Diagnosis device, diagnosis system, diagnosis method, and computer-readable medium

1. A diagnosis device comprising:

a first acquiring unit configured to acquire, from a target device, a piece of context information corresponding to running operation of the target device among pieces of context information determined based on different kinds of operations of the target device;

a second acquiring unit configured to acquire detection information output from a detecting unit that detects a physical quantity that changes depending on operation of the target device;

an extracting unit configured to extract, from the detection information acquired by the second acquiring unit, feature information indicating a feature of a piece of detection information in an interval that includes a specific operation interval of the target device, the specific operation interval being indicated by the piece of context information;

a selecting unit configured to select a piece of reference feature information used as reference based on the feature information, and sequentially select pieces of target feature information to be compared with the piece of reference feature information;

a first calculating unit configured to calculate a likelihood of a process interval for a specific process of the target device based on a comparison between the piece of reference feature information and each of the pieces of target feature information selected by the selecting unit;

a first determining unit configured to determine whether a piece of target feature information corresponding to the likelihood of the process interval is included in the process interval for the specific process, based on the likelihood of the process interval; and

an estimating unit configured to estimate the process interval based on a determination result obtained by the first determining unit.

2. The diagnosis device according to claim 1 , wherein the selecting unit selects, as the piece of reference feature information, a piece of feature information in an interval that is determined as a non-process interval being an interval other than the process interval based on the operation interval.

3. The diagnosis device according to claim 1 , wherein the selecting unit selects, as a new piece of reference feature information, an average of a predetermined number of pieces of information that have been selected as pieces of the reference feature information in the past.

4. The diagnosis device according to claim 1 , wherein the selecting unit selects, as the piece of reference feature information, a piece of information that is generated through machine learning performed on the feature information.

5. The diagnosis device according to claim 1 , wherein

the extracting unit extracts, as the feature information, a frequency spectrum of the piece of detection information in the interval including the operation interval, and

the first calculating unit calculates the likelihood of the process interval based on a difference between a frequency spectrum serving as the piece of reference feature information and a frequency spectrum serving as each of the pieces of the target feature information.

6. The diagnosis device according to claim 1 , wherein when the likelihood of the process interval is greater than a predetermined threshold, the first determining unit determines that the piece of target feature information corresponding to the likelihood of the process interval is included in the process interval for the specific process.

7. The diagnosis device according to claim 1 , further comprising a second determining unit configured to determine presence or absence of an abnormality in a tool serving as the target device in the operation interval, as a result of a comparison between a length of the process interval estimated by the estimating unit and a length of a predetermined interval.

8. The diagnosis device according to claim 7 , wherein the second determining unit determines that the abnormality has occurred in the tool when the length of the process interval estimated by the estimating unit is shorter than the length of the predetermined interval by a predetermined value or more.

9. The diagnosis device according to claim 7 , wherein the predetermined interval is calculated based on a predetermined number of intervals that have been estimated as normal processing intervals in the past by the estimating unit.

10. The diagnosis device according to claim 7 , wherein the predetermined interval is a process interval that was previously estimated by the estimating unit.

11. The diagnosis device according to claim 7 , wherein the predetermined interval is the operation interval.

12. The diagnosis device according to claim 1 , further comprising a second calculating unit configured to calculate a score indicating a change in the feature information corresponding to the process interval estimated by the estimating unit.

13. The diagnosis device according to claim 12 , further comprising a display control unit configured to chronologically display the score calculated by the second calculating unit.

14. The diagnosis device according to claim 1 , further comprising a third determining unit configured to determine whether operation performed by the target device is normal, by using a model that corresponds to the piece of context information acquired by the first acquiring unit among one or more models corresponding to one or more pieces of context information, and by using feature information that corresponds to the piece of context information and corresponds to the process interval estimated by the estimating unit.

15. The diagnosis device according to claim 14 , further comprising a generating unit configured to generate a model from a piece of feature information corresponding to the process interval estimated by the estimating unit among whole feature information corresponding to the piece of context information acquired by the first acquiring unit.

16. A diagnosis system comprising:

the detecting unit, and

the diagnosis device according to claim 1 .

17. A diagnosis method comprising:

acquiring, from a target device, a piece of context information corresponding to running operation of the target device among pieces of context information determined based on different kinds of operations of the target device;

acquiring detection information output from a detecting unit that detects a physical quantity that changes depending on operation of the target device;

extracting, from the detection information acquired at the acquiring, feature information indicating a feature of a piece of detection information in an interval that includes a specific operation interval of the target device, the specific operation interval being indicated by the piece of context information;

selecting a piece of reference feature information used as reference based on the feature information, and sequentially selects pieces of target feature information to be compared with the piece of reference feature information;

calculating a likelihood of a process interval for a specific process of the target device based on a comparison between the piece of reference feature information and each of the pieces of target feature information selected at the selecting;

determining whether a piece of target feature information corresponding to the likelihood of the process interval is included in the process interval for the specific process, based on the likelihood of the process interval; and

estimating the process interval based on a determination result obtained at the determining.

18. A non-transitory computer-readable medium including programmed instructions that cause a computer to execute:

acquiring, from a target device, a piece of context information corresponding to running operation of the target device among pieces of context information determined based on different kinds of operations of the target device;

acquiring detection information output from a detecting unit that detects a physical quantity that changes depending on operation of the target device;

extracting, from the detection information acquired at the acquiring, feature information indicating a feature of a piece of detection information in an interval that includes a specific operation interval of the target device, the specific operation interval being indicated by the piece of context information;

selecting a piece of reference feature information used as reference based on the feature information, and sequentially selects pieces of target feature information to be compared with the piece of reference feature information;

calculating a likelihood of a process interval for a specific process of the target device based on a comparison between the piece of reference feature information and each of the pieces of target feature information selected at the selecting;

determining whether a piece of target feature information corresponding to the likelihood of the process interval is included in the process interval for the specific process, based on the likelihood of the process interval; and

estimating the process interval based on a determination result obtained at the determining.