License plate generation device, diagnostic model generation device, degradation diagnostic device, information processing system and license plate generation method

Abstract

Comprehensive: a training data acquisition unit (11) that acquires training data in which a feature and data relating to an external factor are associated with each other; a partitioning unit (12) for a region of the external factor that classifies a range of values ​​of the data relating to the external factor in the axial direction of the data relating to the external factor into a constant band and a transition band;and an external factor indicator data generation unit (13) which, with respect to the axis of the data relating to the external factor, generates external factor indicator data in which value ranges of the external data are associated with external factor indicators by assigning an external factor indicator, indicating that the range is a range in which the relationship between the data relating to the external factor in the range and the corresponding feature is constant, to the value range of data relating to the external factor which is classified as the constant band, and an external factor indicator, indicating that the range is a range in which the relationship between the data relating to the external factor in the range and the corresponding feature changes frequently, to the value range of data relating to the external factor which is classified as the transition band.

Description

TECHNICAL AREA

[0001] The present disclosure relates to a tag generation device, a diagnostic model generation device, a degradation diagnostic device, an information processing system and a tag generation method that aid in the construction of a diagnostic model to be used in degradation diagnosis on a device or apparatus. STATE OF THE ART

[0002] There is a known technique for performing a degradation diagnosis on a device using a model (hereinafter referred to as the "diagnostic model") that was created based on a feature extracted from data obtained from the device.

[0003] As a technique for performing a diagnosis of an operating state of a device using a diagnostic model, patent literature 1 discloses a diagnostic system that manages devices with similar operating conditions as a group under a common device feature pattern number and performs a diagnosis of a device state that reflects the feature of the device using a diagnostic pattern suitable for the device feature pattern number. REFERENCE LIST PATENT LITERATURE

[0004] Patent Literature 1: JP 2016-91378 A SUMMARY OF THE INVENTIONAL PROBLEM

[0005] A feature used in the degradation diagnosis of a device changes with the operating environment, in addition to the deteriorating condition of the device itself. Therefore, the diagnostic model to be used for degradation diagnosis of the device must be designed as a model capable of diagnosing a normal state as well as a degraded or deteriorated state, taking into account the influence of the operating environment.

[0006] However, the operating environment index likely has continuous values, and there is a possibility that a highly accurate diagnostic result cannot be achieved even if one of the diagnostic models is used before and after the switch at the boundary of switching between the diagnostic models depending on the operating environment.

[0007] The conventional technique has the problem that it is difficult to create a diagnostic model that takes this point into account. Consequently, there is a possibility that the accuracy of the degradation diagnosis result, which is performed on the device with the diagnostic model based on this feature, will decrease.

[0008] It should be noted that, with regard to the operating state of the device, the possibility of the index exhibiting continuous values ​​is low. That is, there is a low probability that the accuracy of the diagnostic result will decrease at the boundary of switching between diagnostic models, depending on the operating state of the device. In a technique for performing a diagnosis of a device state using the diagnostic pattern corresponding to the device feature pattern in the diagnostic system, as disclosed in patent literature 1, the influence of the operating environment on the diagnostic result is not taken into account, and the problem described above has not yet been solved.

[0009] The present disclosure was carried out to solve the problems described above and aims to provide a tagging device capable of supplying data that helps in the construction of a diagnostic model that prevents a decrease in the accuracy of degradation diagnosis on a device due to an operating environment. SOLUTION TO THE TASK

[0010] A feature generation device according to the present disclosure comprises: a training data acquisition unit for acquiring training data in which a feature to be used for generating a diagnostic model, which is used as a degradation diagnosis object in a degradation diagnosis on a target device, is assigned data relating to the external factor relating to an operating environment;a partitioning unit for a region of an external factor, in order to classify, on the basis of the training data procured by the training data acquisition unit, a range of values ​​of the data relating to the external factor in an axial direction of the data relating to the external factor into a constant band, which is a range of values ​​of the data relating to the external factor in which a relationship to the corresponding characteristic is considered constant, and a transition band, which is a range of values ​​of the data relating to the external factor in which the relationship to the corresponding characteristic is considered to change frequently;and a tag data generation unit for the external factor to assign to the range of values ​​of the data relating to the external factor, which is classified as a constant band, a tag of the external factor indicating that a relationship between the data relating to the external factor in the range and the corresponding feature is constant, with respect to an axis of the data relating to the external factor, to assign to the range of values ​​of the data relating to the external factor, classified as a transition band, a tag of the external factor indicating that the relationship between the data relating to the external factor in the range and the corresponding feature changes frequently, and to generate tag data for the external factor in which ranges of values ​​of the data relating to the external factor are associated with tags for the external factor. ADVANTAGEOUS EFFECTS OF THE INVENTION

[0011] According to the present disclosure, with a configuration as described above, it is possible to provide data that aids in the construction of a diagnostic model that prevents a decrease in the accuracy of degradation diagnostics on a device due to an operating environment. As a result, it is possible to construct a diagnostic model that prevents a reduction in the accuracy of the degradation diagnostic result performed on the device using the diagnostic model based on the feature, and it is possible to achieve high-accuracy degradation diagnostics using the diagnostic model. BRIEF DESCRIPTION OF THE DRAWINGS Fig. Figure 1 is a representation showing an example configuration of an information processing system with a license plate generation device according to a first embodiment. Fig. Figure 2 is a representation showing an example configuration of a diagnostic model generation device according to the first embodiment, including the license plate generation device according to the first embodiment. Fig. Figure 3 is a diagram illustrating a method implemented by a partitioning unit for a region of an external factor to divide the range of values ​​of the data relating to the external factor in the axial direction of the data relating to the external factor into a constant band and a transition band in the first embodiment. Fig. Figure 4 is a diagram illustrating the concept of an example of external factor identifiers that are assigned to the external factor data axis by an external factor identifier data generation unit in the first embodiment. Fig. Figure 5 is a representation to illustrate the concept of an example of the content of the identifier data for the external factor, which is generated by the identifier data generation unit for the external factor in the first embodiment. Fig. Figure 6 is a representation showing an example configuration of a degradation diagnostic device according to the first embodiment. Fig. Figure 7 is a diagram illustrating a process example in which a diagnostic unit determines a diagnostic model that corresponds to a feature extracted by a feature extraction unit from among the diagnostic models issued by a model procurement unit in the first embodiment. Fig. Figure 8 is a flowchart to illustrate the processes to be performed by the diagnostic model generation device according to the first embodiment, including the identifier generation device according to the first embodiment. Fig. Figure 9 is a flowchart to explain details of a region allocation process, which is initiated by the allocation unit for a region of an external factor in step ST2 in Fig. 8 is to be carried out. Fig. Figure 10 is a diagram used to illustrate a problem with a conventional technique. Fig. Figure 11 is a flowchart to explain the operation of the degradation diagnostic device according to the first embodiment. Fig. Figure 12 is a diagram illustrating a concept of an example of external factor identifiers that are assigned to the external factor data axis by the external factor identifier data generation unit in the identifier generation device in a case where the identifier generation device generates external factor data for each set of "temperature" and "amplitude value" and one set of "temperature" and "variance" in the first embodiment. Fig. 13A and Fig. Figures 13B are examples of the hardware configuration of the license plate generation device according to the first embodiment. DESCRIPTION OF THE EXAMPLES OF EXECUTION

[0012] In order to explain the present disclosure in more detail, exemplary embodiments of the disclosure are described below with reference to the attached drawings. First embodiment

[0013] A tag generation device according to a first embodiment provides data (hereinafter referred to as "tag data for external factors") that helps in creating a trained model (hereinafter referred to as "diagnostic model") that is used, for example, in the degradation diagnosis of a tag for the external factor (hereinafter referred to as "target device") that is to be subjected to degradation diagnosis.

[0014] The target device is, for example, a platform screen door. It should be noted that this is merely an example and that any device whose degradation is related to the environment or operating environment in which the target device operates can be a target device. Degradation diagnosis on a device is performed using a diagnostic model based on a feature extracted from data collected by the device. The function is influenced by the operating environment. Therefore, if the diagnostic model is not generated under the assumption that the feature is influenced by the operating environment, there is a possibility that the device will be incorrectly diagnosed as degraded due to a change in the feature caused by a change in the operating environment, even though it is not degraded.The identification device according to the first embodiment prevents such a misdiagnosis. More precisely, the identification device according to the first embodiment provides identification data for the external factor, which aids in the creation of a diagnostic model capable of preventing misdiagnoses as described above.

[0015] In the first embodiment, it is assumed that the operating environment is a natural factor that cannot be controlled by humans, such as temperature or precipitation.

[0016] In the first embodiment described below, the operating environment that can influence the deterioration or degradation of the target device described above is also referred to as the "external factor".

[0017] The tag data for the external factor to be supplied by the tag generation device is data that specifies the relationship between the feature and the external factor, which is to be used in degradation diagnosis on the target device.

[0018] If the relationship between the feature and the external factor to be used in the degradation diagnosis on the target device changes, it is assumed that the diagnostic model for the precise execution of the degradation diagnosis on the target device will also change.

[0019] As described above, the operating environment index likely has continuous values, and there is a possibility that a highly accurate diagnostic result cannot be achieved even if one of the diagnostic models is used before and after switching at the boundary between diagnostic models, depending on the operating environment. The external factor's characteristic data indicates that the relationship between the characteristic and the external factor is one where there is a possibility that an accurate diagnostic result cannot be achieved, even if one of the diagnostic models is used.

[0020] Characteristic data for the external factor can contribute to the generation of an accurate diagnostic model. This means that when diagnosing degradation on the target device, an accurate diagnostic result can be expected using a diagnostic model generated based on the characteristic data of the external factor.

[0021] Fig. Figure 1 is a representation showing an example configuration of an information processing system 6 with a license plate generation device 1 according to the first embodiment.

[0022] The information processing system 6 comprises the identifier generation device 1, a diagnostic model generation device 2, a degradation diagnostic device 3, and a storage device 4. It should be noted that in the first embodiment, the storage device 4 is integrated into the information processing system 6, but this is only an example. The information processing system 6 does not necessarily include the storage device 4. The storage device 4 may, for example, be contained in a system that is located outside of the information processing system 6 and is connected to the information processing system 6.

[0023] The identifier generation device 1, the diagnostic model generation device 2 and the degradation diagnostic device 3 are, for example, present in the target device.

[0024] In the first embodiment, the license plate generation device 1 is integrated into the diagnostic model generation device 2.

[0025] The identifier generation device 1 generates identifier data for the external factor based on the training data stored in a training data storage unit 41 of the storage device 4. The training data consists of data to which the features relating to the external factor are assigned. The training data will be described in detail later.

[0026] The license plate generation device 1 stores the generated license plate data for the external factor in a storage unit 42 of the license plate data for the external factor of the storage device 4.

[0027] The diagnostic model generation device 2 acquires the training data to be used in generating a diagnostic model from the training data storage unit 41 of the storage device 4, based on the external factor identifier data stored in the external factor identifier data storage unit 42, by the identifier generation device 1, and generates a diagnostic model using the acquired training data. The diagnostic model generation device 2 stores the generated diagnostic model in a diagnostic model storage unit 43 of the storage device 4.

[0028] The license plate generation device 1 and the diagnostic model generation device 2 will be described in detail later.

[0029] The degradation diagnostic device 3 is connected to a sensor 5. The sensor 5 is located, for example, in the target device, collects data relating to the operating state of the target device (this data is hereinafter referred to as "sensor data") and data relating to external factors, and outputs the sensor data and the data relating to external factors to the degradation diagnostic device 3. The sensor 5 is a vibration sensor, a temperature sensor, a precipitation sensor, or similar.

[0030] The sensor data and the data of the external factors are time series data of sensor measurements over a predetermined period, which are determined, for example, by sensor 5 at predetermined intervals.

[0031] The sensor data and the data from external factors each indicate a value measured by the sensor for vibration, temperature, precipitation, or similar. It should be noted that this is only an example, and the sensor data may contain a control value, such as a command value or a reference value over a predetermined period, as determined, for example, by a plurality of sensors at predetermined intervals.

[0032] Although in Fig. For simplicity, only one sensor 5 is shown in Figure 1; however, there can be multiple sensors 5. The sensor data and the data from external factors can be acquired by different sensors 5. The degradation diagnostic device 3 can be connected to multiple sensors 5.

[0033] The degradation diagnostic device 3 acquires diagnostic data, including sensor data and data on external factors, from sensor 5. Based on the data concerning the external factors, the degradation diagnostic device 3 determines the diagnostic model to be used for degradation diagnosis on the target device and performs the degradation diagnosis on the target device using the determined diagnostic model and the sensor data.

[0034] In the first embodiment, which is described below as an example, the degradation diagnostic device 3 obtains vibration data relating to the vibration of the target device from the sensor 5 as sensor data and temperature data as data relating to the external factors.

[0035] The degradation diagnostic device 3 outputs the result of the degradation diagnosis to an output device (not shown in the drawing). The output device is a display device or a speech output device contained in a personal computer (PC) installed in an administrative room or similar location where an operator or the like performs an administrative operation on the target device. The output device can be a display device or a speech output device, for example, contained in a tablet terminal worn by an operator.

[0036] The degradation diagnostic device 3 will be described in detail later.

[0037] The storage device 4 is formed with a hard disk drive (HDD) or a solid-state drive (SSD) and includes the training data storage unit 41, the external factor identifier data storage unit 42, and the diagnostic model storage unit 43, which have already been mentioned above.

[0038] The training data storage unit 41 stores training data.

[0039] The storage unit 42 for the data relating to the external factor stores the data relating to the external factor generated by the license plate generation device 1.

[0040] The diagnostic model storage unit 43 stores the diagnostic model generated by the diagnostic model generation device 2.

[0041] The storage device 4 can be located in the target device or at a location outside the target device that can be accessed by the tagging device 1, the diagnostic modeling device 2 and the degradation diagnostic device 3.

[0042] The identification device 1, the diagnostic model generation device 2 and the degradation diagnostic device 3 according to the first embodiment are now described in detail.

[0043] First, the diagnostic model generation device 2 according to the first embodiment, comprising the license plate generation device 1 according to the first embodiment, is described.

[0044] Fig. Figure 2 is a representation showing the diagnostic model generation device 2 according to the first embodiment, including the identifier generation device 1 according to the first embodiment.

[0045] The diagnostic model generation device 2 comprises the label generation device 1 and a diagnostic model generation unit 21.

[0046] The tag generation device 1 comprises a training data acquisition unit 11, a partitioning unit 12 for a region of an external factor and a tag data generation unit for the external factor 13.

[0047] The allocation unit 12 for a region of an external factor includes an inflection point extraction unit 121, an adjustment unit 122 for a preliminary region, a generation unit 123 for a relational model, an evaluation unit 124 for the relational model, and a region determination unit 125.

[0048] The training data acquisition unit 11 refers to the training data storage unit 41 and acquires the training data.

[0049] The training data is generated in advance and stored in the training data storage unit 41.

[0050] In the first embodiment, the training data are data in which the features extracted from the sensor data acquired by sensor 5 at the time of the degradation diagnosis performed by the degradation diagnosis device 3 on the target device are linked to the operating environments of the target device, and the data relating to external factors.

[0051] For example, an administrator or similar person first acquires from the target device the features extracted from the sensor data collected by sensor 5 during a preset period, as well as the data relating to the external factor, generates the training data based on these features and the data relating to the external factor(s), and stores the training data in the training data storage unit 41. Thus, the administrator or similar person only needs to acquire the features extracted from the sensor data acquired by sensor 5 via the degradation diagnostic device 3 and the data relating to the external factor, and generate the training data based on these features and the data relating to the external factor. The extraction of the features in the degradation diagnostic device 3 is described later.Alternatively, the degradation diagnostic device 3 can, for example, have the function of generating training data based on the extracted features and storing the training data in the training data storage unit 41. It should be noted that in the first embodiment, the sensor data can be the features that constitute the training data.

[0052] The training data only needs to include features based on sensor data generated when the target device is in a normal state, or features based on sensor data generated assuming the target device is in a normal state. The training data does not necessarily include features based on sensor data generated when the target device is in an abnormal state, or features based on sensor data generated assuming the target device is in an abnormal state.

[0053] It should be noted that the features included in the training data are determined beforehand by the administrator or similar person. The administrator or similar person specifies that the features included in the training data are those relating to faulty states of the device, or, in other words, features that can serve as input for a diagnostic model.

[0054] The training data acquisition unit 11 outputs the acquired training data to the distribution unit 12 for a region of the external factor.

[0055] Based on the training data output by the training data acquisition unit 11, the partitioning unit 12 divides the range of values ​​of the data relating to the external factor for a region of the external factor in the axial direction of the data relating to the external factor into a range of values ​​of data relating to the external factor whose relationship to the corresponding features is considered constant (this range is referred to below as the "constant band"), and a range of values ​​of data of the external factor whose relationship to the corresponding features is considered to change frequently (this range is referred to below as the "transition band").

[0056] As described above, it is assumed that if the relationship between the features and the external factors in the training data changes, the diagnostic model will also change. That is, if the relationship between the features and the external factors in the training data changes frequently, the diagnostic model generated from the training data may also change frequently. In this case, it is possible that even using a frequently changing diagnostic model will not yield an accurate diagnosis. Conversely, if the relationship between the features and the external factor in the training data is constant, the diagnostic model does not change frequently. Therefore, it is assumed that the same diagnostic model can also be used for the diagnosis of degradation.This is based on the concept that, as with an external factor influencing a trait, the influence of the external factor on the trait can be corrected by the same technique or with the same parameter, unless the relationship between the trait and the external factor changes, in which case the same diagnostic model can be used.

[0057] In the first embodiment, the process to be carried out by the partitioning unit 12 for a region of the external factor to divide the value range of the data relating to the external factor in the axial direction into the constant band and the transition band is referred to as the "region partitioning process".

[0058] Here, the division into the constant band and the transition band by the division unit 12 for a region of the external factor is described in detail with reference to the drawings.

[0059] Fig. Figure 3 is a diagram illustrating a method implemented by the partitioning unit 12 for a region of the external factor to partition the range of values ​​of the data relating to the external factor in the axial direction of the data relating to the external factor into the constant band and the transition band in the first embodiment.

[0060] Fig. Figure 3 shows the relationship between a feature (here the amplitude value) based on the training data and an external factor (here the temperature), with the vertical axis representing the values ​​of the feature and the horizontal axis representing the values ​​of the external factor.

[0061] As described above, the allocation unit 12 for the region of the external factor includes the inflection point extraction unit 121, the setting unit 122 for a preliminary region, the generation unit 123 for a relational model, the evaluation unit 124 for a relational model, and the region determination unit 125.

[0062] First, the inflection point extraction unit 121 extracts a multitude of inflection points in the axial direction of the data relating to the external factor associated with the feature in the training data, based on the training data output by the training data acquisition unit 11.

[0063] In the first embodiment, an inflection point refers to a point at which the relationship between the data concerning the external factor and the feature changes in the axial direction of the data concerning the external factor.

[0064] In Fig. The inflection points extracted by the inflection point extraction unit 121 are designated as “F1”, “F2” and “F3”.

[0065] As in Fig. As shown in Figure 3, the inflection point extraction unit extracts 121 inflection points, i.e., the points (“F1” and “F3”) where the value of the feature changes from increasing to decreasing, and the point (“F2”) where the value changes from decreasing to increasing in the axial direction of the data relating to the external factor, for example.

[0066] The Inflection Point Extraction Unit 121 adds data indicating the inflection points to the training data output by the Training Data Acquisition Unit 11 and outputs the training data to which the data indicating the inflection points has been added (this training data is referred to below as "training data with additional inflection points") to the Setting Unit 122 for a preliminary region.

[0067] Next, for a preliminary region, the setting unit 122 selects adjacent inflection points (designated as a first inflection point and a second inflection point) from among the inflection points extracted by the inflection point extraction unit 121, based on the inflection point-supplemented training data output by the inflection point extraction unit 121. For example, for a preliminary region, the setting unit 122 sets "F2" as the first inflection point and "F3" as the second inflection point.

[0068] The setting unit 122 for a preliminary region defines the range of values ​​of the data relating to the external factor between the selected first inflection point and the second inflection point as the inflection point region and defines a preliminary region that is a candidate for the constant band within the inflection point region.

[0069] Setting unit 122 for a preliminary region defines the initial range of the preliminary region, which is the range (the inflection point range) between the value of the data relating to the external factor at the first inflection point and the value of the data relating to the external factor at the second inflection point. Fig. 3 is the initial region of the provisional region, designated with "P".

[0070] Setting unit 122 for a preliminary region outputs the value of the data concerning the external factor in the defined preliminary region and the value of the characteristic corresponding to the value of the data concerning the external factor to generation unit 123 for a relational model. If the initial region "P" is the preliminary region, setting unit 122 for a preliminary region outputs the value of the data concerning the external factor and the value of the characteristic from those at the first turning point "F2" to those at the second turning point "F3" to generation unit 123 for a relational model.

[0071] Next, generation unit 123 creates a relational model that specifies the relationship between the feature and the data concerning the external factor (this model is referred to below as the "relational model"), based on the values ​​of the data concerning the external factor in the preliminary region and the values ​​of the feature that correspond to the values ​​of the data concerning the external factor. The relational model is represented, for example, by a linear expression.

[0072] If, for example, the preliminary region is the initial region “P”, the generating unit 123 for a relational model generates the relational model based on the value of the data concerning the external factor and the value of the feature from those at the first turning point “F2” to those at the second turning point “F3”.

[0073] The generation unit 123 for a relational model outputs the generated relational model to the evaluation unit 124 for a relational model, together with the value of the data concerning the external factor in the preliminary region and the value of the feature that corresponds to the value of the data concerning the external factor, which is output by the setting unit 122 of the preliminary region, which is the training data in the preliminary region.

[0074] Next, the evaluation unit 124 for a relational model assesses the accuracy of the relational model using the value of the data relating to the external factor contained in the preliminary region and the value of the feature corresponding to the value of the data relating to the external factor.

[0075] For example, the relational model is assumed to be represented by the following linear expression (1). Note that in the following linear expression (1), y represents the feature and x represents the data of the external factor. y=x+3

[0076] The evaluation unit 124 for a relational model replaces the value of certain data concerning the external factor in the preliminary region (here, for example, the area between the first turning point "F2" and the second turning point "F3") and the value of the feature that corresponds to the value of the data concerning the external factor in the relational model that was generated by the generation unit 123 for a relational model.

[0077] The evaluation unit 124 for a relational model then determines the error between the value of the feature obtained by the above substitution into the relational model and the value of the substituted feature.

[0078] For example, if the data of the external factor “21°C” and a feature “25 mm” are inserted into the relational model, the resulting feature is “24 mm” and the error is “1 mm”.

[0079] The evaluation unit 124 for the relational model evaluates the relational model by comparing the calculated error with, for example, a predefined threshold (hereinafter referred to as the "evaluation threshold for the relational model"). For example, the evaluation unit 124 for a relational model determines the accuracy of the relational model to be sufficiently high if the error is equal to or less than the threshold for the evaluation of the relational model, and determines the accuracy of the relational model to be low if the error is greater than the threshold for the evaluation of the relational model.

[0080] The evaluation unit 124 for a relational model outputs the result of the evaluation of the relational model together with the value of the data concerning the external factor in the preliminary region and the value of the feature that corresponds to the value of the data concerning the external factor to the region determination unit 125.

[0081] The regional determination unit 125 then determines, based on the result of the evaluation of the relational model, which is issued by the evaluation unit for a relational model 124, whether the relational model is evaluated with a sufficiently high accuracy.

[0082] If the accuracy of the relational model is deemed sufficiently high, the region definition unit 125 classifies the preliminary region in which the accuracy of the relational model is considered high as a constant band. The region definition unit 125 then classifies any range of data values ​​relating to the external factor that differs from the determined constant band as a transition band within the inflection point region.

[0083] In a case where the accuracy of the relational model is deemed insufficient or low, the region determination unit 125 issues an instruction to the preliminary region setting unit 122 to restrict the preliminary region's scope and reset it. Based on the instruction from region determination unit 125, preliminary region setting unit 122 resets the preliminary region. Specifically, preliminary region setting unit 122 limits the range of values ​​for the external factor at both ends of the preliminary region and resets the preliminary region.

[0084] Assuming that the accuracy of the relational model is rated as sufficiently high, the region definition unit 125 determines that the preliminary region, i.e., the area between the first inflection point "F2" and the second inflection point "F3", is classified as a constant band. Since there are no regions other than the constant band within the preliminary region, or in this case, no range of values ​​in the data relating to the external factor, the region definition unit 125 does not define a transition band within the preliminary region that represents the area between the first inflection point "F2" and the second inflection point "F3".

[0085] If the accuracy of the relational model is currently deemed low, the region determination unit 125 issues an instruction to the setting unit 122 for a preliminary region to restrict the scope of the preliminary region and reconfigure it. In this case, for example, the setting unit 122 for a preliminary region sets the scope of the data relating to the external factor that is in Fig. 3 is labelled "D" (the range of values ​​of the data relating to the external factor from the point labelled "E1" to the point labelled "E2" in Fig. 3) as a new provisional region. It should be noted that the range of data relating to the external factor, which is to be set as the provisional range from both ends of the provisional region before the reset, is predetermined when an instruction to reset the provisional region is issued.

[0086] The setting unit 122 for a preliminary region outputs the value of the data of the external factor in the newly defined preliminary region and the value of the feature that corresponds to the value of the data of the external factor to the generation unit 123 for the relational model.

[0087] The generation unit 123 for a relational model regenerates the relational model and outputs the newly generated relational model to the evaluation unit 124 for a relational model. The evaluation unit 124 for a relational model re-evaluates the newly generated relational model and outputs the result of the evaluation of the re-evaluated relational model to the region determination unit 125.

[0088] The setting unit 122 for a preliminary region, the generation unit 123 for a relational model, and the evaluation unit 124 for a relational model each reset the preliminary region, regenerate the relational model, and re-evaluate the accuracy of the relational model until the region determination unit 125 determines that the preliminary region is classified as a constant band, or until the region determination unit 125 determines that the preliminary region cannot be classified as a constant band.

[0089] In a case where the preliminary region set by the setting unit 122 for a preliminary region is smaller than a preset range (hereinafter referred to as the "minimum preliminary range"), the region determination unit 125 determines that the preliminary region cannot be set as a constant band, even if the relational model is evaluated with sufficiently high accuracy, and sets the preliminary region as a transition band. The region determination unit 125 then determines to classify a range of values ​​in the data relating to the external factor, which is not the preliminary region, as a transition band in the inflection area.

[0090] If the accuracy of the relational model is sufficiently high, it can be said that a specific relationship is established between the feature and the external factor within the range of values ​​of the external factor's data for which the relational model was generated. In other words, the influence of the external factor on the feature can be easily estimated. It is assumed that the diagnostic model generated from the feature corresponding to the external factor in this range can be a diagnostic model where the probability of the accuracy of the degradation diagnosis being reduced by the external factor is low when the degradation diagnosis is performed on the target device using this diagnostic model. That is, it is assumed that the same diagnostic model can be used for the degradation diagnosis of the feature corresponding to the external factor in this range.

[0091] On the other hand, it can be said that in a case where the accuracy of the relational model is low, no definite relationship is established between the feature and the external factor within the range of values ​​of the external factor's data for which the relational model was generated. That is, it is estimated that the relationship between the feature and the external factor changes frequently. In other words, it is assumed that it is difficult to estimate the influence of the external factor on the feature. It is assumed that the diagnostic model generated from training data containing the feature corresponding to the external factor within this range is a diagnostic model that offers the possibility of reducing the accuracy of the degradation diagnosis due to the external factor in a case where the degradation diagnosis is performed on the target device using the diagnostic model.Because if the relationship between the feature and the external factor affecting the feature changes frequently, the influence of the external factor on the feature cannot be corrected using the same technique or parameters, and the diagnostic model for performing a degradation diagnosis with high accuracy is likely to change frequently.

[0092] Even if the accuracy of the relational model is sufficiently high, the data for the external factors for which the relational model was generated may be insufficient to assess the relationship between the feature and the external factor if the range of data for which the relational model was generated is smaller than the minimum preliminary region. It is doubtful whether the diagnostic model generated from the training data, which includes the feature corresponding to the external factor within this region, has a low probability of reducing the accuracy of the degradation diagnosis due to the external factor in a case where the degradation diagnosis is performed on the target device using the diagnostic model.Therefore, in this case, the region determination unit 125 determines that it is assumed there is no area that can be classified as a constant band in the inflection point region, or that the diagnostic model generated from training data containing the feature corresponding to the external factor in this inflection point region is a diagnostic model that has a way of reducing the accuracy of the degradation diagnosis due to the external factor in a case where the degradation diagnosis is performed on the target device using the diagnostic model as described above.

[0093] By extracting the inflection points, setting the preliminary region, generating the relational model, evaluating the relational model, and classifying the constant band and the transition band, as described above, between all inflection points, the partitioning unit 12 classifies for a region of the external factor the constant band and the transition band in the range of values ​​of all data relating to the external factor in the axial direction of the data relating to the external factor.

[0094] The partitioning unit 12 for a region of the external factor outputs data (hereinafter referred to as "region setting data") to the indicator data generation unit 13 for an external factor, relating to the constant band or the transition band classified in the value range of the data of the external factor in the axial direction of the data of the external factor.

[0095] The region setting data is data in which data specifying the value range of the external factor data and data specifying whether the range is the constant band or the transition band are mapped to each other.

[0096] Based on the region setting data output by the distribution unit 12 for a region of the external factor, the indicator data generation unit 13 assigns, with respect to the axis of the data relating to the external factor, an external factor indicator that specifies the range in which the relationship to the characteristic corresponding to the data relating to the external factor in the range is constant, or that indicates that the range is the constant band, to the range of values ​​of the data relating to the external factor that are classified as the constant band, and assigns to the range of values ​​of the data relating to the external factor that are classified as the transition band an external factor indicator that specifies the range in which the relationship to the corresponding characteristic is estimated to change frequently, or that indicates that the range is the transition band.

[0097] Fig. Figure 4 is a diagram illustrating the concept of an example of external factor identifiers assigned to the external factor data axis by the identifier data generation unit 13 for an external factor in the first embodiment.

[0098] In Fig. 4 With respect to an external factor (here the temperature), the area below 10 °C is classified as a constant band, the area from 10 °C or more to below 17 °C as a transition band, and the area from 17 °C or more to below 28 °C as a constant band by the division unit 12 for a region of the external factor.

[0099] In this case, the tag data generation unit 13 assigns a tag for an external factor with respect to the axis of the external factor's data, indicating that the range below 10°C is a constant band (where the relationship between the feature and the external factor's data is constant) and is a range where it is estimated that the same diagnostic model (hereafter referred to as the "first diagnostic model") can be used. Fig. 4 is the identifier for the external factor, represented as the "first model application region".

[0100] Furthermore, the tag data generation unit 13 assigns a tag to an external factor indicating that the range from 10 °C or more to less than 17 °C in the axial direction of the external factor data is a transition range or a range in which the relationship between the corresponding feature and the external factor data is expected to change frequently. Fig. 4 shows the external factor being labelled as “not an applicable model”.

[0101] In this case, the tag data generation unit 13 assigns a tag to an external factor with respect to the axis of the external factor's data, indicating that the range from 17 °C or more to less than 28 °C is a constant band (in which the relationship between the feature and the external factor's data is constant) and is a range in which it is estimated that the same diagnostic model (hereafter referred to as the "second diagnostic model"), which differs from the first diagnostic model, can be used. Fig. 4 is the characteristic of the external factor represented as the “second model application region”.

[0102] As in Fig. As shown in Figure 4, the external factor tag data generation unit 13 assigns external factor tags to the respective constant bands in a case where there are multiple constant bands, so that it is clear that training data, which is assumed to have been obtained in the operating environment specified by the external factor data in the range, are estimated in order to be able to use different diagnostic models.

[0103] After the assignment of the tags for the external factor, the tag data generation unit 13 for the external factor generates data relating to the tags for the external factor (this data is referred to below as "tag data for the external factor").

[0104] The indicator data for the external factor consists of data in which the value ranges of the data relating to the external factor are linked to the indicators for the external factor. The indicator data for the external factor can also include data in which, in addition to the value ranges of the external factor data and the indicators for the external factors, data specifying the type of the external factor (here, temperature) and data specifying the type of the characteristic (here, the amplitude value) are linked together.

[0105] Fig. Figure 5 is a representation to illustrate the concept of an example of the content of the identifier data for the external factor, which is generated by the identifier data generation unit 13 for the external factor in the first embodiment.

[0106] The in Fig. The 5 shown indicator data for the external factor are indicator data for the external factor that are generated after the indicator data generation unit 13 for external factors, as shown in Fig. 4 is shown, which has assigned a marker for the external factor.

[0107] It should be noted that in the Fig. In the 5 displayed external factor indicators, indicator "11" is an external factor indicating that the area is a constant band (where the relationship between the trait and the external factor data is constant) and an area where it is estimated that the same first diagnostic model can be used. Meanwhile, external factor indicator "12" indicates that the area is a constant band (where the relationship between the trait and the external factor data is constant) and an area where the same second diagnostic model can be used as an estimate. Furthermore, external factor indicator "00" indicates that the area is a transition band, and in other words, an area where the relationship between the trait and the data relating to the external factor is expected to change frequently.

[0108] The external factor identification data generation unit 13 stores the generated identification data for the external factor in the external factor identification data storage unit 42 of storage device 4.

[0109] Based on the external factor identifier data generated by the external factor identifier data generation unit 13, the diagnostic model generation unit 21 retrieves from the training data storage unit 41 of storage device 4 the training data corresponding to the value range of the external factor data associated with the external factor identifier and indicates that the range is a region in which the relationship between the external factor data in the region and the feature is estimated to be constant. Assuming that the external factor identifier data are identifier data for the external factor whose content is as described in Fig. As shown in Figure 5, the diagnostic model generation unit 21 acquires training data from the training data storage unit 41, to which the data relating to the external factor specifying the range of less than 10°C have been added, as the training data for generating the first diagnostic model. Furthermore, the diagnostic model generation unit 21 acquires training data from the training data storage unit 41, to which the data relating to the external factor specifying the range of 17°C or more to less than 28°C have been added, as the training data for generating the second diagnostic model.

[0110] In this way, in a case where there are value ranges of the data relating to the external factor, to which different identifiers for the external factor are assigned, the diagnostic model generation unit 21 acquires training data for each value range of the data relating to the external factor.

[0111] The diagnostic model generation unit 21 generates a diagnostic model using the acquired training data.

[0112] In the example above, the diagnostic model generation unit 21 generates the first diagnostic model using the training data to which the data relating to the external factor, specifying the range of less than 10 °C, is added, and generates the second diagnostic model using the training data to which the data relating to the external factor, specifying the range of 17 °C or more to less than 28 °C, is added.

[0113] In the first embodiment, a diagnostic model is assumed to be a model that receives a feature as input and outputs a value indicating the state of the target device. The value indicating the state of the target device is a value that specifies the degree of degradation of the target device.

[0114] The diagnostic model generation unit 21 can create a diagnostic model using a known technique such as supervised or unsupervised learning.

[0115] It should be noted that the diagnostic model generation unit 21 does not generate a diagnostic model from the training data that corresponds to the value range of the data concerning the external factor, which is associated with an external factor indicator that indicates that it is a range in which the relationship between the data concerning the external factor in the range and the feature is expected to change frequently.

[0116] A reliability determination unit 211 contained in the diagnostic model generation unit 21 determines the reliability of the diagnostic model generated by the diagnostic model generation unit 21.

[0117] For example, the reliability determination unit 211 determines the reliability of the diagnostic model based on the number of training data used in the generation of the diagnostic model by the diagnostic model generation unit 21.

[0118] For example, the administrator of the device creates a condition in advance (hereafter referred to as the "reliability determination condition") specifying how much training data should be used to determine reliability, and the condition is stored in a location accessible to the diagnostic model generation device 2. The administrator or the device creates the reliability determination condition such that the reliability of a diagnostic model generated with a larger amount of training data is higher. Higher reliability means that the diagnostic model for which reliability was determined is more reliable, i.e., more accurate.

[0119] It should be noted that the reliability determination unit 211 performs the reliability determination for each diagnostic model.

[0120] The diagnostic model generation unit 21 stores the generated diagnostic model in the diagnostic model storage unit 43.

[0121] At this point, the diagnostic model generation unit 21 links the diagnostic model with the data specifying the range of values ​​for the external factor corresponding to the diagnostic model and with the reliability determined by the reliability determination unit 211. The diagnostic model generation unit 21 can further link the diagnostic model with the data specifying the characteristic to be entered.

[0122] The degradation diagnostic device 3 according to the first embodiment is described below.

[0123] Fig. Figure 6 is a representation showing an example configuration of the degradation diagnostic device 3 according to the first embodiment.

[0124] The degradation diagnostic device 3 comprises a diagnostic data acquisition unit 31, a model determination unit 32, a model acquisition unit 33, a feature extraction unit 34, a diagnostic unit 35 and a diagnostic result output unit 36.

[0125] The diagnostic data acquisition unit 31 obtains diagnostic data from sensor 5, including sensor data (here vibration data) relating to the operating state of the target device, and data relating to the external factor (here temperature data) relating to the operating environment of the target device.

[0126] The diagnostic data acquisition unit 31 outputs the data relating to the external factor among the acquired diagnostic data to the model determination unit 32, and outputs the sensor data to the feature extraction unit 34.

[0127] The model determination unit 32 determines the diagnostic model to be used in the degradation diagnosis on the basis of the data procured by the diagnostic data acquisition unit 31 concerning the external factor and the data that specify the range of values ​​of the data associated with the diagnostic model concerning the external factor.

[0128] The model determination unit 32 refers to the diagnostic model storage unit 43 of the storage device 4 and determines the diagnostic model.

[0129] For example, it is assumed that the external factor data are data relating to the external factor, indicating an external factor (here, temperature) of 18°C. It is also assumed that the second diagnostic model, stored in diagnostic model storage unit 43, is associated with data indicating a range of 17°C to 28°C for the data relating to the external factor.

[0130] In this case, the model determination unit 32 determines the second diagnostic model as the diagnostic model to be used in the degradation diagnosis.

[0131] It should be noted that, as described above, the diagnostic model generation unit 21 of the diagnostic model generation device 2 does not generate a diagnostic model based on the training data that corresponds to the value range of the data concerning the external factor to which an external factor indicator is assigned, where the external factor indicator shows that the range is a transition band or a range in which the relationship between the data concerning the external factor and the feature is estimated to change frequently. That is to say, depending on the value of the data concerning the external factor, the corresponding diagnostic model may not be generated.

[0132] In this case, the model determination unit 32 determines that the diagnostic model to be used for degradation diagnosis has not yet been generated.

[0133] The model determination unit 32 outputs data (hereinafter referred to as "model identification data") to the model procurement unit 33, which can be used to identify the diagnostic model intended for degradation diagnosis. It should be noted that data (e.g., an ID) is added to the diagnostic model, which allows the diagnostic model generation unit 21 to identify the diagnostic model at the time of its generation.

[0134] If it is determined that the diagnostic model to be used in the degradation diagnosis has not been generated, the model determination unit 32 outputs data to the model procurement unit 33 (hereinafter referred to as the "model not generated message") to indicate that the diagnostic model has not been generated.

[0135] In a case where the model identification data is output by the model determination unit 32, the model acquisition unit 33 acquires the diagnostic model determined by the model determination unit 32 for use in degradation diagnosis from the diagnostic model storage unit 43 of the storage device 4.

[0136] When the model generation message is issued by the model determination unit 32, the model procurement unit 33 generates the diagnostic model to be used in the degradation diagnosis.

[0137] The generation of the diagnostic model by the model procurement unit 33 will now be described in detail.

[0138] Based on the value of the data concerning the external factor acquired by the diagnostic data acquisition unit 31, the model acquisition unit 33 refers to the training data storage unit 41 of the storage device 4 to acquire the training data in which data concerning the external factor and the features are mapped. It should be noted that the model acquisition unit 33 can acquire the data concerning the external factor from the diagnostic data acquisition unit 31 via the model determination unit 32.

[0139] The condition (hereinafter referred to as the “condition for obtaining data about external factors”), regarding which data concerning the external factor are to be obtained by the diagnostic data acquisition unit 31, will be generated beforehand by the administrator or the like and stored in a location to which the degradation diagnostic device 3 can refer.

[0140] For example, a condition is set for acquiring data about external factors that "external factor data within the range of ±0.1°C of the value of the external factor data acquired by diagnostic data acquisition unit 31 are defined as data around the external factor." If, in this case, the value of the external factor data acquired by diagnostic data acquisition unit 31 is, for example, 16°C, model acquisition unit 33 acquires the training data from training data storage unit 41, in which the external factor data, showing values ​​between 15.9°C and 16.1°C, and the feature are assigned.

[0141] Alternatively, a condition for acquiring the data of the external factor can be set, for example, that "data of the external factor with ten values, ordered by their proximity to the value of the data of the external factor acquired by the diagnostic data acquisition unit 31, are defined as data around the external factor." If, in this case, the value of the data acquired by the diagnostic data acquisition unit 31 concerning the external factor is, for example, 16°C, the model acquisition unit 33 identifies ten pieces of data concerning the external factor with values ​​around and close to 16°C and acquires from the training data storage unit 41 the training data in which the identified data concerning the external factor and the feature are mapped to each other.

[0142] After acquiring the training data, the model acquisition unit 33 generates a diagnostic model based on the acquired training data.

[0143] The model procurement unit 33 can generate a diagnostic model using the same procedure by which the diagnostic model generation unit 21 of the diagnostic model generation device 2 generates a diagnostic model.

[0144] After the diagnostic model has been generated, the model acquisition unit 33 determines the reliability of the generated diagnostic model. The model acquisition unit 33 can determine the reliability of the diagnostic model using the same procedure by which the reliability determination unit 211 of the diagnostic model generation device 2 determines the reliability of a diagnostic model.

[0145] As described above, in a case where the diagnostic model to be used in the degradation diagnosis has not yet been generated, the model procurement unit 33 generates a diagnostic model based on the local training data in accordance with the condition for obtaining the data concerning the external factor.

[0146] It should be noted that the data concerning the external factors, as described above, are time series data.

[0147] Model procurement unit 33 carries out a process of procuring the appropriate diagnostic model for the data concerning the external factor in time series.

[0148] Furthermore, after the procurement of the corresponding diagnostic model for the data relating to the external factor, the model procurement unit 33 generates data (hereinafter referred to as "applicable model domain data") in which the procurement date and procurement time of the data relating to the external factor and the data for identifying the corresponding diagnostic model are linked together in time series.

[0149] Note that the procurement date and time of the data relating to the external factor will be added to the data relating to the external factor.

[0150] The model acquisition unit 33 outputs to the diagnostic unit 35 the diagnostic model that was acquired by reference to the diagnostic model storage unit 43, or the diagnostic model that was acquired by generation based on the training data acquired by the training data storage unit 41, together with the generated applicable model domain data.

[0151] Once the diagnostic model has been generated, the model procurement unit 33 assigns the diagnostic model to the data that specifies the value of the data concerning the external factor corresponding to the diagnostic model and the determined reliability. The model procurement unit 33 can also link the diagnostic model to the data that specifies the characteristic to be entered.

[0152] The feature extraction unit 34 extracts the feature from the sensor data acquired by the diagnostic data acquisition unit 31.

[0153] More precisely, the feature extraction unit 34 extracts the feature to be entered into the diagnostic model from the sensor data acquired by the diagnostic data acquisition unit 31.

[0154] It should be noted that the administrator or similar entity generates data (hereinafter referred to as "feature definition data") to identify the feature to be entered into the diagnostic model to be used in the degradation diagnosis on the target device in the degradation diagnosis device 3, and holds the data in a buffer or similar of the feature extraction unit 34. The feature extraction unit 34 determines the type of feature to be extracted in accordance with the feature definition data.

[0155] The feature to be entered into the diagnostic model is the amplitude value. Feature extraction unit 34 extracts the vibration amplitude value as a feature.

[0156] The feature extraction unit 34 can extract the feature from the sensor data using a known feature extraction method.

[0157] The feature extraction unit 34 outputs the extracted feature to the diagnostic unit 35.

[0158] Based on the feature extracted by the feature extraction unit 34, the diagnostic unit 35 performs a degradation diagnosis on the target device using the diagnostic model acquired by the model acquisition unit 33.

[0159] First, the diagnostic unit 35 determines which diagnostic model corresponds to the feature extracted by the feature extraction unit 34 among the diagnostic models issued by the model procurement unit 33.

[0160] In particular, based on the feature extracted by the feature extraction unit 34, the diagnostic unit 35 establishes the diagnostic model procured by the model procurement unit 33 as the diagnostic model corresponding to the feature extracted by the feature extraction unit 34, or as the diagnostic model to be used in the degradation diagnosis on the target device based on the feature extracted by the feature extraction unit 34, with respect to the data concerning the external factor corresponding to the feature.

[0161] Fig. Figure 7 is a diagram illustrating a method in which the diagnostic unit 35 determines the diagnostic model that corresponds to the feature extracted by the feature extraction unit 34 from among the diagnostic models output by the model procurement unit 33 in the first embodiment.

[0162] As described above, the sensor data and the data relating to the external factor are linked to time series data. This means that the sensor data and the data relating to the external factor are linked to the acquisition dates and times of the data. It is important to note that the acquisition dates and times of the data are assigned to the respective sensor data and the data relating to the external factor.

[0163] The diagnostic unit 35 compares the procurement dates and times of the data contained in the applicable model space data with the procurement dates and times of the data added to the characteristic based on the applicable model space data issued by the model procurement unit 33, in order to determine which diagnostic model is used for the characteristic and during which period.

[0164] As in Fig. Figure 7 shows, for example, that in the applicable model domain data a diagnostic model “B” is associated with the data concerning the external factor in the time span from time t1 to time t2.

[0165] In this case, the diagnostic unit 35 determines that the diagnostic model to be used in the degradation diagnosis on the target device is the diagnostic model “B” by inputting the feature of the period from time t1 to time t2.

[0166] In this way, the diagnostic unit 35 determines that the diagnostic model acquired on the basis of the data relating to the external factor of the corresponding data acquisition date and the corresponding data acquisition time for the feature extracted by the feature extraction unit 34 is the diagnostic model to be used in the degradation diagnosis of the target device on the basis of the feature.

[0167] After determining the diagnostic model to be used in the degradation diagnosis of the target device based on the feature extracted by the feature extraction unit 34, the diagnostic unit 35 performs a degradation diagnosis of the target device using the diagnostic model. It should be noted that the diagnostic model was issued by the model acquisition unit 33. The diagnostic unit 35 can identify the diagnostic model based on the diagnostic model identification data contained in the applicable model domain data.

[0168] For example, diagnostic unit 35 compares the value indicating the condition of the target device, as obtained by inputting the feature into the diagnostic model, with a preset threshold (hereinafter referred to as the "degradation diagnostic threshold") and diagnoses that the target device has degraded if the value indicating the condition of the target device is equal to or greater than the degradation diagnostic threshold. If the value indicating the condition of the target device is less than the degradation diagnostic threshold, diagnostic unit 35 diagnoses that the target device has not degraded.

[0169] The diagnostic unit 35 can perform a degradation diagnosis for the targeting device by determining the degree of degradation of the targeting device, rather than performing a degradation diagnosis for the targeting device depending on whether the targeting device has deteriorated.

[0170] For example, diagnostic unit 35 determines the degree of device degradation, which indicates the degree of deterioration or degradation of the target device, depending on the value representing the target device's state, obtained by inputting the feature into the diagnostic model. It is assumed that a condition specifying the degree of target device degradation is generated in advance and stored in a buffer or similar within diagnostic unit 35.

[0171] The diagnostic unit 35 outputs a result of the degradation diagnosis at the target device (the result is hereinafter referred to as the “degradation diagnosis result”) to the diagnostic result output unit 36.

[0172] At this time, the diagnostic unit 35 outputs the diagnostic model used in the degradation diagnosis at the target device to the diagnostic result output unit 36, whereby the diagnostic model is linked to the degradation diagnosis result.

[0173] The diagnostic result output unit 36 ​​outputs the degradation diagnostic result to an output device.

[0174] At this point, the diagnostic result output unit 36 ​​can output the reliability of the diagnostic model used in the degradation diagnosis at the target device, along with the degradation diagnosis result, to the output device.

[0175] The diagnostic result output unit 36 ​​can output the data for identifying the diagnostic model used in the degradation diagnosis at the target device, along with the degradation diagnosis result, to the output device.

[0176] The reliability of the diagnostic model and the data for identifying the diagnostic model have been assigned to the diagnostic model. Based on the diagnostic model output by diagnostic unit 35, diagnostic result output unit 36 ​​can determine the reliability of the diagnostic model and the data for identifying the diagnostic model.

[0177] When the diagnostic result output unit 36 ​​outputs the degradation diagnostic result to the output device, the output device displays the degradation diagnostic result or outputs it audibly.

[0178] The operator or similar can, for example, use the displayed or audibly output result of the degradation diagnosis on the targeting device to determine whether the targeting device has deteriorated or what the degree of degradation of the targeting device is.

[0179] When the diagnostic result output unit 36 ​​outputs the reliability, the output device can display or audibly announce the reliability along with the result of the degradation diagnosis at the target device. As a result, in a case where the reliability is low, the operator or similar can ascertain the certainty of the degradation diagnosis result and take appropriate action, such as an early inspection of the target device, even if the target device was diagnosed as not degraded.

[0180] When the diagnostic output unit 36 ​​outputs the diagnostic model used in the degradation diagnosis, the output device can display or audibly output the data for identifying the diagnostic model along with the result of the degradation diagnosis at the target device. This allows, for example, the operator or similar person to evaluate the diagnostic model.

[0181] The processes to be carried out by the diagnostic model generation device 2, including the label generation device 1, and the degradation diagnostic device 3 are now described.

[0182] First, the processes performed by the license plate generation device 1 and the diagnostic model generation device 2 are described.

[0183] Fig. Figure 8 is a flowchart to illustrate the processes to be carried out by the diagnostic model generation device 2 according to the first embodiment, including the identifier generation device 1 according to the first embodiment.

[0184] Upon receiving a command to start the process, the diagnostic model generation device 2 starts a process as shown in the flowchart in Fig. Figure 8 illustrates this. For example, the operator or a similar person enters an instruction to start the process from a PC in the control room. Upon receiving the command to start operation, a control unit (not shown) of the diagnostic model generation device 2 causes the license plate generation device 1 and the diagnostic model generation unit 21 to commence operation.

[0185] What the flowchart in Fig. With regard to the processes shown in Figure 8, a process that includes the processes in steps ST1 to ST3 is a process to be performed by the label generation device 1, and a process that includes the process in step ST4 is a process to be performed by the diagnostic model generation device 2.

[0186] The training data acquisition unit 11 refers to the training data storage unit 41 and acquires training data (step ST1).

[0187] The training data acquisition unit 11 outputs the acquired training data to the distribution unit 12 for a region of the external factor.

[0188] Based on the training data output by the training data acquisition unit 11 in step ST1, the partitioning unit 12 performs a process for partitioning a region of the external factor, in which a constant band and a transition band are classified in the value range of the data relating to the external factor in the axial direction of the data relating to the external factor (step ST2).

[0189] The distribution unit 12 for a region of the external factor outputs the region setting data to the indicator data generation unit 13 for the external factor.

[0190] Based on the region setting data output by the splitting unit 12 for a region of the external factor in step ST2, the indicator data generation unit 13 assigns an external factor indicator, indicating that the area is a constant band, to the range of values ​​of the data relating to the external factor that are classified as a constant band, and an external factor indicator, indicating that the area is a transition band, to the range of values ​​of data relating to the external factor that are classified as a transition band, with respect to the axis of the data relating to the external factor.

[0191] After the assignment of the indicators for the external factor, the indicator data generation unit 13 generates the data for the indicators for the external factor (step ST3).

[0192] The external factor identification data generation unit 13 stores the generated identification data for the external factor in the external factor identification data storage unit 42 of storage device 4.

[0193] Based on the external factor identifier data generated by the external factor identifier data generation unit 13 in step ST3, the diagnostic model generation unit 21 acquires from the training data storage unit 41 of storage device 4 the training data corresponding to the value range of the external factor data associated with the external factor identifier, indicating that the range is one in which the relationship between the external factor data and the feature is constant. The diagnostic model generation unit 21 then generates a diagnostic model using the acquired training data (step ST4).

[0194] It should be noted that in step ST4, the reliability determination unit 21 contained in the diagnostic model generation unit 211 determines the reliability of the diagnostic model.

[0195] The diagnostic model generation unit 21 stores the generated diagnostic model in the diagnostic model storage unit 43.

[0196] At this point, the diagnostic model generation unit 21 links the diagnostic model with the data specifying the range of values ​​for the external factor corresponding to the diagnostic model and with the reliability determined by the reliability determination unit 211. The diagnostic model generation unit 21 can further link the diagnostic model with the data specifying the characteristic to be entered.

[0197] Fig. Figure 9 is a flowchart to explain details of a region allocation process carried out by allocation unit 12 for a region of an external factor in step ST2 in Fig. 8 is to be carried out.

[0198] First, the inflection point extraction unit 121 extracts a multitude of inflection points in the axial direction of the data relating to the external factor associated with the feature in the training data, based on the training data output by the training data acquisition unit 11 (step ST101).

[0199] The inflection point extraction unit 121 outputs the training data supplemented by the inflection point to the partitioning unit 12 for a region of the external factor.

[0200] Based on the training data augmented with the inflection point, output by the inflection point extraction unit 121, the preliminary setting unit 122 selects the first and second inflection points. Preliminary setting unit 122 defines the range of values ​​in the data relating to the external factor between the selected first and second inflection points as the inflection point region and establishes a preliminary region that is a candidate for a constant band within the inflection point region (step ST101).

[0201] The setting unit 122 for a preliminary region outputs the value of the data concerning the external factor in the defined preliminary region and the value of the characteristic that corresponds to the value of the data concerning the external factor to the generation unit 123 for a relational model.

[0202] Based on the values ​​of the data concerning the external factor in the preliminary region and the values ​​of the feature that corresponds to the values ​​of the data concerning the external factor, the generating unit 123 for a relational model generates a relational model (step ST103).

[0203] The generation unit 123 for a relational model outputs the generated relational model to the evaluation unit 124 for a relational model, together with the values ​​of the data concerning the external factor in the preliminary region and the values ​​of the characteristic that correspond to the values ​​of the data concerning the external factor, which are output by the setting unit 122 for the preliminary region.

[0204] The evaluation unit 124 for the relational model evaluates the accuracy of the relational model using the values ​​of the data relating to the external factor contained in the preliminary region and the values ​​of the feature corresponding to the values ​​of the data relating to the external factor, and outputs the result of the evaluation of the relational model together with the values ​​of the data relating to the external factor in the preliminary region and the values ​​of the feature corresponding to the values ​​of the data relating to the external factor, or, in other words, together with the training data in the preliminary region, to the region determination unit 125.

[0205] Based on the result of the evaluation of the relational model issued by the evaluation unit 124 for the relational model, the region determination unit 125 determines whether the relational model is evaluated with a sufficiently high accuracy (step ST104).

[0206] If the accuracy of the relational model is determined to be sufficiently high (if "YES" in step ST104), the region determination unit 125 classifies the preliminary region in which the accuracy of the relational model is rated as high as a constant band. The region determination unit 125 then classifies any range of values ​​in the data relating to the external factor that differs from the determined constant band as a transition band within the inflection point region (step ST105).

[0207] The distribution unit 12 for a region of the external factor outputs the region setting data to the indicator data generation unit 13 for the external factor.

[0208] If the accuracy of the relational model is deemed insufficient or low (if "NO" in step ST104), the region determination unit 125 issues an instruction to the preliminary region setting unit 122 to restrict the preliminary region's scope and reset it. Based on the instruction from region determination unit 125, preliminary region setting unit 122 resets the preliminary region. Specifically, preliminary region setting unit 122 limits the range of values ​​for the external factor data at both ends of the preliminary region and resets the preliminary region. The operation of the external factor region splitting unit 12 then reverts to the procedure in step ST102.

[0209] It should be noted that in step ST104, the region determination unit 125 determines whether the preliminary region set by the preliminary region setting unit 122 is smaller than the minimum range for the preliminary region. If the preliminary region set by the preliminary region setting unit 122 is smaller than the minimum range of the preliminary region, the region determination unit 125 determines that the preliminary region cannot be set as a constant band, even if the relational model is evaluated to have sufficiently high accuracy, and classifies the preliminary region as a transition band. The region determination unit 125 then determines to classify a range of values ​​in the data relating to the external factor, which is not the preliminary region, as a transition band in the inflection point region.The distribution unit 12 for a region of the external factor outputs the region setting data to the indicator data generation unit 13 for the external factor. The operation of distribution unit 12 for a region of the external factor then completes the process, as shown in the flowchart in [reference]. Fig. 9 shown.

[0210] As with the technique disclosed in the aforementioned patent literature 1, in a case where a model for an operating state of a device is generated, the index of the operating state is less likely to have continuous values, and accordingly, the accuracy of a diagnostic result is less likely to decrease at the model change boundary. That is, depending on the sensor data acquired by the device, several models can be uniquely switched. (See the upper diagram in the upper part of Fig. 10)

[0211] However, the operating environment index will likely exhibit continuous values. In this case, it is possible that a highly accurate diagnostic result cannot be achieved, even if one of the models is used before and after the model change at the model change boundary. (See the lower diagram in the lower part of Fig. 10. It should be noted that the point in time at which there is a possibility that even when using one of the models a very accurate diagnostic result cannot be obtained is represented as a transition band.

[0212] Using conventional techniques, it is difficult to construct a model when considering the possibility that an accurate diagnostic result cannot be obtained due to the operating environment, even when one of the models is used before and after the switchover.

[0213] On the other hand, the identifier generation device 1 acquires training data that includes a feature and associated data relating to the external factor, and classifies, on the basis of the acquired training data, the range of values ​​of the data relating to the external factor in the axial direction of the data relating to the external factor into a constant band, which is a range of values ​​of data relating to the external factor whose relationship to the corresponding feature is considered constant, and a transition band, which is a range of values ​​of data relating to the external factor whose relationship to the corresponding feature is considered to change frequently.With respect to the axis of the data concerning the external factor, the indicator generation device 1 then assigns to the value range of the data concerning the external factor classified as a constant band an indicator for the external factor, which indicates that the range is a range in which the relationship to the feature corresponding to the data concerning the external factor in the range is constant; assigns to the value range of the data concerning the external factor classified as a transition band an indicator for the external factor, which indicates that the range is a range in which the relationship to the feature corresponding to the data concerning the external factor in the range changes frequently; and generates indicator data for the external factor in which the value ranges of the data concerning the external factor are linked to the indicators for the external factor.

[0214] In this way, the label generation device 1 can provide data that aids in the creation of a diagnostic model, preventing a reduction in diagnostic accuracy when the target device degrades due to the operating environment. Consequently, the label generation device 1 can contribute to the creation of a diagnostic model that prevents a reduction in diagnostic accuracy when the target device degrades due to the operating environment and can help achieve an accurate degradation diagnosis with the diagnostic model. Furthermore, the label generation device 1 can help reduce the time required to design a diagnostic model.

[0215] Furthermore, when generating the tag data for the external factor, tag generation device 1 focuses on the relationship between the feature based on sensor data in a normal state of the target device and the external factor. In other words, the training data used by tag generation device 1 to generate the tag data for the external factor only needs to be data that at least corresponds to the relationship between the feature based on sensor data in a normal state of the target device and the external factor. This reduces the effort required to generate training data. Normally, preparing training data based on sensor data in an abnormal state requires more effort than preparing training data based on sensor data in a normal state.

[0216] Furthermore, by supplying the data concerning the external factor, the label generation device 1 can ensure that the operator or the like clearly recognizes which external factor, for example, causes the diagnostic model to switch at what time.

[0217] Based on the external factor identifier data generated by identifier generation device 1, diagnostic model generation device 2 acquires the training data corresponding to the range of values ​​of the data relating to the external factor associated with the external factor identifier, indicates that the range is a range in which the relationship between the data relating to the external factor in the range and the corresponding feature is constant, and generates a diagnostic model using the acquired training data.

[0218] By generating a diagnostic model based on the tag data for external factors generated by tag generation device 1, diagnostic model generation device 2 can create a diagnostic model that prevents a reduction in diagnostic accuracy in the event of deterioration or degradation of a device due to the operating environment.

[0219] Next, the functionality of the degradation diagnostic device 3 will be described.

[0220] Fig. Figure 11 is a flowchart to explain the operation of the degradation diagnostic device 3 according to the first embodiment.

[0221] Upon receiving a command to start degradation diagnostics at the target device, the degradation diagnostic device 3 begins a process as shown in the flowchart in Fig. Figure 11 illustrates this. For example, the operator or a similar person issues an instruction from a PC in the control room to start a degradation diagnosis on the target device. Upon receiving the instruction to start the degradation diagnosis on the target device, a control unit (not shown) of the degradation diagnosis device 3 causes the diagnostic data acquisition unit 31, the model determination unit 32, the model acquisition unit 33, the feature extraction unit 34, the diagnostic unit 35, and the diagnostic result output unit 36 ​​to commence operation.

[0222] The diagnostic data acquisition unit 31 acquires diagnostic data from sensor 5 (step ST21).

[0223] The diagnostic data acquisition unit 31 outputs the data relating to the external factor among the acquired diagnostic data to the model determination unit 32, and outputs the sensor data to the feature extraction unit 34.

[0224] The model determination unit 32 determines the diagnostic model to be used in the degradation diagnosis on the basis of the data concerning the external factor acquired by the diagnostic data acquisition unit 31 in step ST21 and the data that specify the range of values ​​of the data concerning the external factor assigned to the diagnostic model (step ST22).

[0225] Model determination unit 32 outputs the model identification data to model procurement unit 33. If it is determined that the diagnostic model to be used for degradation diagnosis has not been created, model determination unit 32 outputs a message to model procurement unit 33 indicating that the model has not been created.

[0226] In a case where the model identification data is output by the model determination unit 32 in step ST22, the model acquisition unit 33 acquires the diagnostic model that was determined by the model determination unit 32 for use in degradation diagnosis from the diagnostic model storage unit 43 of the storage device 4. In a case where a message indicating that a model was not generated is output by the model determination unit 32 in step ST22, and in a case where data is output indicating that a diagnostic model was not generated, the model acquisition unit 33 acquires a diagnostic model by generating the diagnostic model to be used in degradation diagnosis (step ST23).

[0227] After the procurement of the corresponding diagnostic model for the data concerning the external factor has been completed, the model procurement unit 33 generates the corresponding model domain data.

[0228] The model acquisition unit 33 outputs to the diagnostic unit 35 the diagnostic model that was acquired by reference to the diagnostic model storage unit 43, or the diagnostic model that was acquired by generation based on the training data acquired by the training data storage unit 41, together with the generated applicable model domain data.

[0229] The feature extraction unit 34 extracts the feature from the sensor data acquired by the diagnostic data acquisition unit 31 in step ST21 (step ST24).

[0230] The feature extraction unit 34 outputs the extracted feature to the diagnostic unit 35.

[0231] Based on the feature extracted by the feature extraction unit 34 in step ST24, the diagnostic unit 35 performs a degradation diagnosis on the target device using the diagnostic model acquired by the model acquisition unit 33 in step ST23 (step ST25).

[0232] The diagnostic unit 35 outputs the result of the degradation diagnosis to the diagnostic result output unit 36.

[0233] At this time, the diagnostic unit 35 outputs the diagnostic model used in the degradation diagnosis at the target device to the diagnostic result output unit 36, whereby the diagnostic model is linked to the degradation diagnosis result.

[0234] The diagnostic result output unit 36 ​​outputs the degradation diagnostic result to the output device (step ST26).

[0235] At this point, the diagnostic result output unit 36 ​​can output the reliability of the diagnostic model used in the degradation diagnosis at the target device, along with the degradation diagnosis result, to the output device.

[0236] The diagnostic result output unit 36 ​​can output the data for identifying the diagnostic model used in the degradation diagnosis at the target device, along with the degradation diagnosis result, to the output device.

[0237] It should be noted that in the flowchart in Fig. 11 the processes in steps ST22 and ST23 and the process in step ST24 are carried out in parallel, but this is just an example.

[0238] For example, the process in step ST24 can be carried out after the processes in steps ST22 and ST23, or vice versa.

[0239] It is only necessary to complete the operations in steps ST22 and ST23 and the operation in step ST24 before performing the operation in step ST25.

[0240] As described above, the diagnostic model generation device 2 in the degradation diagnostic device 3 performs a degradation diagnosis on the target device, using the diagnostic model generated based on the tag data for the external factor produced by the tag generation device 1. The degradation diagnostic device 3 acquires the diagnostic data, including sensor data and data relating to the external factor, and determines the diagnostic model to be used in the degradation diagnosis based on the acquired data relating to the external factor and the data specifying the value range of the data relating to the external factor associated with the diagnostic model. The degradation diagnostic device 3 acquires the diagnostic model intended for use in the degradation diagnosis.

[0241] Furthermore, the degradation diagnostic device 3 extracts a feature from the acquired sensor data.

[0242] Based on the extracted feature, the degradation diagnostic device 3 then performs a degradation diagnosis on the target device, using the diagnostic model that was acquired for the data corresponding to the feature concerning the external factor. The degradation diagnostic device 3 outputs the result of the degradation diagnosis performed on the target device.

[0243] The degradation diagnostic device 3 performs a degradation diagnosis on the target device using the diagnostic model, which was generated taking into account the possibility that an accurate diagnostic result cannot be obtained before and after switching due to the operating environment, and can thus perform an accurate degradation diagnosis.

[0244] If the diagnostic model for degradation diagnosis has not yet been generated, the degradation diagnosis device 3 generates a diagnostic model each time based on local training data in accordance with the condition for obtaining data concerning the external factor.

[0245] By generating a diagnostic model that uses only the local training data, the degradation diagnostic device 3 can create a diagnostic model that is compatible with the properties of a transition band in which the optimal diagnostic model changes frequently.

[0246] It should be noted that in the first embodiment described so far, for the sake of simplicity, the identifier generation device 1 was described as a device that generates identifier data for the external factor for a data set of one type of feature (amplitude value) and one type of external factor (temperature), but this is only an example.

[0247] The training data can be, for example, data where the features are linked to data relating to the external factor, which can relate to a variety of different types of external factors, such as temperature and precipitation.

[0248] The characteristic can also be several different types of characteristics, e.g., amplitude value, variance, and standard deviation.

[0249] In the identifier generation device 1, the training data to be procured by the training data acquisition unit 11 can contain data relating to the external factor, which can be linked to a plurality of different types of external factors, with the characteristic for each external factor. Similarly, in the training data to be procured by the training data acquisition unit 11, a plurality of characteristics can be assigned to one external factor.

[0250] The identifier generation device 1 generates the identifier data for the external factor for each set of data relating to the external factor and a feature.

[0251] In particular, in the identifier generation device 1, the partitioning unit 12 classifies a constant band and a transition band in the value range of the data relating to the external factor for each set of data relating to the external factor and a feature, and the identifier data generation unit 13 for the external factor generates the identifier data for the external factor by assigning an identifier for the external factor, indicating the constant band or the transition band, to the axis of the data relating to the external factor for each set of data relating to the external factor and a feature.

[0252] For example, it is assumed that features “amplitude value” and “variance” with respect to the external factor “temperature” serve as inputs to the diagnostic model.

[0253] In this case, the identifier generation device 1 generates data relating to the external factor for each set of "temperature" and "amplitude value" and one set of "temperature" and "variance".

[0254] Fig. Figure 12 is a diagram illustrating a concept of an example of external factor identifiers assigned to the external factor data axis by the external factor identifier data generation unit 13 in the identifier generation device 1 in a case where the identifier generation device 1 generates external factor data for each set of "temperature" and "amplitude value" and a set of "temperature" and "variance" in the first embodiment.

[0255] In Fig. With respect to an external factor (here the temperature) in the set “temperature” and “amplitude value”, the area below 10 °C is classified as a constant band, the area from 10 °C or more to below 17 °C as a transition band and the area from 17 °C or more to below 28 °C as a constant band by the division unit 12 for a region of the external factor.

[0256] In this case, the indicator data generation unit 13 assigns an indicator for an external factor with respect to the axis of the data relating to the external factor, indicating that the range below 10°C is a constant band (in which the relationship between the feature and the data relating to the external factor is constant) and is a range in which it is estimated that the same diagnostic model (hereinafter referred to as the "first diagnostic model") can be used. Fig. 4 is the identifier for the external factor, represented as the "first model application region". Fig. 12 is the identifier for the external factor, represented as the "first model-a application region".

[0257] Furthermore, the identifier data generation unit 13 assigns an identifier for an external factor, indicating that the range from 10°C or more to less than 17°C in the axial direction of the data relating to the external factor is a transition range. Fig. 12 is the indicator for the external factor as "no applicable model".

[0258] In the axial direction of the data relating to the external factor, the indicator data generation unit 13 assigns an indicator for the external factor, indicating that the range from 17°C or more to less than 28°C is a constant band (in which the relationship between the feature and the data relating to the external factor is constant) and is a range in which it is assumed that the same diagnostic model (second diagnostic model), which differs from the first diagnostic model, can be used. Fig. 12 is the identifier for the external factor, represented as the “second model application region”.

[0259] Regarding the external factor (temperature) in the set "temperature" and "variance", the range below 8°C is classified as a constant range, the range from 8°C or more to below 17°C as a transition range, and the range from 17°C or more to below 25°C as a constant range by the division unit 12 for a region of the external factor.

[0260] In this case, the indicator data generation unit 13 assigns an indicator for the external factor, with respect to the axis of the data relating to the external factor, indicating that the range below 8°C is a constant band (in which the relationship between the feature and the data relating to the external factor is constant) and is a range in which it is estimated that the same diagnostic model (first diagnostic model) can be used. Fig. 12 is the identifier for the external factor, represented as the "first model b application region".

[0261] Furthermore, the indicator data generation unit 13 assigns an indicator for an external factor, indicating that the range of 8°C or more and less than 17°C is a transition range with respect to the axis of data concerning the external factor. Fig. 12 is the indicator for the external factor as "no applicable model".

[0262] With respect to the axis of data concerning the external factor, the indicator data generation unit 13 assigns an external factor indicator that shows that the range from 17°C or more to less than 25°C is a constant band (in which the relationship between the feature and the data concerning the external factor is constant) and is a range in which it is estimated that the same diagnostic model (second diagnostic model), which differs from the first diagnostic model, can be used. Fig. 12 is the identifier for the external factor, represented as the “second model b application region”.

[0263] It should be noted that in the case of the above example, the range of values ​​of the data relating to the external factor to which the external factor indicator showing a constant band is assigned, and the range of values ​​of the data relating to the external factor to which the external factor indicator showing a transition band is assigned, will vary depending on the type of feature.

[0264] A concrete example: For the range of 8°C or more and less than 10°C, the external factor flag “first model-a application region” is assigned in the set “Temperature” and “Amplitude Value”, but the external factor flag “no applicable model” is assigned in the set “Temperature” and “Variance”.

[0265] In this way, in a case where there is external factor identifier data corresponding to a plurality of features, and external factor identifier data to which external factor identifiers with different contents in a range of values ​​of the same data relating to the external factor along the axis of the same data relating to the external factor are assigned, the range of values ​​of the data relating to the external factor that has a constant relationship to the corresponding feature is readjusted for the generation of a diagnostic model based on the external factor identifier data generated by identifier generation device 1.

[0266] For example, the administrator or similar checks the tag data for the external factor based on the tag data for the external factor, as in Fig. 12 is shown, and determined to generate the first diagnostic model for “temperature” less than 10°C, which is an external factor, for a diagnostic model with “amplitude value” and “variance” as inputs.

[0267] The administrator or the like sets the range of less than 10°C as the range of values ​​of the data concerning the external factor in which it is estimated that the first diagnostic model can be used.

[0268] The administrator or the like then creates a condition (hereinafter referred to as the "diagnostic model generation condition") for the generation of a diagnostic model based on the range of values ​​of the data relating to the external factor, assuming that the newly set first diagnostic model can be used.

[0269] The diagnostic model generation condition consists of data in a tabular format, in which the newly set value range of the data relating to the external factor, the external factor indicator in the newly set value range of the data relating to the external factor, data specifying the features (“amplitude value” and “variance” in the example above) that serve as inputs for the target diagnostic model, are defined as the application region by an external factor indicator, and data (hereinafter referred to as the “feature-specific region data”) that indicate whether the newly set value range of the data relating to the external factor is a constant band or a transition band in the external factor indicator data originally generated by the indicator generation device 1 for the features, are connected, for example.

[0270] In the example above, the administrator or similar creates a diagnostic model creation condition in which the data indicating that the range of "Temperature" as an external factor is less than 10°C, the flag for the external factor "First Diagnostic Model Application Region," and the feature-specific region data are linked together. In the feature-specific region data, the feature "Amplitude Value" is linked to data indicating that the range below 10°C is a constant band, and the feature "Variance" is linked to data indicating that the range below 8°C is a constant band and the range of 8°C or more and below 10°C is a transition band (see Fig. 12).

[0271] The administrator or the like stores the diagnostic model generation condition in storage unit 42 of identifier data for the external factor.

[0272] In the diagnostic model generation device 2, the diagnostic model generation unit 21 generates a diagnostic model in accordance with the diagnostic model generation condition if the diagnostic model generation condition is stored in the storage unit 42 of identifier data for the external factor.

[0273] In the example above, the diagnostic model generation unit 21 determines to generate the first diagnostic model for the data concerning the external factor of less than 10°C in accordance with the diagnostic model generation condition, and generates the first diagnostic model using the training data to which the data concerning the external factor of less than 10°C are added.

[0274] At this point, the reliability determination unit 211 can determine the reliability of the diagnostic model based on the feature-specific region data defined under the diagnostic model generation condition.

[0275] In the example above, the first diagnostic model is generated from the training data, to which the data concerning the external factor of less than 10°C are added. However, with respect to the "variance" feature extracted from the training data, the range of 8 degrees or more and less than 10°C is originally the range of values ​​of the data concerning the external factor, to which the external factor indicator of a transition band is assigned in the indicator generation device 1. For example, the reliability determination unit 211 determines the reliability to be low for a diagnostic model generated from the training data to which the data concerning the external factor with the external factor indicator of a transition band, assigned to them by the indicator generation device 1, were added, as described above.The reliability determination unit 211 can determine how low the reliability is to be determined, for example, on the basis of the data concerning the external factor to which the identifier of the external factor of a transition band was assigned by the identifier generation device 1, and the proportion of the given training data.

[0276] It should be noted that the above example is only an example and the reliability determination unit 211 can generate a different diagnostic model by defining the range of less than 8°C as a range in which "amplitude" and "variance" are used, and the range of 8°C or more and less than 10°C as a range in which only "amplitude" is used.

[0277] It should be noted that in a case where there is a set of a feature and an external factor, the administrator or the like can also create a diagnostic model generation condition after checking the feature data for the external factor.

[0278] The administrator or similar can generate a suitable diagnostic model generation condition based on the identifier data for the external factor.

[0279] In the first embodiment described above, the diagnostic model generation unit 21 of the diagnostic model generation device 2 also includes the reliability determination unit 211, but this is only an example. The diagnostic model generation device 2 does not necessarily have the task of determining the reliability of a diagnostic model.

[0280] In a case where the diagnostic model generation device 2 does not have the function of determining the reliability of a diagnostic model, the diagnostic model generation device 2 may be designed such that it does not include the reliability determination unit 211.

[0281] Furthermore, the model procurement unit 33 in the degradation diagnostic device 3 does not necessarily have the function of determining the reliability in the generation of a diagnostic model.

[0282] In the first embodiment described so far, the license plate generation device 1 is integrated into the diagnostic model generation device 2, but this is only an example.

[0283] The license plate generation device 1 and the diagnostic model generation device 2 can be independent devices.

[0284] In this case, after receiving an operating start command from an operator or the like and accepting the operating start command, a control unit (not shown) of the tag generation device 1 causes the training data acquisition unit 11, the distribution unit 12 for a region of the external factor and the tag data generation unit 13 for the external factor to be put into operation.

[0285] In the first embodiment described so far, the identifier generation device 1, the diagnostic model generation device 2 and the degradation diagnostic device 3 are included in the target device, but this is only an example.

[0286] For example, the tag generation device 1 may contain some or all of the training data acquisition units 11, the partitioning unit 12 for a region of the external factor, and the tag data creation unit with external factor 13 in a server or similar outside the target device.

[0287] Furthermore, the diagnostic model generation unit 21 of the diagnostic model generation device 2 can, for example, be located in a server or the like outside the target device.

[0288] Furthermore, in the degradation diagnostic device 3, some or all of the diagnostic data acquisition unit 31, the model determination unit 32, the model acquisition unit 33, the feature extraction unit 34, the diagnostic unit 35 and the diagnostic result output unit 36 ​​may, for example, be housed in a server or the like outside the target device.

[0289] Fig. 13A and Fig. Figures 13B are illustrations, each showing an example hardware configuration of the license plate generation device 1 according to the first embodiment.

[0290] In the first embodiment, the functions of the training data acquisition unit 11, the partitioning unit 12 for a region of the external factor, the label data generation unit 13 for the external factor, and the control unit (not shown) are performed by a processing circuit 101. That is, the label generation device 1 contains the processing circuit 101 for performing the control in order to provide data that helps in creating a diagnostic model, which prevents a decrease in the accuracy of the diagnosis in the event of device degradation due to the operating environment.

[0291] The processing circuit 101 can be dedicated hardware, as shown in Fig. 13A shown, or a processor 104 executing a program stored in memory, as in Fig. 13B shown.

[0292] In the case that the processing circuit 101 is dedicated hardware, the processing circuit 101 can be a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination thereof.

[0293] In a case where the processing circuit is the processor 104, the functions of the training data acquisition unit 11, the partitioning unit 12 for a region of the external factor, the tag data generation unit 13 for the external factor, and the control unit (not shown) are implemented by software, firmware, or a combination of software and firmware. The software or firmware is written as a program and stored in a memory 105. The processor 104 reads the program stored in memory 105 and executes it to perform the functions of the training data acquisition unit 11, the partitioning unit 12 for a region of the external factor, the tag data generation unit 13 for the external factor, and the control unit (not shown). That is, the tag generation unit 1 contains memory 105 for storing the program to execute steps ST1 to ST4 described above. Fig. 8, when the program is executed by processor 104. It can also be said that the program stored in memory 105 causes a computer to implement the processing procedures or methods performed by the training data acquisition unit 11, the partitioning unit 12 for a region of the external factor, the external factor identifier data generation unit 13, and the control unit (not shown). Here, memory 105 is a non-volatile or volatile semiconductor memory, such as RAM, read-only memory (ROM), flash memory, erasable programmable read-only memory (EPROM), or electrically erasable programmable read-only memory (EEPROM) (registered trademark, not mentioned hereafter), a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini-disk, a digital versatile disc (DVD), or the like.

[0294] It should be noted that the functions of the training data acquisition unit 11, the partitioning unit 12 for a region of the external factor, the tag data generation unit 13 for the external factor, and the control unit (not shown) can be implemented partly by dedicated hardware and partly by software or firmware. For example, the functions of the training data acquisition unit 11 can be implemented by the processing circuit 101 as dedicated hardware, and the functions of the partitioning unit 12 for a region of the external factor, the tag data generation unit 13 for the external factor, and the control unit (not shown) can be implemented by the processor 104, which reads and executes the program stored in memory 105.

[0295] Furthermore, the license plate generation device 1 comprises an input interface device 102 and an output interface device 103, which perform wired or wireless communication with a device such as the diagnostic model generation device 2 or the storage device 4.

[0296] Examples of the hardware configuration of the diagnostic model generation device 2 according to the first embodiment are also given in the Fig. 13A and Fig. 13B shown.

[0297] In the first embodiment, the functions of the diagnostic model generation unit 21 and the control unit (not shown) are performed by the processing circuit 101. That is, the diagnostic model generation device 2 contains the processing circuit 101 for carrying out the control to generate a diagnostic model based on the data generated by the diagnostic model generation device 1 concerning the external factor.

[0298] The processing circuit 101 can be dedicated hardware, as shown in Fig. 13A shown, or a processor 104 executing a program stored in memory, as in Fig. 13B shown.

[0299] In the case that the processing circuit 101 is dedicated hardware, the processing circuit 101 can be a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination thereof.

[0300] If the processing circuit is processor 104, the functions of the diagnostic model generation unit 21 and the control unit (not shown) are implemented by software, firmware, or a combination of software and firmware. The software or firmware is written as a program and stored in memory 105. Processor 104 reads the program stored in memory 105 and executes it to perform the functions of the diagnostic model generation unit 21 and the control unit (not shown). That is, the diagnostic model generation unit 2 contains memory 105 for storing the program to perform step ST4 described above. Fig. 8 to be executed when the program is run by processor 104. It can also be said that the program stored in memory 105 causes a computer to implement the processing operations or procedure to be carried out by the diagnostic model generation unit 21 and the control unit (not shown).

[0301] It should be noted that the functions of the diagnostic model generation unit 21 and the control unit (not shown) can be implemented partly by special hardware and partly by software or firmware.

[0302] Furthermore, the diagnostic model generation device 2 includes the input interface device 102 and the output interface device 103, which perform wired or wireless communication with a device such as the license plate generation device 1 or the storage device 4.

[0303] Examples of the hardware configuration of the degradation diagnostic device 3 according to the first embodiment are also given in the Fig. 13A and Fig. 13B shown.

[0304] In the first embodiment, the functions of the diagnostic data acquisition unit 31, the model determination unit 32, the model acquisition unit 33, the feature extraction unit 34, the diagnostic unit 35, the diagnostic result output unit 36, and the control unit (not shown) are implemented by the processing circuit 101. That is, in the degradation diagnostic device 3, the diagnostic model generation device 2 contains the processing circuit 101 for controlling the degradation diagnosis on the target device using a diagnostic model generated on the basis of the data concerning the external factor produced by the feature generation device 1.

[0305] The processing circuit 101 can be dedicated hardware, as shown in Fig. 13A shown, or a processor 104 executing a program stored in memory, as in Fig. 13B shown.

[0306] In the case that the processing circuit 101 is dedicated hardware, the processing circuit 101 can be a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination thereof.

[0307] In a case where the processing circuit is the processor 104, the functions of the diagnostic data acquisition unit 31, the model determination unit 32, the model acquisition unit 33, the feature extraction unit 34, the diagnostic unit 35, the diagnostic result output unit 36, and the control unit (not shown) are implemented by software, firmware, or a combination of software and firmware. The software or firmware is written as a program and stored in a memory 105. The processor 104 reads the program stored in the memory 105 and executes it to perform the functions of the diagnostic data acquisition unit 31, the model determination unit 32, the model acquisition unit 33, the feature extraction unit 34, the diagnostic unit 35, the diagnostic result output unit 36, and the control unit (not shown).This means that the degradation diagnostic device 3 uses memory 105 to store the program for executing steps ST21 to ST26 described above. Fig. 11 contains, when the program is executed by processor 104. It can also be said that the program stored in memory 105 causes a computer to implement the processing operations or procedure to be carried out by the diagnostic data acquisition unit 31, the model determination unit 32, the model acquisition unit 33, the feature extraction unit 34, the diagnostic unit 35, the diagnostic result output unit 36 ​​and the control unit (not shown).

[0308] It should be noted that the functions of the diagnostic data acquisition unit 31, the model determination unit 32, the model acquisition unit 33, the feature extraction unit 34, the diagnostic unit 35, the diagnostic result output unit 36 ​​and the control unit (not shown) may be implemented partly by special hardware and partly by software or firmware.

[0309] Furthermore, the degradation diagnostic device 3 includes the input interface device 102 and the output interface device 103, which perform wired or wireless communication with a device such as the sensor 5 or the storage device 4.

[0310] As described above, the identifier generation device 1 according to the first embodiment comprises: the training data acquisition unit 11, which acquires training data in which a feature that can be used to generate a diagnostic model to be used as the object of degradation diagnosis in the degradation diagnosis of the target device is associated with data relating to the external factor relating to the operating environment;the partitioning unit 12 for a region of the external factor, which, on the basis of the training data acquired by the training data acquisition unit 11, classifies a range of values ​​of the data relating to the external factor in the axial direction of the data relating to the external factor into a constant band, which is a range of values ​​of the data relating to the external factor in which the relationship to the corresponding characteristic is considered constant, and a transition band, which is a range of values ​​of the data relating to the external factor in which the relationship to the corresponding characteristic is considered to be frequently changing;The external factor indicator data generation unit 13 generates the external factor indicator data in which value ranges of the data relating to the external factor are linked to external factor indicators by assigning an external factor indicator, indicating that a relationship between the data relating to the external factor in the range and the corresponding feature is constant, to the value range of the data relating to the external factor, which is classified as the constant band, and an external factor indicator, indicating that the relationship between the data relating to the external factor in the range and the corresponding feature changes frequently, to the value range of the external factor data, which is classified as the transition band, with respect to the axis of the data relating to the external factor.

[0311] Accordingly, the label generation device 1 can provide data that aids in the creation of a diagnostic model, preventing a reduction in diagnostic accuracy when the target device degrades due to the operating environment. Consequently, the label generation device 1 can contribute to the creation of a diagnostic model that prevents a reduction in diagnostic accuracy when the target device degrades due to the operating environment and can help achieve accurate degradation diagnosis with the diagnostic model. Furthermore, the label generation device 1 can help reduce the time required to design a diagnostic model.

[0312] Furthermore, when generating the tag data for the external factor, tag generation device 1 focuses on the relationship between the feature based on sensor data in a normal state of the target device and the external factor. In other words, the training data used by tag generation device 1 to generate the tag data for the external factor only needs to be data that at least corresponds to the relationship between the feature based on sensor data in a normal state of the target device and the external factor. This reduces the effort required to generate training data.

[0313] Furthermore, by supplying the data concerning the external factor, the label generation device 1 can ensure that the operator or the like clearly recognizes which external factor, for example, causes the diagnostic model to switch at what time.

[0314] The diagnostic model generation device 2 according to the first embodiment also includes: the feature generation device 1, which includes: the training data acquisition unit 11, which acquires training data in which a feature that can be used to generate a diagnostic model to be used as the object of degradation diagnosis in the degradation diagnosis of the target device is associated with data relating to the external factor relating to the operating environment;the partitioning unit 12 for a region of the external factor, which, on the basis of the training data acquired by the training data acquisition unit 11, classifies a range of values ​​of the data relating to the external factor in the axial direction of the data relating to the external factor into a constant band, which is a range of values ​​of the data relating to the external factor in which the relationship to the corresponding characteristic is considered constant, and a transition band, which is a range of values ​​of the data relating to the external factor in which the relationship to the corresponding characteristic is considered to be frequently changing;The external factor indicator data generation unit 13 generates the external factor indicator data in which value ranges of the data relating to the external factor are linked to external factor indicators by assigning an external factor indicator that indicates that a relationship between the data relating to the external factor in the range and the corresponding feature is constant, to the value range of the data relating to the external factor that is classified as the constant band, and to the value range of the data relating to the external factor that is classified as the transition band, assigns an external factor indicator that indicates that the relationship between the data relating to the external factor in the range and the corresponding feature changes frequently, with respect to the axis of the data relating to the external factor;and the diagnostic model generation unit 21, which, based on the external factor identifier data generated by the identifier data generation unit 13 in the identifier data generation device 1, acquires the training data corresponding to the value range of the data relating to the external factor associated with the identifier for the external factor, indicating that the area is a range in which the relationship between the data relating to the external factor in the area and the corresponding feature is constant, and generates a diagnostic model using the acquired training data.

[0315] By generating a diagnostic model based on the tag data for external factors generated by tag generation device 1, diagnostic model generation device 2 can create a diagnostic model that prevents a reduction in diagnostic accuracy in the event of deterioration or degradation of a device due to the operating environment.

[0316] Furthermore, according to the first embodiment, the degradation diagnostic device 3 performs a degradation diagnosis on the target device, using a diagnostic model generated by the diagnostic model generation device 2, which comprises: the feature generation device 1, which comprises: the training data acquisition unit 11, which acquires training data in which a feature that can be used to generate a diagnostic model to be used as the object of the degradation diagnosis on the target device is linked to data relating to the external factor that relates to the operating environment;the partitioning unit 12 for a region of the external factor, which, on the basis of the training data acquired by the training data acquisition unit 11, classifies a range of values ​​of the data relating to the external factor in the axial direction of the data relating to the external factor into a constant band, which is a range of values ​​of the data relating to the external factor in which the relationship to the corresponding characteristic is considered constant, and a transition band, which is a range of values ​​of the data relating to the external factor in which the relationship to the corresponding characteristic is considered to be frequently changing;The external factor indicator data generation unit 13 generates the external factor indicator data in which value ranges of the data relating to the external factor are linked to external factor indicators by assigning an external factor indicator that indicates that a relationship between the data relating to the external factor in the range and the corresponding feature is constant, to the value range of the data relating to the external factor that is classified as the constant band, and to the value range of the data relating to the external factor that is classified as the transition band, assigns an external factor indicator that indicates that the relationship between the data relating to the external factor in the range and the corresponding feature changes frequently, with respect to the axis of the data relating to the external factor;and the diagnostic model generation unit 21, which, based on the external factor identifier data generated by the identifier data generation unit 13 in the identifier data generation device 1, acquires the training data corresponding to the value range of the data relating to the external factor associated with the identifier for the external factor, indicating that the area is a range in which the relationship between the data relating to the external factor in the area and the corresponding feature is constant, and generates the diagnostic model using the acquired training data.

[0317] The degradation diagnostic device 3 according to the first embodiment comprises: the diagnostic data acquisition unit 31, which acquires diagnostic data including sensor data relating to an operating state of a target device and data relating to the external factor relating to an operating environment of the target device; the model determination unit 32, which determines a diagnostic model to be used in the degradation diagnosis based on the data relating to the external factor acquired by the diagnostic data acquisition unit 31 and data specifying a range of values ​​of the data relating to the external factor assigned to the diagnostic model; the model acquisition unit 33, which acquires the diagnostic model determined for use in the degradation diagnosis from the model determination unit 32;the feature extraction unit 34, which extracts a feature from the sensor data acquired by the diagnostic data acquisition unit 31; the diagnostic unit 35, which performs a degradation diagnosis on the target device based on the feature extracted by the feature extraction unit 34, using the diagnostic model acquired by the model acquisition unit 33 for the data relating to the external factor corresponding to the feature; and the diagnostic result output unit 36, which outputs a result of the degradation diagnosis performed by the diagnostic unit 35 on the target device.

[0318] The degradation diagnostic device 3 performs a degradation diagnosis on the target device using the diagnostic model, which was generated taking into account the possibility that an accurate diagnostic result cannot be obtained before and after switching due to the operating environment, and can thus perform an accurate degradation diagnosis.

[0319] Furthermore, the information processing system 6 according to the first embodiment comprises the identifier generation device 1, the diagnostic model generation device 2 and the degradation diagnostic device 3, which have been described above.

[0320] In this way, the information processing system 6 can provide data that aids in the creation of a diagnostic model, preventing a decrease in diagnostic accuracy due to degradation of the target device caused by the operating environment. Furthermore, the information processing system 6 can create a diagnostic model that prevents a reduction in diagnostic accuracy when the target device degrades due to the operating environment, and can use this diagnostic model to achieve an accurate degradation diagnosis.

[0321] It should be noted that in the present disclosure it is possible to modify any of the components of the embodiment or to omit any of the components in the embodiment. INDUSTRIAL APPLICABILITY

[0322] A label generation device according to the present disclosure can provide data that helps in creating a diagnostic model that prevents a reduction in diagnostic accuracy when a device degrades due to the operating environment. REFERENCE MARK LIST

[0323] 1: Tag generation device, 11: Training data acquisition unit, 12: Splitting unit for an external factor region, 121: Inflection point extraction unit, 122: Setting unit for a preliminary region, 123: Generation unit for a relational model, 124: Evaluation unit for a relational model, 125: Region determination unit, 13: Tag data generation unit for an external factor, 2: Diagnostic model generation device, 21: Diagnostic model generation unit, 211: Reliability determination unit, 3: Degradation diagnostic device, 31: Diagnostic data acquisition unit, 32: Model determination unit, 33: Model acquisition unit, 34: Feature extraction unit, 35: Diagnostic unit, 36: Diagnostic result output unit, 4: Storage device, 41: Training data storage unit, 42: External factor tag data storage unit, 43: Diagnostic model storage unit 5: Sensor, 6: Information processing system, 101: Processing circuit,102: Input interface device, 103: Output interface device, 104: Processor, 105: Memory, QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] JP 2016-91378 A

[0004]

Claims

[1] License plate generating device comprising: a training data acquisition unit for acquiring training data in which a feature to be used to generate a diagnostic model, which is used as a degradation diagnosis object in a degradation diagnosis on a target device, is linked to data concerning the external factor relating to an operating environment; a partitioning unit for a region of an external factor, in order to classify, on the basis of the training data procured by the training data acquisition unit, a range of values ​​of the data relating to the external factor in an axial direction of the data relating to the external factor into a constant band, which is a range of values ​​of the data relating to the external factor in which a relationship to the corresponding feature is considered constant, and a transition band, which is a range of values ​​of the data relating to the external factor in which the relationship to the corresponding feature is considered to change frequently; and A flag data generation unit for an external factor, to assign a flag for the external factor to the value range of the data relating to the external factor, which is classified as a constant band, indicating that a relationship between the data relating to the external factor in the range and the corresponding characteristic is constant with respect to an axis of the data relating to the external factor; to assign a flag for the external factor to the value range of the data relating to the external factor, classified as a transition band, with respect to the axis of the data relating to the external factor, indicating that the relationship between the data relating to the external factor in the range and the corresponding characteristic changes frequently; and to generate flag data for the external factor in which value ranges of the data relating to the external factor are associated with flags for the external factor. [2] Label generation device according to claim 1, wherein the partitioning unit for a region of an external factor comprises: an inflection point extraction unit to extract a plurality of inflection points at which a relationship between the feature and the data concerning the external factor changes in the axial direction of the data concerning the external factor; a setting unit for a preliminary region to select a first inflection point and a second inflection point that are adjacent to each other from the plurality of inflection points extracted by the inflection point extraction unit, and to define a preliminary region as a candidate for the constant band in an inflection point range, wherein the inflection point range is a range of values ​​of the data relating to the external factor between the selected first inflection point and the selected second inflection point; a generation unit for a relational model to generate a relational model that indicates a relationship between the feature and the data concerning the external factor based on the value of the data concerning the external factor and the value of the feature that corresponds to the value of the data concerning the external factor in the preliminary region; a rating unit for the relational model to assess the accuracy of the relational model, using the value of the data relating to the external factor contained in the preliminary range and the value of the feature corresponding to the value of the data relating to the external factor; and a region determination unit to determine, if the accuracy of the relational model is rated as low, that the preliminary region in which the accuracy of the relational model is rated as high is set as the constant band, and that another range of values ​​of the data relating to the external factor is set as the transition band, rather than the determined constant band, and to instruct, if the accuracy of the relational model is rated as low, the setting unit for the preliminary region to reset the preliminary region by narrowing the area of ​​the preliminary region, and to instruct the generating unit for the relational model to recreate the relational model, and to instruct the rating unit for the relational model to re-evaluate the accuracy of the relational model. [3] License plate generating device according to claim 1 or 2, wherein the data concerning the external factor, which includes data concerning the external factor relating to a variety of operating environments, The feature extraction unit extracts a large number of features. the division unit for the region of the external factor classifies the range of values ​​of the data relating to the external factor into the constant band and the transition band for each set of data relating to the external factor and the characteristic, and The identifier data generation unit for the external factor generates the identifier data for the external factor for each set of data relating to the external factor and the characteristic. [4] Diagnostic model generation device comprising: the license plate generating device according to one of claims 1 to 3; and a diagnostic model generation unit to obtain training data corresponding to the value range of the data relating to the external factor associated with the external factor identifier, which indicates a range in which the relationship between the data relating to the external factor in the region and the corresponding feature is constant, based on the external factor identifier data generated by the external factor identifier data generation unit, and to generate the diagnostic model using the obtained training data. [5] Diagnostic model generation device according to claim 4, wherein the diagnostic model generation unit connects the diagnostic model with data that specifies a range of values ​​of the data relating to the external factor that corresponds to the training data used in generating the diagnostic model. [6] Diagnostic model generation device according to claim 4 or 5, wherein the diagnostic model generation unit includes a reliability determination unit to determine the reliability of the diagnostic model based on the number of pieces of training data used in generating the diagnostic model. [7] Degradation diagnostic device which performs the degradation diagnosis on the target device using the diagnostic model generated by the diagnostic model generation device according to one of claims 4 to 6, wherein the diagnostic model is linked to data that specifies a range of values ​​of the data concerning the external factor, which corresponds to the training data used in generating the diagnostic model, and The degradation diagnostic device includes: a diagnostic data acquisition unit for acquiring diagnostic data including sensor data relating to an operating state of the target device and data relating to the external factor relating to the operating environment of the target device; a model determination unit for determining the diagnostic model to be used in degradation diagnosis based on the data procured by the diagnostic data acquisition unit concerning the external factor and data that specify a range of values ​​of the data assigned to the diagnostic model concerning the external factor; a model procurement unit for procuring the diagnostic model that was determined by the model determination unit for use in degradation diagnosis; a feature extraction unit for extracting a feature from the sensor data procured by the diagnostic data acquisition unit; a diagnostic unit for performing degradation diagnosis on the target device based on the feature extracted by the feature extraction unit, using the diagnostic model procured by the model acquisition unit for the data corresponding to the feature concerning the external factor; and A diagnostic output unit for outputting a result of the degradation diagnosis performed by the diagnostic unit on the target device. [8] Degradation diagnostic device according to claim 7, wherein Data indicating the reliability, which is determined based on the number of training data portions used in generating the diagnostic model, are added to the diagnostic model and The diagnostic result output unit outputs the result of the degradation diagnosis and data indicating the reliability of the diagnostic model used in the degradation diagnosis. [9] Information processing system, including: the license plate generating device according to one of claims 1 to 3; the diagnostic model generation device according to any one of claims 4 to 6; and the degradation diagnostic device according to claim 7 or 8. [10] License plate generation process, which includes: Procurement of training data by a training data acquisition unit in which a feature to be used to generate a diagnostic model, which is used as a degradation diagnostic object in a degradation diagnosis on a target device, is linked to data concerning the external factor relating to an operating environment; Classifying a range of values ​​of the data relating to the external factor in an axial direction of the data relating to the external factor, based on the training data procured by the training data acquisition unit, into a constant band, which is a range of values ​​of the data relating to the external factor, by a partition unit for a region of an external factor in which a relationship to the corresponding characteristic is assessed as constant, and into a transition band, which is a range of values ​​of the data relating to the external factor in which the relationship to the corresponding characteristic is assessed as frequently changing; and Generating indicator data for the external factor, in which value ranges of the data concerning the external factor are linked to indicators for the external factor, by assigning an indicator for the external factor, which indicates that a relationship between the data concerning the external factor in the area and the corresponding characteristic is constant, to the value range of the data concerning the external factor, which is classified as the constant band, by an indicator data generation unit, and assigning an indicator for the external factor, which indicates that the relationship between the data concerning the external factor in the area and the corresponding characteristic changes frequently, to the value range of the data concerning the external factor, which is classified as the transition band, with respect to an axis of the data concerning the external factor.

Images