and the second at least one diagnosis and prognosis based on the comparison.
6. The method of claim 5 , wherein the generated at least one diagnosis and prognosis and the second at least one diagnosis and prognosis are in a same format prior to the comparison.
7. The method of claim 1 , wherein generating the at least one diagnosis and prognosis further comprises:
generating, via the machine health module, the diagnosis for the installed product based on the received non-parametric data.
8. The method of claim 7 , wherein the diagnosis is generated based on:
detecting at least one anomaly in the non-parametric data.
9. The method of claim 1 , wherein generating the at least one diagnosis and prognosis further comprises:
generating the prognosis for the installed product based on the received non-parametric data.
10. The method of claim 1 , wherein generating the at least one diagnosis and prognosis further comprises:
pre-processing the received non-parametric data.
11. The method of claim 10 , further comprising:
applying feature engineering to the pre-processing data.
12. The method of claim 11 , wherein feature engineering further comprises:
extracting one or more features from the pre-processed data; and
determining one or more patterns to associate with each extracted feature.
13. A system comprising:
a machine health module;
a memory storing processor-executable process steps; and
a non-parametric data processor coupled to the memory, and in communication with the machine health module and operative to execute the processor-executable steps to cause the system to:
receive, at a machine health module, non-parametric data associated with operation of an installed product;
generate, via the machine health module, at least one of a diagnosis for a failure type and a prognosis of a remaining useful life of the installed product, based on the received non-parametric data; and
operate the installed product based on the generated at least one of the diagnosis and the prognosis.
14. The system of claim 13 , wherein the non-parametric data is at least one of an operator log, a machine calibration log, a machine software log, an event history data log, and a maintenance and repair log.
15. The system of claim 13 , wherein the at least one of the diagnosis and the prognosis for the installed product is generated based on non-parametric data associated with two or more installed products.
16. The system of claim 13 , wherein the non-parametric data is any data beyond machine sensor data which cannot be stored in relational databases.
17. The system of claim 13 , wherein generating the at least one of the diagnosis and the prognosis further comprises processor-executable steps to cause the system to:
pre-process the received non-parametric data.
18. The system of claim 17 , further comprising processor-executable steps to cause the system to:
apply feature engineering to the pre-processed data.
19. The system of claim 18 , wherein the application of feature engineering further comprises processor-executable steps to cause the system to:
extract one or more features from the pre-processed data; and
determine one or more patterns to associate with each extracted feature.
20. A non-transitory computer-readable medium storing program code, the program code executable by a computer system to cause the computer system to:
receive, at a machine health module, non-parametric data associated with operation of an installed product;
generate, via the machine health module, at least one of a diagnosis for a failure type and a prognosis of a remaining useful life of the installed product, based on the received non-parametric data; and
operating the installed product based on the generated at least one of the diagnosis and the prognosis.
21. The medium claim 20 , wherein the non-parametric data is at least one of an operator log, a machine calibration log, a machine software log, an event history data log, and a maintenance and repair log.
22. The medium of claim 21 , wherein the non-parametric data is any data beyond machine sensor data which cannot be stored in relational databases.