Diagnosis method, diagnostic processing device, terminal, and diagnosis system

Abstract

This diagnostic processing device is provided with a processor and a communication device. The processor generates, on the basis of data obtained via the communication device from a plurality of sensors included in an apparatus in a normal operating state thereof, a prediction model for predicting a first physical quantity measured by a first sensor among the plurality of sensors. The processor uses a predicted value of the first physical quantity predicted using the prediction model and a measured value of the first physical quantity measured by the first sensor to determine abnormality of information related to the first physical quantity. If it is determined that the information is abnormal, the processor calculates a contribution value for the degree of abnormality of the information with respect to each of the sensors other than the first sensor. The processor uses the contribution values to estimate an abnormal output value of a portion of the apparatus, and estimates an abnormal portion in the apparatus.

Description

Diagnostic method, diagnostic processing device, terminal, and diagnostic system

[0001] The present invention relates to a diagnostic method, a diagnostic processing device, a terminal, and a diagnostic system.

[0002] There are known techniques for estimating power consumption and predicting component failures. Patent Literature 1 discloses a power consumption estimation device that estimates the power consumption of each device or each application from the total power consumption by using nonnegative matrix factorization.

[0003] Patent Document 2 focuses on the pattern of harmonic currents generated by electrical equipment installed at power consumers, and discloses a technology for determining the phase difference between the fundamental wave of the total load current and the harmonic current and voltage from the total load current and voltage measured near the power feeder entrance, and estimating the electrical equipment used indoors and the power consumption of each individual electrical equipment from the pattern.

[0004] Patent Literature 3 discloses a failure prediction device equipped with a machine learning device including: a state observation unit that observes operation status data indicating the operation status of managed devices and device configuration data indicating the device configuration of the managed devices as state variables representing the current state of the environment, a label data acquisition unit that acquires maintenance history data indicating the maintenance history of the managed devices as label data, and a learning unit that uses the state variables and the label data to associate and learn the failure timing of printed circuit boards that constitute the managed devices with the operation status data and the device configuration data. Patent Literature 3 discloses that highly accurate failure prediction is possible by updating a failure estimation model as needed through machine learning.

[0005] Japanese Patent Application Laid-Open No. 2020-205684 International Publication No. 2001 / 077696 Japanese Patent Application Laid-Open No. 2019-8675

[0006] A known maintenance method involves monitoring the status of equipment (devices) in real time and performing maintenance based on the status in order to prevent breakdowns. This method is sometimes called condition-based maintenance (CBM). Based on the CBM approach, identifying a part of a device that is producing an abnormal output (e.g., a part that is producing more output than normal) can provide the following advantages, for example. For example, adjusting the output of the part in question can enable the use of the device with reduced running costs. Furthermore, for example, it can be possible to make the device usable by replacing only the part in question, without replacing the entire device.

[0007] Therefore, there is a need to provide a technique for evaluating a part of a device that produces an abnormal output.

[0008] According to a first aspect of the present invention, there is provided a diagnostic method as described below. The diagnostic method includes a prediction model generating step, an abnormality determination step, a contribution value calculating step, and an estimation step. In the prediction model generating step, a processor generates a prediction model that predicts a first physical quantity measured by a first sensor of a plurality of sensors included in the device based on data obtained from the plurality of sensors when the device is operating normally. In the abnormality determination step, the processor determines an abnormality in information related to the first physical quantity by using a predicted value of the first physical quantity predicted using the prediction model and an actual measured value of the first physical quantity measured by the first sensor. In the contribution value calculating step, if the processor determines that the information is abnormal, the processor calculates a contribution value to the abnormality degree of the information for each sensor other than the first sensor. In the estimation step, the processor uses the contribution values ​​to estimate an abnormal output value of a part of the device and an abnormal part in the device.

[0009] According to a second aspect of the present invention, there is provided a diagnostic processing device. The diagnostic processing device includes a processor and a communication device. The processor generates a prediction model for predicting a first physical quantity measured by a first sensor of a plurality of sensors included in the device, based on data obtained via the communication device from the plurality of sensors included in the device when the device is operating normally. The processor determines an abnormality in information related to the first physical quantity using a predicted value of the first physical quantity predicted using the prediction model and an actual measured value of the first physical quantity measured by the first sensor. If the processor determines that the information is abnormal, it calculates a contribution value to the abnormality degree of the information for each sensor other than the first sensor. The processor uses the contribution value to estimate an abnormal output value of a portion of the device and estimate an abnormal portion of the device. Also provided is, for example, a terminal that provides a user with information based on data input from the diagnostic processing device via communication. Also provided is, for example, a diagnostic system that includes a diagnostic processing device that outputs information related to the estimated abnormal portion of the device to an external device via the communication device, and a terminal that provides a user with information based on the data input from the diagnostic processing device. Furthermore, a diagnostic processing device may be provided that determines whether the information relating to the first physical quantity is abnormal using a predetermined threshold value.

[0010] According to the present invention, a technique for evaluating a part of a device that produces an abnormal output is provided. Note that problems, configurations, and effects other than those described above will become apparent from the following description of the preferred embodiment of the invention.

[0011] FIG. 1 is a diagram illustrating an example of the configuration of a diagnostic system. FIG. 2 is a flowchart illustrating an example of overall processing in this embodiment. FIG. 3 is a diagram illustrating an example of setting a virtual current based on a current abnormality degree. FIG. 4 is a diagram illustrating an example of a correspondence table linking parts of an apparatus with sensors. FIG. 5 is a diagram illustrating an example of information provided to a user.

[0012] Embodiments of the present invention will be described below with reference to the drawings. The embodiments are merely examples for explaining the present invention, and appropriate omissions and simplifications have been made for clarity of explanation. The present invention can be implemented in various other forms. Unless otherwise specified, each component may be singular or plural. The position, size, shape, and range of each component shown in the drawings may not represent the actual position, size, shape, and range in order to facilitate understanding of the invention. Therefore, the present invention is not necessarily limited to the position, size, shape, and range disclosed in the drawings. While various types of information may be described using terms such as "table," "list," and "queue," various types of information may also be expressed using other data structures. For example, various types of information such as "XX table," "XX list," and "XX queue" may also be referred to as "XX information." When describing identification information, terms such as "identification information," "identifier," "name," "ID," and "number" are used, but these terms are interchangeable. When there are multiple components having the same or similar functions, they may be described using the same reference numeral with different subscripts. Furthermore, when it is not necessary to distinguish between these multiple components, the subscripts may be omitted. In some embodiments, processing performed by executing a program will be described. Here, a computer executes the program using a processor (e.g., a CPU or a GPU) and performs processing defined by the program using storage resources (e.g., memory) and interface devices (e.g., communication ports). Therefore, the entity performing the processing by executing the program may be the processor. Similarly, the entity performing the processing by executing the program may be a controller, device, system, computer, or node having a processor. The entity performing the processing by executing the program may be any computing unit, and may include a dedicated circuit that performs specific processing. Here, the dedicated circuit may be, for example, an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), or a CPLD (Complex Programmable Logic Device).A program may be installed on a computer from a program source. The program source may be, for example, a program distribution server or a computer-readable storage medium. When the program source is a program distribution server, the program distribution server may include a processor and storage resources for storing the program to be distributed, and the processor of the program distribution server may distribute the program to be distributed to other computers. In addition, in an embodiment, two or more programs may be realized as one program, or one program may be realized as two or more programs.

[0013] In the embodiments, an example of a technology for quantitatively estimating an abnormal output value of a part of the equipment and estimating an abnormal part of the equipment (a part producing an abnormal output) by performing a causal analysis when an abnormality in a sensor value is determined based on data acquired from an equipment (device) equipped with multiple sensors will be described. Note that in the embodiments, an example of acquiring a current value from a current sensor and performing a causal analysis will be described in detail. However, the abnormal part of the equipment may also be estimated by acquiring sensor values ​​from other types of sensors and performing a causal analysis based on a similar technique.

[0014] First, an example of a diagnostic system will be described with reference to Fig. 1. As shown in Fig. 1, the diagnostic system 101 includes a diagnostic processing device 111 and a device 131 (apparatus). In this embodiment, the diagnostic processing device 111 is a computer that exists in a cloud environment, but it may also be a computer that exists in an on-premises environment.

[0015] The diagnostic system 101 also includes an information output device 151 used to provide information to a user. The information output device 151 is, for example, a user terminal (terminal). This user terminal is configured appropriately and is connected to the diagnostic processing device 111 via communication. The diagnostic processing device 111 then outputs data related to the processing results to the information output device 151.

[0016] The diagnostic processing device 111 includes a processor 112, a memory 113, a storage 114, and a communication device 115. The processor 112 is a main body that executes predetermined processing, and is, for example, a CPU (Central Processing Unit). However, as long as it is capable of executing the predetermined processing, the processor 112 may be configured using other semiconductor devices.

[0017] The memory 113 is a main storage device, such as a RAM (Random Access Memory). The processor 112 stores data in the memory 113 during processing.

[0018] The storage 114 is an auxiliary storage device, such as a hard disk drive (HDD), and can store, for example, data acquired from outside, processing results, programs, and the like.

[0019] The device 131 is, for example, an industrial machine installed in a factory, and examples of the type of device include a compressor, a water supply unit, an air shower, an inkjet printer, a hoist, etc. The device 131 is provided with various sensors according to the type, and the device 131 is configured with a sensor unit 132 that includes the sensors provided in the device 131.

[0020] In this embodiment, the diagnostic processing device 111 is configured to be able to acquire various information in the diagnostic system 101. The diagnostic processing device 111 collects, for example, sensor values ​​from sensors provided in the devices 131 via a network.

[0021] The diagnostic processing device 111 also includes, as its functional configuration, a sensor value acquisition unit 121, a prediction model generation unit 122, an abnormality determination unit 123, a contribution value calculation unit 124, an abnormality portion estimation unit 125, and an information output unit 126. Each of these units (121 to 126) is realized by the processor 112 storing a predetermined program in the memory 113 and executing data processing.

[0022] The sensor value acquisition unit 121 acquires the sensor values ​​of the sensors from each of the devices 131 .

[0023] The prediction model generation unit 122 generates a prediction model that predicts the normal value of the first sensor of the device 131. Here, the prediction model is a model that receives a sensor value of a sensor other than the first sensor of the device 131 as input and outputs a predicted value of the first sensor. The prediction model generation unit 122 generates the prediction model by performing a learning process using the sensor value of the device 131 in a normal operating state as training data. Note that in this embodiment, the first sensor is a current sensor, and the prediction model generation unit 122 generates a prediction model that predicts the current value output by the current sensor. Note that when the diagnostic system 101 includes multiple devices 131, the diagnostic processing device 111 can generate a prediction model for each of the devices 131.

[0024] The abnormality determination unit 123 acquires operation data of the device 131 and acquires an actual measurement value of the first sensor based on the operation data. The abnormality determination unit 123 also inputs operation data of a sensor other than the first sensor into a prediction model and acquires a predicted value of the first sensor.

[0025] The abnormality determination unit 123 determines whether the information about the first physical quantity (information about a value based on the actual measurement value and the predicted value of the first sensor) is abnormal based on the actual measurement value and the predicted value of the first sensor. In this embodiment, the abnormality determination unit 123 determines whether the information about the current sensor is abnormal.

[0026] When an abnormality is determined, the contribution value calculation unit 124 performs a factor analysis. Here, the factor analysis is a process of analyzing the influence of other sensors on the abnormality level of the first sensor in the device 131, and will be described in detail later. In this embodiment, the influence of other sensors on the abnormality level of the current sensor is analyzed.

[0027] Based on the result of the factor analysis, the abnormal part inferring unit 125 infers an abnormal part in the device 131. The processing of the abnormal part inferring unit 125 will be described in detail later.

[0028] The information output unit 126 generates and outputs data for providing information on an abnormal portion in the device 131. The processing of the information output unit 126 will be described in detail later.

[0029] Next, an example of the overall processing will be described with reference to Fig. 2. Fig. 2 is a flowchart showing an example of the overall processing in this embodiment. In particular, the processing is performed by the processor 112 of the diagnostic processing device 111.

[0030] The diagnostic processing device 111 acquires the sensor values ​​of each sensor from the sensor unit 132 of the device 131 in the sensor value acquisition unit 121 (S201). The timing of acquiring the sensor values ​​is not particularly limited as long as appropriate processing can be performed. For example, the diagnostic processing device 111 may acquire the sensor values ​​in real time while the device 131 is operating.

[0031] The diagnostic processing device 111 extracts normal data from the sensor value data acquired in S201 in the prediction model generation unit 122 (S202). The diagnostic processing device 111 can extract the normal data based on an appropriate method. For example, the diagnostic processing device 111 may store data indicating normal sensor values ​​of each sensor provided in the device 131 in the storage 114 in advance, and extract data similar to the data stored in the storage 114 from the data acquired in S201. The data stored in advance in the storage 114 may be, for example, data for the first year after delivery of the device 131, with no abnormality alarms.

[0032] In the diagnostic processing device 111, the prediction model generation unit 122 uses the normal data extracted in S202 as training data to generate a prediction model that predicts the sensor value of the first sensor by inputting sensor values ​​of multiple sensors other than the first sensor (S203).

[0033] Here, the first sensor is one of the sensors included in the device 131. Accordingly, in S202, normal data from the first sensor and each sensor different from the first sensor is extracted, and in S203, a prediction model for predicting the sensor value of the first sensor is generated. Note that in this embodiment, the first sensor is a current sensor, and a prediction model for predicting the sensor value of the current sensor as the first physical quantity is generated.

[0034] When a predictive model is generated and the diagnostic processing device 111 performs diagnostic processing on the equipment 131, the diagnostic processing device 111, in the abnormality determination unit 123, acquires the sensor values ​​of each sensor equipped in the equipment 131 that is the target of diagnostic processing from the sensor value data acquired in S201 as data to be analyzed (S204).

[0035] The diagnostic processing device 111 acquires, in the abnormality determination unit 123, the prediction model for the device 131 that is the target of the diagnostic processing generated in S203 (S205). Then, the diagnostic processing device 111 inputs the sensor values ​​of a sensor other than the first sensor from the analysis target data acquired in S204 into the prediction model, and acquires an estimated output value of the first sensor for the device 131 that is the target of the diagnostic processing (S206). In this embodiment, the first sensor is a current sensor, and the diagnostic processing device 111 acquires the estimated output value of the current sensor as the first physical quantity.

[0036] Then, the diagnostic processing device 111, in the abnormality determination unit 123, performs an abnormality determination for the device 131 that is the target of the diagnostic processing (S207). This abnormality determination is performed using the estimated value of the first sensor obtained from the prediction model and the actual measurement value of the first sensor. Here, the actual measurement value of the first sensor (i.e., the actual measurement value of the first physical quantity) is obtained from the analysis target data obtained in S204. In this embodiment, the first sensor is a current sensor, and therefore the diagnostic processing device 111 obtains the actual measurement value of the current sensor as the first physical quantity.

[0037] In this embodiment, the diagnostic processing device 111 determines that the information regarding the first physical quantity is normal if the actual measurement value minus the predicted value (the value obtained by subtracting the predicted value from the actual measurement value) falls within a predetermined range, and determines that the information regarding the first physical quantity is abnormal if the value does not fall within the predetermined range.

[0038] For example, the diagnostic processing device 111 uses a value based on the prediction model as a reference point in the state space, and determines that the information related to the first physical quantity is normal when an actual measurement value is present within a predetermined distance from the reference point along an axis related to the first sensor. On the other hand, the diagnostic processing device 111 determines that the information related to the first physical quantity is abnormal when an actual measurement value is not present within a predetermined distance from the reference point along an axis related to the first sensor.

[0039] Then, the diagnostic processing device 111 outputs the result of the abnormality determination made in S207 in the abnormality determination unit 123 (S208). In this embodiment, since the first sensor is a current sensor, the determination result based on the value of the current sensor is output.

[0040] If an abnormality is determined, the diagnostic processing device 111 defines the difference between the actual measurement value and the estimated output value of the prediction model as the abnormality degree in the contribution value calculation unit 124, and evaluates the influence of the other sensor values ​​on the abnormality degree of the first sensor (S209). If a normality is determined, the process of S209 is omitted.

[0041] In this embodiment, the diagnostic processing device 111 uses SHAP (Shapley Additive exPlanation), an example of an analytical method for machine learning models, to evaluate the influence of other sensor values ​​on the abnormality level of the first sensor and calculate the contribution value of each of the other sensors to the abnormality level of the first sensor.

[0042] For example, in a state space such as VQC (Vector Quantization Clustering) or LSC (Local Subspace Classifier), the degree of anomaly can be calculated from the difference in position coordinates in the state space. For example, in an N-dimensional space in which axes corresponding to each sensor of the device 131 are set, the degree of anomaly for the first sensor is defined based on the distance between a reference point along the axis for the first sensor and a point indicating the actual measurement value. Here, it is not possible to estimate the breakdown of the degree of anomaly (values ​​obtained by allocating the degree of anomaly to each sensor) from the N-dimensional space. On the other hand, by calculating the SHAP for the degree of anomaly of the first sensor as described above without calculating the SHAP for the first sensor, it is possible to calculate the breakdown of the degree of anomaly.

[0043] In S209, after evaluating the influence of the values ​​of the other sensors on the abnormality level of the first sensor, the diagnostic processing device 111 performs processing to estimate an abnormal part in the device 131. First, the diagnostic processing device 111 estimates an abnormal output value by setting a virtual value for the abnormality level of the first sensor for each sensor using the contribution values ​​of each of the other sensors to the abnormality level of the first sensor in the abnormal part estimation unit 125 (S210). Note that the virtual value may be the contribution value itself, or may be a value that represents the influence on the abnormality level based on the contribution value.

[0044] In this embodiment, since the first sensor is a current sensor, a virtual current value for the degree of current abnormality is set as shown in FIG. 3. In the example of FIG. 3, the contribution value of each sensor is calculated in S209, and a virtual value (virtual current value) is set using the contribution values ​​of these sensors in S210, indicating that the sensor values ​​of sensors A to D among these sensors affect the degree of current abnormality. In other words, it is estimated that the sensor values ​​of sensors A to D are abnormal output values. Here, the virtual current value may be, for example, the contribution value itself for the degree of current abnormality. Alternatively, the virtual current value may be, for example, a value representing the influence on the degree of current abnormality based on the contribution value.

[0045] The diagnostic processing device 111 then estimates the abnormal part by performing an analysis based on the virtual value in the abnormal part estimation unit 125. As an example, as shown in Fig. 4, a correspondence table containing information on parts of the device 131 and sensors associated with the parts is held in the storage 114 of the diagnostic processing device 111. The diagnostic processing device 111 then refers to the correspondence table and identifies, as the abnormal part, a part associated with a sensor that is evaluated to be outputting a sensor value that affects the degree of abnormality.

[0046] 3 and 4, if sensor A is a sensor that detects tank pressure, sensor B is a sensor that detects discharge pressure, sensor C is a sensor that detects fan frequency, and part A is a separate element filter, the diagnostic processing device 111 will determine that the sensor values ​​of the sensors (sensors A to C) affect the degree of abnormality and identify part A (separate element filter) to which these sensors A to C are linked as the abnormal part. Also, if sensor D is a sensor that detects motor surface temperature, sensor E is a sensor that detects fan frequency, and part B is a bearing unit, the diagnostic processing device 111 will determine that the sensor value of sensor (sensor D) affects the degree of abnormality and identify part B (bearing unit) to which sensor (sensor D) is linked as the abnormal part.

[0047] In this way, by using a correspondence table (data linking parts of the equipment 131 with each sensor different from the first sensor), it becomes possible to appropriately estimate the abnormal part even when, for example, multiple sensors are associated with one part.

[0048] Furthermore, the diagnostic processing device 111 generates data indicating the results of the diagnostic processing in the information output unit 126 and outputs the data to the information output device 151 (S211). The user can grasp the results of the diagnostic processing using the information output device 151. The diagnostic processing device 111 generates and outputs data for providing information about an abnormal portion, for example. The diagnostic processing device 111 also generates and outputs data for providing information about the influence of the abnormal portion on the degree of abnormality, using the contribution values ​​of each sensor related to the abnormal portion.

[0049] An example of the information to be provided is shown in Fig. 5. In the example of Fig. 5, the first sensor is a current sensor, and sensors A to D and parts A and B are the same as those described above.

[0050] Therefore, in this example, the user can determine that the separate element filter (site A) is the abnormal site. The user can also determine the excess power consumption (i.e., the abnormal output value) of the separate element filter. The excess power consumption value (9A in FIG. 5 ) of the separate element filter is a value based on the virtual current values ​​set for sensors A to C associated with the separate element filter using the contribution values ​​of sensors A to C. The user can also determine that the bearing unit (site B) is the abnormal site. The user can also determine the excess power consumption (i.e., the abnormal output value) of the bearing unit. The excess power consumption value (2A in FIG. 5 ) of the bearing unit is a value based on the virtual current values ​​set for sensors D to E associated with the bearing unit using the contribution values ​​of sensors D to E.

[0051] Information is provided to the user in an appropriate manner. For example, the information output device 151 may display the results of the diagnostic processing on a display included in the information output device 151. The display may be similar to the example shown in FIG. 5 , for example. That is, as shown in FIG. 5 , a display may be provided that includes a graph showing the influence of each sensor on the abnormality level, information indicating the abnormal part, and text explaining the abnormal output value of the part. However, as long as the results of the diagnostic processing can be appropriately presented, the display manner can be changed as appropriate. For example, the influence of the abnormal part (power consumption in the example shown in FIG. 5 ) may be displayed as a percentage of the overall abnormality level. For example, the influence of the abnormal part on the overall abnormality level may be displayed as a percentage. Furthermore, the results of the diagnostic processing may include, for example, the amount of power consumption (electricity bill) for each part. The amount of power consumption can be calculated using a known method using power consumption values.

[0052] For example, the result of the diagnostic process may be provided by voice, and the information output device 151 may output the result of the diagnostic process from a voice output device included in the information output device 151.

[0053] Also, for example, the provision of information regarding the influence of each sensor on the abnormality level may be omitted, and information regarding the abnormal portion may be provided instead.

[0054] If normality is determined in S208 and S209 is omitted, the virtual value is not set, and the diagnostic processing device 111 generates data for providing a diagnosis result that the device 131 is normal in the information output unit 126, and outputs the data to the information output device 151. Then, the information output device 151 can provide information indicating normality.

[0055] According to the diagnostic system 101 of this embodiment, the diagnostic processing device 111 monitors the state of the device 131, and when, for example, the first sensor value of the device 131 is abnormal, the diagnostic processing device 111 evaluates the degree of abnormality for each part of the device 131, identifies the abnormal part of the device 131, and provides a service that provides information on the abnormal part of the device 131. Then, maintenance is performed based on the provided information, and improvements are made to, for example, energy costs and running costs. Therefore, the diagnostic system 101 can contribute from an environmental perspective and an economic perspective.

[0056] For example, by evaluating each part for the degree of current anomaly, parts that are consuming more current than normal are identified and quantitatively evaluated. The results are then provided, and maintenance measures are taken based on the information provided, making it possible to use the device 131 with reduced power consumption.

[0057] Furthermore, according to this embodiment, there is no need to create a learning model for each part of the device 131, thereby reducing the calculation load.

[0058] Although the embodiments have been described above, the present invention is not limited to the above-described embodiments and includes various modifications and equivalent configurations within the spirit of the appended claims. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, for example, other configurations may be added, deleted, or replaced with part of the configuration of the embodiment.

[0059] In the diagnostic system 101, there may be a single device 131 or a plurality of devices 131. Furthermore, the devices 131 may be of the same type or different types.

[0060] The contribution value may be any value that allows appropriate evaluation of the influence on the abnormality level, and may be, for example, a contribution rate (i.e., a value related to the proportion to the abnormality level).

[0061] In the present embodiment, an example has been described in which an industrial machine is used in the diagnostic system 101. However, the device 131 in the diagnostic system 101 is not limited to an industrial machine. In the diagnostic system 101, for example, the device 131 to be diagnosed may be a household device (such as a home appliance).

[0062] In the abnormality determination unit 123 in S204 to S207, the diagnostic processing device 111 may determine whether the acquired actual measurement value is equal to or greater than a predetermined threshold. Here, the predetermined threshold can be set appropriately. For example, the predetermined threshold may be set based on the average value of the past abnormality levels.

[0063] 101 diagnostic system 111 diagnostic processing device 131 device 151 information output device

Claims

1. A diagnostic method comprising: a prediction model generation step in which a processor generates a prediction model for predicting a first physical quantity measured by a first sensor of a plurality of sensors equipped in the apparatus based on data obtained from the plurality of sensors when the apparatus is in a normal operating state; an anomaly determination step in which the processor determines an anomaly in information regarding the first physical quantity using a predicted value of the first physical quantity predicted using the prediction model and an actual value of the first physical quantity measured by the first sensor; a contribution value calculation step in which the processor, when determining that the information is abnormal, calculates a contribution value to the degree of anomaly of the information for each sensor other than the first sensor; and an estimation step in which the processor uses the contribution value to estimate an abnormal output value of a portion of the apparatus and estimates an abnormal portion of the apparatus.

2. A diagnostic method according to claim 1, further comprising an output step in which the processor outputs data for providing information regarding the abnormal portion in the device estimated in the estimation step.

3. A diagnostic method as claimed in claim 2, characterized in that in the output step, the processor outputs data for providing information on the abnormal output value of the abnormal part, the data being generated using the contribution values ​​of each sensor related to the abnormal part.

4. A diagnostic method as described in claim 1, characterized in that in the estimation step, a processor estimates the abnormal part using data associated with a part of the device and a sensor other than the first sensor.

5. A diagnostic method according to claim 4, further comprising an output step in which the processor outputs data for providing information regarding the abnormal portion in the device estimated in the estimation step.

6. A diagnostic method as claimed in claim 5, characterized in that in the output step, the processor outputs data for providing information on the abnormal output value of the abnormal part, the data being generated using the contribution values ​​of each sensor related to the abnormal part.

7. A diagnostic processing device comprising: a processor; and a communication device, wherein the processor generates a prediction model that predicts a first physical quantity measured by a first sensor of a plurality of sensors equipped in the device based on data obtained via the communication device from the plurality of sensors equipped in the device when the device is in a normal operating state, the processor determines an abnormality in information related to the first physical quantity using a predicted value of the first physical quantity predicted using the prediction model and an actual measured value of the first physical quantity measured by the first sensor, and when the processor determines that the information is abnormal, calculates a contribution value to the degree of abnormality of the information for each sensor other than the first sensor, and the processor uses the contribution value to estimate an abnormal output value of a part of the device and estimates an abnormal part in the device.

8. A diagnostic processing device according to claim 7, characterized in that the processor outputs data for providing information relating to the estimated abnormal part in the device to the outside via the communication device.

9. A diagnostic processing device as claimed in claim 8, characterized in that the processor outputs, via the communication device, data for providing information relating to the abnormal output value of the abnormal part, the data being generated using the contribution values ​​of each sensor related to the abnormal part.

10. A diagnostic processing device according to claim 7, characterized in that the processor estimates an abnormal part using data in which a part of the device is linked to a sensor other than the first sensor.

11. A diagnostic processing device according to claim 10, characterized in that said processor outputs data for providing information relating to an estimated abnormal part in said device to the outside via said communication device.

12. A diagnostic processing device as described in claim 11, characterized in that the processor outputs data to the outside via the communication device, the data being generated using the contribution values ​​of each sensor related to the abnormal area, to provide information regarding the abnormal output value of the abnormal area.

13. A terminal which provides a user with information based on data input from the diagnostic processing device according to claim 8.

14. A diagnostic system comprising: a diagnostic processing device according to claim 8; and a terminal for providing a user with information based on data input from said diagnostic processing device.

15. A diagnostic processing device comprising: a processor; and a communication device, wherein the processor acquires an actual measurement value of a first physical quantity measured by a first sensor of a plurality of sensors equipped in the device based on data obtained from the plurality of sensors via the communication device, the processor determines an abnormality of information regarding the first physical quantity based on whether the acquired actual measurement value is equal to or greater than a predetermined threshold, and when the processor determines that the information is abnormal, calculates a contribution value to the degree of abnormality of the information for each sensor other than the first sensor, and the processor uses the contribution value to estimate an abnormal output value of a part of the device and estimate an abnormal part in the device.

Images