TW201730085A - Apparatus classification device using the value of the feature quantity to classify equipment by an apparatus unit - Google Patents

Abstract

This invention is an apparatus classification device in which a data acquisition unit (101) obtains the apparatus classification index data from an apparatus classification index database (201). A feature quantity conversion unit (102) converts the apparatus classification index data into a feature quantity indicating a feature of an apparatus. An apparatus classification unit (103) uses the value of the feature quantity to classify equipment by an apparatus unit.

Description

Machine sorting device

The present invention relates to a machine sorting device for classifying devices in units of machines constituting the device.

It is useful to classify devices having the same characteristics in a device in which a lift or an air conditioner equivalent is present in a plurality of environments. For example, in the conventional system described in Patent Document 1, the purpose of the construction equipment is to save energy. In the air conditioning control, the elevators are classified according to the same characteristics. The system described in Patent Document 1 uses the elevator operation information to pattern the flow rate of the shared portion, the number of people in the cabin, and the existence rate on a certain day or every time interval, and plan the control schedule. Here, for the construction in which the information on the operation of the elevator cannot be obtained, the classification of the building is carried out in order to transfer the analysis results of similar structures holding the same features.

[Previous Technical Literature] [Patent Literature]

Patent Document 1 Japanese Patent Laid-Open Publication No. 2005-104635

When a fault or abnormality is analyzed for a machine constituting a device such as an elevator or an air conditioner, the machine of the same type and the same feature is used for a plurality of devices. The classification and analysis of the machine is expected to be compared with the analysis of only a single device. Acceleration accuracy of faults or abnormalities. However, in the conventional method, since it is classified according to the equipment unit such as a lift, even if the machine holding different characteristics is the same kind and the same feature in terms of the equipment unit, there is a problem that the machine cannot be classified. . For example, in the case where the door device constituting the elevator A and the door device constituting the elevator B are of the same type but different in characteristics, it is conventionally known that when the elevator A and the elevator B are determined to have the same characteristics as a device called an elevator, both The door machine will also be classified as the same feature. Therefore, for example, in the analysis of a malfunction or abnormality of a machine, a machine that holds a feature that is abnormal under a certain condition and a machine that has a feature that has no abnormality under the same condition may be treated as the same classification, and the like. Problems related to the analysis of abnormal cause or the accuracy of abnormality detection.

The present invention has been made to solve the above problems, and an object thereof is to provide a machine sorting device capable of performing analysis of a failure or an abnormality of a machine with high accuracy.

According to the machine classification device of the present invention, the data acquisition unit includes the device classification index data, which is equivalent to each device obtained from the monitoring data of each of the plurality of devices including the singular or plural devices. Information; and the Machine Classification Department, which uses machine classification indicator data to classify equipment by machine unit.

According to the machine sorting apparatus of the present invention, the machine classification index data is used, and the equipment is classified by machine using the machine classification index data corresponding to the information inherent to each machine. Thereby, the analysis of the malfunction or abnormality of the machine can be performed with high accuracy.

11‧‧‧ Processor

12‧‧‧ memory

13‧‧‧Communication I/F device

14‧‧‧ memory

15‧‧‧Output device

100, 100a, 100b, 100c‧‧‧ machine sorting devices

101, 101a‧‧‧Information Acquisition Department

102‧‧‧Characteristics Conversion Department

103, 103a‧‧‧Machine Classification Department

200, 200a, 200b‧‧‧ data collection and management device

201, 201a‧‧‧ Machine Classification Indicators Database

202‧‧‧ Machine Classification Indicators Database Group

202a‧‧‧Sensor database

202b‧‧‧Maintenance Performance Database

300‧‧‧Network

400‧‧‧Monitoring objects

500‧‧‧Fault risk calculation department

600‧‧‧ input device

701, 702, 703‧‧‧ feature magnitude

Classification of 704, 705‧‧

901, 902‧‧‧ equipment feature quantity

Fig. 1 is a configuration diagram of a machine sorting apparatus according to a first embodiment of the present invention.

Fig. 2 is an explanatory view showing sensor data used in the machine sorting device of the first embodiment of the present invention.

Fig. 3 is an explanatory view showing maintenance performance data used in the machine sorting device of the first embodiment of the present invention.

Fig. 4 is a view showing the hardware configuration of the machine sorting apparatus of the first embodiment of the present invention.

Fig. 5 is a flow chart showing the machine sorting process of the machine sorting apparatus according to the first embodiment of the present invention.

Fig. 6 is an explanatory view showing a feature table used in the machine sorting device of the first embodiment of the present invention.

Fig. 7 is an explanatory view showing the classification of the feature amounts of the machine sorting device according to the first embodiment of the present invention.

Fig. 8 is a configuration diagram of an additional failure risk prediction device in the case of the machine classification device of the first embodiment of the present invention.

Fig. 9 is an explanatory view showing the feature amounts of the respective machines of the machine sorting device according to the first embodiment of the present invention.

Fig. 10 is a view showing the configuration of a machine sorting apparatus according to a second embodiment of the present invention.

Fig. 11 is a view showing the configuration of a machine sorting apparatus according to a second embodiment of the present invention.

Hereinafter, the present invention will be described in more detail, and the form for carrying out the invention will be described based on the drawings.

Implementation type 1

Fig. 1 is a configuration diagram of a monitoring system including a machine sorting device 100 according to the present embodiment. In the illustrated monitoring system, the device sorting device 100 is connected to the data collection management device 200, and the data collection management device 200 is connected to the monitoring target 400 via the network 300.

The machine classification device 100 includes a material acquisition unit 101, a feature amount conversion unit 102, and a device classification unit 103. The data acquisition unit 101 is a processing unit that acquires the device classification index data from the device classification index database 201 managed by the data collection management device 200. The feature amount conversion unit 102 is a processing unit that converts the machine classification index data into a feature amount indicating the characteristics of the machine. The machine classifying unit 103 uses the value of the feature amount converted by the feature amount converting unit 102 to classify the processing unit of the device by machine.

The data collection management device 200 collects monitoring data from the monitoring target 400, and accumulates the monitoring data as machine classification index data in the machine classification index database 201 and manages the monitoring data. The machine classification index data accumulated in the machine classification index database 201 indicates the data (sensor data) obtained from the sensor of the monitoring object 400, the data (maintenance performance data), etc. based on the inspection by the maintenance staff or the equipment information. Monitoring data obtained directly or indirectly from the monitoring object 400. As the machine classification index data accumulated in the machine classification index database 201, an elevator is taken as an example, an example of the sensor data is shown in Fig. 2, and an example of the maintenance performance data is shown in Fig. 3.

In Figure 2, the sensor as a machine from a device An example of the obtained data, exemplifying the sensor data. In the sensor data shown in the figure, as an example of the data item, temperature, vibration, rotation speed, contact 1 current, contact 1 voltage, contact 2 current, contact 2 voltage, and the like are indicated. The value of the data item example of the sensor data is an example. The data item can be changed in order to store the items of the sensor data collected from the actual equipment and the machine. In addition, if you can distinguish between equipment and machines, you can also collect data from a number of equipments and machines and make a table. If the connection of the machines is possible, the data of one machine of one device can also be divided into a plurality of tables. The data items common to each machine such as temperature and humidity can also be managed by a table other than the machine data.

In Fig. 3, maintenance performance data is exemplified as an example of information obtained from inspection of a machine self-maintenance person or equipment information of a device. In the example of the maintenance performance data, as an example of the data item, the device ID, the model ID, the device ID, the operator name, the maintenance work content, and the presence or absence of an abnormality are described. The value of these data items is an example. The data item can be changed in order to store items of maintenance performance data collected from actual equipment and equipment. In addition, if you can distinguish between equipment and machines, you can also collect data from a number of equipments and machines and make a table. Furthermore, if the association of the equipments of the equipment is possible, the data of one machine of one equipment can be divided into a plurality of tables. In addition, it is also possible to divide and maintain the maintenance performance data of maintenance work that is different from the maintenance work at the time of normal operation and the maintenance work at the time of abnormality. In addition, the sensor data example shown in FIG. 2 and the maintenance performance data example shown in FIG. 3 may be managed as one table. That is, the machine classification index data stored in the machine classification index database 201 may be any information if it is information inherent to the machine.

The monitoring object 400 is, for example, a lift or an air conditioner. A device consisting of a single or multiple machines. The monitoring object 400 is assumed to have more than two devices consisting of the same machine. It is also possible to directly connect the monitoring object 400 and the material collection management device 200 without being connected to the network 300. Regardless of the connection mode of the monitoring object 400 and the data collection management device 200, the network connection data collection management device 200 and the device classification device 100 may be configured.

Fig. 4 is a block diagram showing a hardware configuration for realizing the machine sorting apparatus of the present embodiment. In Fig. 4, the machine sorting device 100 and the data collection management device 200 of Fig. 1 are illustrated as being formed on one piece of hardware. The machine classification device 100 and the data collection management device 200 include a processor 11, a memory 12, a communication I/F (interface) device 13, a memory 14, and an output device 15. The processor 11 is a processor for realizing the functions of the machine sorting device 100 and the material collection management device 200. The memory 12 is a program memory that stores various programs corresponding to the functions of the device sorting device 100 and the data collection management device 200, a working memory used for data processing by the processor 11, and a memory for expanding the signal data. Memory unit such as ROM and RAM used. The communication I/F device 13 is an external communication interface with the network 300 or the like. The memory 14 is a memory device for accumulating various materials and programs. The output device 15 is a device for outputting the processing result to the outside.

The processing performed by the material acquisition unit 101, the feature amount conversion unit 102, and the device classification unit 103 in Fig. 1 is executed by the processor 11 reading the program stored in the memory 12. The data accumulated by the machine classification index database 201 is stored in the memory 14 from the monitoring object 400 via the network 300 via the communication I/F device 13. The processing result of the machine classifying unit 103 is stored in the memory if necessary. And outputted to the outside by the output device 15. Further, the machine classification device 100 and the material collection management device 200 may be configured on different hardware.

Next, the operation of the machine sorting apparatus 100 of the present embodiment will be described. The data collection management device 200 continuously or intermittently accumulates the machine classification index data obtained from the monitoring target 400 to the machine classification index database 201. The machine classification device 100 performs processing for acquiring the machine classification index data from the machine classification index database 201. Fig. 5 is a flow chart showing the processing of the machine sorting device 100. First, the material acquisition unit 101 acquires the device classification index data from the device classification index database 201 (step ST1). In addition, when a plurality of data items are included in the machine classification index data, the flow of Fig. 5 is executed for each data item. For example, if the machine ID is input as an index of the machine classification index data, a list of the classified machine IDs is output. The form of the list is not limited. For example, the classification ID is assigned to each category, and each device ID and the corresponding category ID are stored in a table format output of one line. In addition, as an example of other lists, there is a method of preparing an address for each category and storing the machine ID belonging to the corresponding category in the address.

The feature amount conversion unit 102 converts the input machine classification index data into a feature amount in order to quantify the characteristics of each device. First, the feature amount conversion unit 102 determines that the input device classification index data is data of a numerical value or data other than the numerical value, and performs subsequent processing (step ST2). In the case of data other than data and values mixed in a table in combination with the sensor data of Fig. 2 and the maintenance performance data of Fig. 3, the data items and numerical values of the data divided into numerical values are excluded. The data item and perform the next processing on the splitter.

When the input device classification index data is numerical data such as sensor data (step ST2: YES), the feature amount conversion unit 102 executes the processing of step ST3. In step ST3, the numerical data is converted into a feature amount. The feature quantity may be, for example, a multivariate analysis method such as principal component of principal component analysis, distance under MT method, regression coefficient and error under regression analysis, similarity in pattern matching method, and the like. Methods. The feature quantity is shown in Fig. 6.

In Fig. 6, a feature amount table is exemplified as a feature quantity obtained for one machine of a plurality of devices. In the feature amount table, feature quantities 1 to 4 are described as an example of the feature amount of the device 1 of each device. The value of each item data of the feature amount table is an example. The feature amount can be changed in order to store the feature amount calculated from the actual data. In addition, if you can distinguish between equipment and machines, you can also collect data from a number of equipments and machines and make a table. Furthermore, if the association of the machines is possible, the data of one machine of one device can be divided into a plurality of tables.

Further, when the input device classification index data is not numerical data such as sensor data (step ST2: NO), the feature amount conversion unit 102 executes the processing of step ST4. That is, conventionally, the indicators used in the classification are only information such as the operation information of the elevator, the use of the elevator, the scale, etc., in the case of analysis of malfunctions or abnormalities, due to the sensor data or A lot of information such as past fault history is used as an object of analysis, and by combining these pieces of information into classification indicators in the classification of features, it is expected to improve the accuracy of analysis such as failures or abnormalities. Therefore, in the present embodiment, when the input device classification index data is data other than the numerical value such as the character data, the feature amount conversion unit 102 executes the processing of step ST4. As data other than the numerical value, for example, it is a character string data. Essay In the example of the character string, in step ST4, the character data is digitized and converted into a feature amount. The feature quantity is the same as the case of the numerical data, and can be, for example, multivariate analysis of each principal component of principal component analysis, distance under the MT method, regression coefficient and error under regression analysis, and similarity under the pattern comparison method. Method and other general methods. Alternatively, you can convert text data directly into feature quantities. The conversion method may be, for example, a method of allocating the same digital material to the same text data of the model, a method of digitizing the similarity of the morphemes obtained by dividing the character string, and an N-gram model for numerically similarity from the coincidence frequency. A general text column parsing method such as a method. The feature amount is stored in the same form as the feature scale of Fig. 6.

Then, in order to classify each machine by a device having similar features, the machine classifying unit 103 inputs a feature quantity converted by the feature quantity converting unit 102 of a plurality of machine parts, and sorts the machine whose feature quantity is an approximate value (step ST5). In the classification processing of step ST5, a general multivariate analysis method such as a hierarchical cluster analysis such as a dendrogram or a non-hierarchical cluster analysis such as a k-means method, or a support vector machine (support) can be used. General machine learning methods such as vector machine). An example of classification is shown in Figure 7.

Fig. 7 is a graphical representation of the feature quantity 1 and the feature quantity 2 in the second to fourth column data shown in the feature scale of Fig. 6 on the two-dimensional scattergram. The feature amount value 701 of Fig. 7 indicates the first column data of the feature amount table of Fig. 6, the feature amount value 702 indicates the second column data of the feature amount table, and the feature amount value 703 indicates the third column data of the feature amount table. The classifications 704 and 705 are used as an example of hierarchically classifying the distances between the respective data. The feature magnitude 701 and the feature magnitude 702 indicate that their combination becomes a category 704 due to the closer distance on the scatter plot. The feature quantity value 703 is in the scatter map due to the feature quantity value 701 and the feature quantity value 702. The upper distance is further, indicating that it is a different classification 705 than the classification 704.

As one of the uses of the present invention, it is a malfunction of the machine. Abnormal resolution. For example, in the case of predicting the future failure time to establish a maintenance plan for the machine, based on the data obtained from the machine, the probability of occurrence of future failure (fault risk) is predicted from the statistical failure frequency and deterioration tendency, and it is assumed that maintenance must be performed. The method of time.

The configuration of the machine classification device 100a to which the above-described function for predicting the risk of failure is added is shown in Fig. 8. The monitoring system shown in the figure is a configuration shown in Fig. 1 in the additional failure risk calculation unit 500. The failure risk calculation unit 500 is a processing unit that uses the result classified by the machine classification unit 103 to predict the risk of failure. As one method of calculating the risk of failure, the cumulative failure rate for the number of days elapsed since the maintenance work day is calculated using the failure occurrence frequency. By calculating the result by a function, the value obtained by inputting the number of days from the maintenance work day to the approximate function is used as a future failure risk. As a function of approximation, a general risk function utilized in reliability engineering such as Weibull distribution can be utilized. In addition, the risk of failure can be corrected in accordance with the degree of deterioration of the machine. Here, machines having similar characteristics are also likely to be similar due to the probability of failure or the tendency to deteriorate, and it is also useful to classify the machine by similar characteristics for predicting the risk of failure. Sorting by a more similar machine results in an increased accuracy of the risk of failure.

The data used by the failure risk calculation unit 500 may be the same as the material used by the data acquisition unit 101 to the machine classification unit 103, and may be used by including other materials. When estimating the risk of failure, although it is possible to predict the risk of failure according to the machine unit, it is also possible to comprehensively judge the failure risk of a plurality of machines, such as the correlation between a plurality of machines constituting the same equipment, and predict the failure of the equipment unit. risk. In other cases, when there is a possibility that a failure occurring in one machine occurs in another machine of the same classification, it can be used to develop or change a maintenance plan such as an early maintenance work. For other purposes, it can be used to allocate maintenance work for a plurality of machines of the same classification to maintenance work such as the same maintenance person. For other purposes, it can be used to indicate the same classification of machines in state monitoring such as remote monitoring of the machine.

Next, the effect of the first embodiment will be described. In the ninth figure, as an example of the feature quantity of each machine, for the two machines of the machine 1, the machine 2, data is collected from the three devices of the device a, the device b, and the device c, and the feature quantities are calculated and expressed in two characteristics. The amount is made on a 2-dimensional scatter plot. The device feature quantity of the machine 1 is indicated at 901, and the device feature quantity of the machine 2 is indicated as 902. In the conventional method, since it is a classification of equipment units, when equipment a, equipment b, and equipment c are classified as equipments having the same characteristics, the characteristics of the machines 1 and 2 are the same classification. On the other hand, in the embodiment 1, for example, in the device feature amount 901 of the machine 1, even in the case where the device a and the device b are one classification and the device c is another classification, the device feature amount 902 of the device 2 is the device. It is possible to classify a with equipment c as a classification and equipment b for other classifications and other machine units.

By classifying machines that hold the same characteristics, the prediction accuracy and failure of the machine's failure risk can be expected. The accuracy of the abnormality detection, etc. is improved. In addition, faults are found in a machine by sorting by similar machines. In the case of abnormality, etc., it is expected to prevent malfunctions in other machines by extracting the machine holding the same features and performing maintenance. Maintenance is more efficient due to abnormality and maintenance work on each machine scheduled. For example, because of the lift switch A's door switch horse When the torque is low and the shutdown occurs, the door switch motor of the other elevators having the same characteristics can be inspected and maintained without any indication of a low torque, and it is expected that the malfunction or accident can be reduced. As another example, when the torque of the door opening and closing motor of the elevator A is detected to be low, if the door opening and closing motor of the other elevator having the same characteristics has a possibility that the torque is lowered, it is not necessary to deal with it immediately, by appropriate Ground scheduling maintenance operations can be expected to be efficient. In addition, when the machine included in a specific device predicts the risk of future failure from past data, the past data obtained from the machine is small, and it is estimated that the risk of failure cannot be calculated. If the machine sorting device according to the embodiment 1 is used, data is obtained from a machine of another device having the same characteristics as the machine that predicts the risk of failure, and the acquired data is used, and it is possible to predict the failure of the machine with less past data. risk.

As described above, the machine classification device according to the first embodiment includes a machine classification index that acquires information unique to each device obtained from monitoring data of each of a plurality of devices composed of a single number or a plurality of devices. The data acquisition unit of the data, the feature quantity conversion unit that converts the machine classification index data into the feature quantity indicating the characteristics of the machine, and the machine that approximates the feature quantity, and the machine classification unit that classifies the device according to the machine unit. The machine can be analyzed for faults or abnormalities with high accuracy.

Further, according to the machine classification device of the first embodiment, the failure risk calculation unit that predicts the risk of failure of the machine that is determined to be similar by the machine classification unit is used, and the machine of each device can be surely executed. Maintenance work.

Implementation type 2

In the first embodiment, the machine classification index database 201 stores, for example, the sensor data shown in FIG. 2 or the maintenance performance data shown in FIG. 3, but a plurality of databases may be formed according to the data category. The case is described as Embodiment 2. Fig. 10 is a view showing the configuration of a monitoring system to which the machine classification device of the embodiment 2 is applied. The data collection management device 200a of the implementation type 2 includes a sensor database 202a and a maintenance performance database 202b as the machine classification index database group 202. The sensor database 202a is a database for accumulating sensor data of the monitoring target 400, and the data collection management device 200a is acquired from the monitoring target 400 via the network 300. The maintenance performance database 202b is a database for accumulating maintenance performance data of the monitoring object 400. The maintenance performance database 202b is composed of the maintenance performance data of the input device 600 for inputting the result of the maintenance of the monitoring target 400 via the network 300 by the data collection management device 200a. Further, the input device 600 is composed of a personal computer or the like, and is a device for inputting maintenance performance data to a maintenance person of the monitoring target 400.

The machine classification device 100b of the implementation type 2 is composed of a material acquisition unit 101a, a feature amount conversion unit 102, and a device classification unit 103. Here, the data acquisition unit 101a is configured to acquire sensor data from the sensor database 202a, and obtain maintenance performance data from the maintenance performance database 202b. Since the configuration and operation of the feature amount conversion unit 102 and the device classification unit 103 are the same as those of the first embodiment, the description thereof will be omitted.

Further, in the above example, the sensor data and the maintenance performance data are described as the data type, but the data type is not limited thereto, and the monitoring data obtained from the monitoring object 400 can be divided according to any category. In addition, it may be a decentralized management of a plurality of databases based on the viewpoint of load dispersion. Cheng, or, is not a database form but a file management machine as a component of the classification indicator data. Further, the machine classification index database group 202 shown in FIG. 10 is a conceptual group, and for example, each database may be distributed to different devices.

In this way, in the machine sorting apparatus of the implementation type 2, by dividing the machine classification index database as the machine classification index database group 202, the amount of data of each database can be reduced. Therefore, it is expected that the access speed from the machine sorting device 100b to the machine classification index data group 202 is increased. In addition, by arranging the respective databases to be distributed and arranged on different devices, the load on each device can be reduced.

As described above, according to the machine sorting device of the embodiment 2, the machine classification index data can be accessed to the machine classification index data more quickly by providing the data library according to the data type.

Implementation type 3

In the first embodiment, the device classification index data acquired by the data acquisition unit 101 is converted into the feature amount by the feature amount conversion unit 102, and the device classification unit 103 performs classification. However, the device classification may be based on the content of the device classification index data. The part 103 directly classifies the data from the machine classification index data. Such an example is explained as Embodiment 3.

Fig. 11 is a view showing the configuration of a monitoring system including the machine sorting device of the third embodiment. The machine classification device 100c shown in the figure includes a material acquisition unit 101 and a machine classification unit 103a. The data of the device classification index database 201a managed by the data collection management device 200b includes, for example, the usage frequency of the device in the monitoring target 400, the installation environment such as the region, the temperature, the operating temperature, and the like. Machine division The class part 103a is configured to directly extract the characteristics of each machine from such machine classification index data, and classify the devices by machine. The classification processing is the same as the implementation type 1, and a general multivariate analysis method such as a hierarchical cluster analysis such as a tree diagram or a non-hierarchical cluster analysis such as a k-means method, or a general support vector machine or the like can be used. Machine learning method.

As described above, the machine classification device according to the third embodiment includes a machine classification index that acquires information unique to each device obtained from monitoring data of each of a plurality of devices composed of a single number or a plurality of devices. The data acquisition unit of the data and the machine classification index data are used to classify the machine classification unit of the equipment according to the machine unit, and the failure or abnormality of the machine can be analyzed with high accuracy. In addition, since the processing of the feature amount is not calculated, the processing time can be shortened.

Further, according to the machine classification device of the third embodiment, the machine classification index data is such that the frequency of use of each device is information unique to each device, and the machine classification unit can directly use the machine classification index data for classification processing.

Further, according to the machine classification device of the third embodiment, the machine classification index data can be directly classified using the machine classification index data by performing the classification information by setting the machine environment information of each device to the information unique to each device.

Further, the present invention is not limited to any combination of the embodiments, or any of the constituent members of the respective embodiments, or the omission of any constituent members in the respective embodiments within the scope of the invention.

Industrial utilization possibility

As described above, the machine sorting apparatus according to the present invention classifies each apparatus by a machine held by these apparatuses for a plurality of apparatuses, and is suitable for use in a device in which an elevator or an air conditioner equivalent item exists in a plurality of environments.

100‧‧‧Machine sorting device

101‧‧‧Information Acquisition Department

102‧‧‧Characteristics Conversion Department

103‧‧‧Machine Classification Department

200‧‧‧Data collection management device

201‧‧‧ Machine Classification Indicators Database

300‧‧‧Network

400‧‧‧Monitoring objects

Claims (6)

A machine classification device comprising: a material acquisition unit that acquires machine classification index data, which is equivalent to each machine obtained from monitoring data of each of the plurality of devices each composed of a single or a plurality of devices; Intrinsic information; and the Machine Classification Department uses the above-mentioned machine classification indicator data to classify equipment according to machine units. A machine classification device comprising: a material acquisition unit that acquires machine classification index data, which is equivalent to each machine obtained from monitoring data of each of the plurality of devices each composed of a single or a plurality of devices; The intrinsic information; the feature quantity conversion unit converts the machine classification index data into a feature quantity indicating characteristics of the machine; and the machine classification unit, the machine having the feature quantity close to the machine having similar characteristics to classify the device according to the machine unit . The machine classification device according to claim 2, wherein the machine classification index data is provided as a database according to a data category. The machine sorting device according to claim 1, wherein the machine classification index data sets the frequency of use of each machine as information inherent to each machine. The machine classification device according to claim 1, wherein the machine classification index data sets the installation environment information of each machine as information inherent to the respective machines. The machine classification device as described in claim 2, further comprising: The failure risk calculation unit uses the failure occurrence frequency of each of the above-described devices to predict the risk of failure of the machine determined to be similar by the machine classification unit.