WO2019077679A1

WO2019077679A1 - Data processing device, data processing system, data processing method, data processing program, and storage medium

Info

Publication number: WO2019077679A1
Application number: PCT/JP2017/037555
Authority: WO
Inventors: 琢鴫原
Original assignee: 三菱電機株式会社
Priority date: 2017-10-17
Filing date: 2017-10-17
Publication date: 2019-04-25
Also published as: CN111213162B; KR102343793B1; JP6676210B2; KR20200051758A; CN111213162A; JPWO2019077679A1

Abstract

A data processing device (2) is provided with an algorithm selection unit (26) and a life prediction processing unit. The algorithm selection unit (26) selects, from an algorithm storage unit (30) that stores algorithms for predicting the lives of the components constituting a device, an algorithm associated with a target component, which is one of the components constituting the device, and the life of which is to be predicted. A prevention maintenance processing unit (25) constituting the life prediction processing unit performs a process for predicting the life of the target component on the basis of the algorithm selected by the algorithm selection unit (26).

Description

Data processing apparatus, data processing system, data processing method, data processing program and storage medium

The present invention relates to a data processing apparatus, data processing system, data processing method, data processing program, and storage medium which perform data processing for preventive maintenance of the apparatus.

Based on the data obtained by measuring the condition of the device, the indication of failure is monitored to predict the remaining life until the level of the indication exceeds a certain reference level, thereby precluding failure occurring in the device. Methods of preventing preventive maintenance are known.

Patent Document 1 discloses a control system which predicts the life of each part based on information from sensors provided in a plurality of parts constituting the engine, and predicts the life of the entire engine from the life of each part. ing. The control system of Patent Document 1 predicts the trend of phenomena causing failure such as breakage or loss of coating, and determines the remaining life of parts from the prediction result of the incidence of failure due to the occurrence of such phenomena. The control system of Patent Document 1 is programmed to calculate the remaining life of the parts according to a life prediction algorithm using data on the operating state of the engine.

Japanese Patent Application Publication No. 2014-518974

In the production site of products, in order to enable planned and stable production, preventive maintenance of production equipment operated at the production site is required. Maintenance or replacement of parts to the user of the production apparatus by calculating the lifetime of each part that constitutes the production apparatus using an algorithm that performs life prediction based on data obtained by measuring the state of the production apparatus. We can tell when we should do it.

In general, the parts that make up the production device are arbitrarily selected by the manufacturer of the production device. When the technology of Patent Document 1 is applied to a production apparatus, the provider of the application for preventive maintenance incorporated in the production apparatus constructs an algorithm specialized for the configuration of the production apparatus for each manufacturer of the production apparatus. It will be done. Also, when there is an addition or replacement of parts in a production device, the algorithm will be rebuilt. Therefore, according to the technology of Patent Document 1, there is a problem that the load required for constructing an application used for data processing for preventive maintenance of the device may increase.

The present invention has been made in view of the above, and it is an object of the present invention to obtain a data processing apparatus capable of reducing a load required for construction of an application used for data processing for preventive maintenance of the apparatus.

In order to solve the problems described above and to achieve the object, a data processing apparatus according to the present invention is a component that constitutes an apparatus from an algorithm storage unit in which an algorithm for performing lifetime prediction of parts constituting the apparatus is stored. An algorithm selection unit that selects an algorithm corresponding to a target component to be subjected to life prediction among them, and a life prediction processing unit that executes processing of life prediction of the target component based on the algorithm selected by the algorithm selection unit; , And.

The data processing apparatus according to the present invention has the effect of being able to reduce the load required to construct an application used for data processing for preventive maintenance of the apparatus.

Block diagram of data processing system according to embodiment 1 of the present invention Configuration diagram of the preventive maintenance application installed in the data processing apparatus shown in FIG. 1 Block diagram showing the functional configuration of the data processing apparatus shown in FIG. 1 Block diagram showing the hardware configuration of the data processing apparatus shown in FIG. 1 A flowchart showing the procedure of processing by the task handler shown in FIG. 2 Block diagram showing the functional configuration of the preventive maintenance processing unit shown in FIG. 3 The figure which shows the representative life curve and failure threshold value which are selected by the representative life curve selection part shown in FIG. A first diagram illustrating generation of a life prediction curve by the life prediction curve generation unit shown in FIG. A second diagram illustrating generation of a life prediction curve by the life prediction curve generation unit shown in FIG. A third diagram illustrating generation of a life prediction curve by the life prediction curve generation unit shown in FIG. The 4th figure explaining the production | generation of the life prediction curve by the life prediction curve generation part shown in FIG. A flowchart showing the procedure of processing by the data processing device after the preventive maintenance algorithm shown in FIG. 2 is selected The flowchart which shows the procedure of the processing which forms the life prediction curve with the life prediction curve formation department which is shown in Figure 6

Hereinafter, a data processing apparatus, a data processing system, a data processing method, a data processing program and a storage medium according to an embodiment of the present invention will be described in detail based on the drawings. The present invention is not limited by the embodiment.

Embodiment 1
FIG. 1 is a block diagram of a data processing system according to a first embodiment of the present invention. A data processing system 1 shown in FIG. 1 includes a data processing device 2, a device 4A connected to the data processing device 2, and devices 4B and 4C connected to the device 4A. The

devices

4A, 4B, 4C are devices for acquiring industrial data. Industrial data is data such as temperature, voltage, current, distance, speed or position information, and is any data about the condition of a production apparatus or a production site.

The device 4B is a production device, and is a driving device such as a numerical control (NC) device, a servomotor, an inverter or the like. The device 4A is a controller that controls the device 4B, and is a programmable logic controller (PLC). The device 4C is a sensor attached to the device 4B which is a production device, and is a vibration sensor, a sound collection microphone, a current clamp meter, a temperature sensor or the like. The number of

devices

4A, 4B and 4C provided in the data processing system 1 is arbitrary. The data processing system 1 shown in FIG. 1 includes one device 4A, two devices 4B, and one device 4C. The

devices

4A, 4B and 4C are not limited to the above specific examples, and may be devices that acquire industrial data.

The data processing device 2 is a computer in which a preventive maintenance application 10 which is a data processing program is installed. The data processing device 2 collects the industrial data transmitted from the

devices

4A, 4B, 4C, and performs a series of functional processing on the industrial data. The functional processing performed by the data processing device 2 includes processing for predicting the life of the parts constituting the device 4B. The data processing device 2 is connected to the cloud server 3 which is an external server. The display device 5 is connected to the data processing device 2. The display device 5 displays the result of the life prediction obtained by the data processing device 2.

The device 4B is provided with a mechanism for transmitting the driving force of the motor. One of the main causes of the failure of the device 4B is an abnormality of the rotating mechanism that rotates in response to the driving force of the motor. In the first embodiment, the data processing device 2 carries out the preventive maintenance of the device 4B by performing the life prediction of at least one of a bearing, a ball screw, a gear and a belt, which are main components constituting the rotation mechanism. . In FIG. 1, the illustration of the motor, the operation mechanism, the rotation mechanism, and the parts is omitted.

FIG. 2 is a block diagram of the preventive maintenance application 10 installed in the data processing apparatus 2 shown in FIG. The preventive maintenance application 10 which is a data processing program includes a task handler 11 and a preventive maintenance algorithm 12. The preventive maintenance algorithm 12 is a program in which an algorithm for preventive maintenance is implemented and which implements the implemented algorithm. The preventive maintenance algorithm 12 may be information in which a calculation procedure for preventive maintenance is described instead of the program. The information in which the calculation procedure is described may include information indicating a calculation formula. In this case, the algorithm for preventive maintenance is realized by the preventive maintenance application 10 referring to the stored preventive maintenance algorithm 12. In the following description, the preventive maintenance algorithm 12 is assumed to be a program.

The preventive maintenance algorithm 12 is prepared for each type of part. In the first embodiment, at least four preventive maintenance algorithms 12 for bearings, ball screws, gears, and belts are used. More finely divided part-type preventive maintenance algorithms 12 may be used, such as for large bearings, medium bearings, and small bearings. The data processing apparatus 2 can predict the lifetime in which the influence of the product specification such as the difference in the dimension of each part is folded by referring to the specification parameter 15 described later. For example, when performing life prediction for bearings having different dimensions, a common algorithm corresponding to both bearings may be used, and the specification parameters 15 may be made different.

The user of the device 4 B downloads the preventive maintenance application 10 in which the preventive maintenance algorithm 12 is incorporated, from a store selling the application on a website, etc., and installs the application on the data processing device 2. The user of the device 4B can change the preventive maintenance algorithm 12 incorporated in the preventive maintenance application 10. The user of the device 4B additionally acquires the preventive maintenance algorithm 12 by downloading from a store or the like that sells the application on the website. The user of the device 4B may obtain the preventive maintenance algorithm 12 by reading the preventive maintenance algorithm 12 from the storage medium in which the preventive maintenance algorithm 12 is stored. The user of the device 4B may acquire the preventive maintenance application 10 by reading the preventive maintenance application 10 from the storage medium in which the preventive maintenance application 10 is stored. The user of device 4 B may configure preventive maintenance application 10 by arbitrarily combining preventive maintenance algorithm 12.

The provider of the preventive maintenance application 10 provides the user of the device 4B with the preventive maintenance application 10 capable of adding and replacing the preventive maintenance algorithm 12. The task handler 11 is standard equipment in the preventive maintenance application 10 provided by the provider of the preventive maintenance application 10. The provider of the preventive maintenance algorithm 12 is assumed to be the manufacturer of the part, the manufacturer of the device 4B or the provider of the preventive maintenance application 10, but may be others.

The task handler 11 reads the setting information 14 and the specification parameter 15. The setting information 14 includes identification information for identification of the preventive maintenance algorithm 12 for each part, identification information for identification of the specification parameter 15 for each part, information for identifying a part for which life prediction is to be performed, and life It is a file including information on an execution cycle of life prediction for each part that executes prediction. The user of the device 4B can set which of the parts included in the preventive maintenance application 10 the corresponding preventive maintenance algorithm 12 executes the life prediction. Among the plurality of preventive maintenance algorithms 12 included in the preventive maintenance application 10, the preventive maintenance algorithm 12 to be used can be arbitrarily selected by the user.

The specification parameter 15 is a file that defines component-specific information. The specification parameter 15 is referred to at the time of specifying a part specific failure mode. The specification parameters 15 include information such as dimensions of parts. As a specific example, the specification parameters 15 for the bearing include the numerical values of the diameter of the rolling element, the pitch circle diameter of the rolling element, the number of rolling elements, and the contact angle. It is assumed that the user of the device 4B can obtain the specification parameters 15 created for each of the parts manufacturers and parts types via the web or a storage medium. The specification parameter 15 is created by the manufacturer of the part, but may be created by another person.

The identification information of the preventive maintenance algorithm 12 is a file name given to the file of the preventive maintenance algorithm 12. The identification information of the specification parameter 15 is a file name assigned to the file of the specification parameter 15. The identification information of the preventive maintenance algorithm 12 may be any information that can identify the preventive maintenance algorithm 12 for each part, and may be information other than the file name. The identification information of the specification parameter 15 may be information that can identify the specification parameter 15 of each part, and may be information other than the file name.

The task handler 11 manages the process of preventive maintenance in the preventive maintenance application 10. Management of the process of preventive maintenance by the task handler 11 includes management of an execution cycle of life prediction of each part. The task handler 11 recognizes the execution cycle of each component based on the information of the execution cycle included in the setting information 14. The data processing apparatus 2 can execute the life prediction with independent timing for each of the parts constituting the device 4B by managing the execution cycle in the task handler 11.

When there is a part whose execution cycle has arrived, the task handler 11 executes life prediction for the part. In addition, the information which shows not performing lifetime prediction may be settable to the information of an execution period. When the information indicating that the life prediction is not performed is set, the task handler 11 may not perform the life prediction for the part corresponding to the setting. The data processing apparatus 2 executes the preventive maintenance algorithm 12 corresponding to a part of the parts included in the preventive maintenance application 10 with the corresponding preventive maintenance algorithm 12 and the preventive maintenance algorithm corresponding to the other parts. It is good not to execute twelve. The task handler 11 activates a thread 13 for executing processing on a target part which is a target of life prediction. The thread 13 selects the preventive maintenance algorithm 12 corresponding to the target component based on the identification information included in the setting information 14 on the target component.

The thread 13 executes processing according to the selected preventive maintenance algorithm 12 using the specification parameter 15 of the target part. The task handler 11 performs parallel processing by a plurality of threads 13 when there are a plurality of parts for which the execution cycle has arrived. The thread 13 executes the process according to the preventive maintenance algorithm 12 using the specification parameter 15 to execute the process of life prediction. The functions of the preventive maintenance algorithm 12 include functions of calculation of failure frequency, calculation of actual measurement value, specification of failure mode, selection of representative life curve, selection of life prediction curve, and calculation of prediction life. Each function of the preventive maintenance algorithm 12, the failure frequency, and the failure mode will be described later.

FIG. 3 is a block diagram showing the functional configuration of the data processing apparatus 2 shown in FIG. Each functional unit shown in FIG. 3 is realized by the execution of the preventive maintenance application 10 on a computer which is hardware.

The data processing device 2 includes a control unit 20 that is a functional unit that controls the data processing device 2, a storage unit 21 that stores information, a communication unit 22 that is a functional unit that communicates information, and a function that inputs information. And an input unit 23 which is a unit.

The control unit 20 includes a preventive maintenance management unit 24 which is a functional unit that manages the process of preventive maintenance. The preventive maintenance manager 24, which is an execution cycle manager, manages an execution cycle of life prediction. In addition, the control unit 20 is a functional unit that executes the process of life prediction, and from the preventive maintenance algorithm 12 for each part, target parts to be subjected to life prediction among the parts constituting the apparatus And an algorithm selection unit 26 which is a functional unit that selects the preventive maintenance algorithm 12 corresponding to The preventive maintenance processing unit 25 is a life prediction processing unit that executes processing of life prediction of the target part based on the algorithm selected by the algorithm selection unit 26. The function of the preventive maintenance manager 24 and the function of the algorithm selector 26 are realized by the processing of the task handler 11. The function of the preventive maintenance processing unit 25 is realized by the process of the preventive maintenance algorithm 12 executed by using the specification parameter 15 of the target part.

The storage unit 21 includes an algorithm storage unit 30 for storing the preventive maintenance algorithm 12, an industrial data storage unit 31 for storing industrial data of all parts acquired from the

devices

4A, 4B and 4C, and specification parameters 15 for parts. And a specification parameter storage unit 32 for storing

The preventive maintenance algorithm 12 incorporated in the preventive maintenance application 10 is stored in the algorithm storage unit 30. The industrial data storage unit 31 stores industrial data acquired every one second along with time information. The industrial data storage unit 31 may store industrial data acquired at intervals of millisecond order or microsecond order, and may store industrial data acquired at other intervals. As a specific example, the industrial data acquired by the device 4B for the bearings includes the values of motor current, encoder position, motor speed, and temperature. The preventive maintenance processing unit 25 calculates the vibration frequency of the bearing based on the industrial data on the bearing. The calculation of the vibration frequency will be described later. Industrial data acquired by the device 4C, which is a sensor provided in the bearing, includes values of vibration acceleration and sound pressure level. The specification parameter storage unit 32 stores specification parameters 15 for each part of the device 4B.

Further, the storage unit 21 stores the life data storage unit 33 storing the result of the life prediction by the preventive maintenance processing unit 25, the representative life curve storage unit 34 storing the representative life curve and the failure threshold, and the setting information 14. And a setting information storage unit 35. Specifically, the life data storage unit 33 stores the failure mode of each part, the remaining life of the part, and the identification score. The setting information 14 is input to the data processing apparatus 2 by the manufacturer of the device 4B. The representative life curve, the failure threshold, the failure mode and the identification score will be described later.

The communication unit 22 performs communication between the data processing device 2 and the

devices

4A, 4B, 4C, which are devices outside the data processing device 2, the display device 5, and the cloud server 3. The input unit 23 inputs the setting information 14 into the data processing device 2.

FIG. 4 is a block diagram showing the hardware configuration of the data processing apparatus 2 shown in FIG. The data processing apparatus 2 includes a central processing unit (CPU) 40 that executes various processes, a random access memory (RAM) 41 including a program storage area and a data storage area, and an external storage device. A hard disk drive (Hard Disk Drive, HDD) 42 is provided. In addition, the data processing device 2 includes a communication circuit 43 which is a connection interface with an external device of the data processing device 2 and an input device 44 which receives an input operation to the data processing device 2. The respective units of the data processing device 2 shown in FIG. 4 are mutually connected via a bus 45. The external storage device may be a semiconductor memory.

The HDD 42 stores a preventive maintenance application 10, industrial data, component parameters 15 of parts, life data as a result of life prediction, a representative life curve, and setting information 14. The function of the storage unit 21 shown in FIG. 3 is realized using the HDD 42.

The preventive maintenance application 10 is loaded into the RAM 41. The CPU 40 develops the preventive maintenance application 10 in a program storage area in the RAM 41 and executes various processes. A data storage area in the RAM 41 is a work area in the execution of various processes. The functions of the control unit 20 shown in FIG. 3 are realized using the CPU 40. The function of the communication unit 22 is realized using the communication circuit 43. The input device 44 includes a keyboard or pointing device. The function of the input unit 23 shown in FIG. 3 is realized using the input device 44.

The preventive maintenance application 10 may be stored in a storage medium readable by a computer. The data processing device 2 may store the preventive maintenance application 10 stored in the storage medium in the HDD 42. The storage medium may be a portable storage medium, which is a flexible disk, or a flash memory, which is a semiconductor memory. The preventive maintenance application 10 may be installed on the data processing device 2 from another computer or server device via a communication network.

FIG. 5 is a flowchart showing the procedure of processing by the task handler 11 shown in FIG. The task handler 11 is activated in response to the activation of the computer which is the data processing device 2 and maintains the activated state until the computer is shut down.

In step S1, the task handler 11 determines whether or not there is a component for which the life prediction execution period has arrived, based on the information of the execution period included in the setting information 14 read from the setting information storage unit 35. Do. If there is no component for which the execution cycle has arrived (step S1, No), the task handler 11 waits in step S2 until it is next determined whether there is a component for which the execution cycle has arrived. After waiting in step S2, the task handler 11 returns the process to step S1.

If there is a component for which the execution cycle has arrived (step S1, Yes), the task handler 11 activates the thread 13 for the target component which is the component for which the execution cycle has arrived in step S3. In step S4, the thread 13 selects the preventive maintenance algorithm 12 for the target part based on the identification information included in the setting information 14 for the target part. Thereby, the task handler 11 selects the preventive maintenance algorithm 12 of the target part from the preventive maintenance algorithm 12 stored in the algorithm storage unit 30. The task handler 11 is not limited to one that executes the process of step S3 when there is a component for which the execution cycle of the life prediction has arrived in step S1. The task handler 11 may execute the process of step S3 when an instruction to execute the preventive maintenance process is input from the input unit 23 by the user.

The thread 13 selects, based on the identification information of the specification parameter 15, the specification parameter 15 for the part to be subjected to life prediction among the specification parameters 15 read from the specification parameter storage unit 32. In step S5, the thread 13 inputs the specification parameter 15 of the part into the preventive maintenance algorithm 12. In step S6, in the thread 13, the preventive maintenance algorithm 12 executes a life prediction process which is a preventive maintenance process. At the end of the process in step S6, the task handler 11 ends the process shown in FIG.

FIG. 6 is a block diagram showing a functional configuration of the preventive maintenance processing unit 25 shown in FIG. The preventive maintenance processing unit 25 selects a representative life curve, and a measured value calculation unit 51 that is a functional unit that calculates measured values, a failure mode calculation unit 52 that is a functional unit that calculates a failure mode value for each failure mode. A representative life curve selection unit 53 which is a functional unit, a failure mode identification unit 54 which is a functional unit which specifies a failure mode, and a life prediction curve generation unit 55 which is a functional unit which generates a life prediction curve. Further, the preventive maintenance processing unit 25 includes a life predicting unit 56 which is a functional unit that calculates a predicted life of the target component based on the preventive maintenance algorithm 12 selected by the algorithm selecting unit 26.

The failure mode indicates the cause of component failure. The failure mode value is a numerical value used to identify the failure mode. The preventive maintenance algorithm 12 stored in the algorithm storage unit 30 includes a failure model which is a calculation formula of failure mode values for each failure mode of the part. In the first embodiment, the failure mode is a failure cause that can be monitored for signs of failure by observing the frequency of vibration generated in the component. The failure mode calculation unit 52 calculates a failure frequency which is a failure mode value. The failure frequency is the frequency of vibration that is a sign of failure, and is a unique frequency for each failure mode. The failure mode calculation unit 52 calculates the failure frequency for each failure mode.

Here, calculation of the failure frequency by the failure mode calculation unit 52 will be described by taking a bearing which is one of the components as an example. Bearing failure can occur due to anomalies in the inner ring, outer ring, cage and rolling elements. The failure modes of the bearing include first to fifth failure modes described below.

In the following formulas (1) to (5), “d” is the diameter of the rolling element, “D” is the diameter of the pitch circle of the rolling element, “Z” is the number of rolling elements, and “α” is It is a contact angle. The unit of “d” and “D” is millimeter, and the unit of “α” is radian. The failure mode calculation unit 52 acquires each value of “d”, “D”, “Z”, and “α” from the specification parameter 15 read from the specification parameter storage unit 32. “F ₀ ” is the rotational frequency of the inner ring. The unit of “f ₀ ” is hertz. The failure mode calculation unit 52 calculates the value of “f ₀ ” based on the industrial data 16 read from the industrial data storage unit 31.

The first failure mode is a failure of the cage, and the symptom of failure can be monitored by observing the rotational frequency f _m of the cage. The failure mode calculation unit 52 calculates the rotational frequency f _m which is the failure frequency of the first failure mode by the following equation (1).

The second failure mode is the failure of the cage, and the symptom of failure can be monitored by observing the relative rotational frequency f _m-i of the cage relative to the inner ring. The failure mode calculation unit 52 calculates the relative rotational frequency f _m-i which is the failure frequency of the second failure mode by the following equation (2).

The third failure mode is a scratch or peeling of the race surface of the inner ring, and the indication of failure can be monitored by observing the passing frequency f _i of the rolling element with respect to the inner ring. The failure mode calculation unit 52 calculates the passing frequency f _i which is the failure frequency of the third failure mode by the following equation (3).

The fourth failure mode is a scratch or separation of the race surface of the outer ring, and the indication of failure can be monitored by observing the passing frequency f _O of the rolling element with respect to the outer ring. Failure mode calculator 52, the following equation (4), and calculates the pass frequency f _O is faulty frequency of the fourth failure mode.

The fifth failure mode is a scratch or peeling of the rolling element, and the indication of the failure can be monitored by observing the rotation frequency f _b of the rolling element. Failure mode calculator 52, by the following equation (5), and calculates the rotation frequency f _b is the failure frequency of the fifth failure mode.

When the target component is a bearing of the servo motor, failure mode calculating unit 52, based on the speed monitoring value is a industry data obtained from the servo motor, it may be calculated rotation frequency f ₀ of the inner ring. Or, failure mode calculating unit 52, based on the number of pulses is industrial data acquired from the device 4C is a sensor provided outside the pulse encoder may calculate the rotation frequency f ₀ of the inner ring. The rotational frequency f ₀ is a value that fluctuates in a specific cycle. The failure mode calculation unit 52 acquires the rotational frequency f ₀ at a constant timing in the cycle.

By rotation frequency f ₀ at a constant timing is acquired, the value of the rotation frequency f ₀ obtained in failure mode calculating unit 52 at each timing is constant. Therefore, the value of the rotation frequency f _0, in place of the value calculated on the basis of the industrial data may be preset values in specifications parameter 15. It should be noted that even if the rotational frequency f ₀ is acquired at a fixed timing, the rotational frequency f ₀ may slightly fluctuate depending on the situation of the device 4 B, so the rotational frequency f ₀ is calculated based on the industrial data. By calculating, it is possible to obtain a value in which the change in the status of the device 4B is more reflected than the value set in advance. Therefore, by obtaining the rotation frequency f ₀ by calculation based on industry data, failure mode calculating unit 52 can calculate accurately fault frequency.

The failure mode may be a failure cause that can monitor a symptom of failure by observing a phenomenon other than vibration. The failure mode calculation unit 52 may calculate failure mode values other than the failure frequency. If the failure mode is a failure cause that can be monitored for symptoms of failure by observing the temperature of the gearbox, then the failure mode value is the failure temperature which is the temperature of the gearbox. The failure mode calculation unit 52 acquires a failure temperature.

The actual measurement value calculation unit 51 reads out the industrial data 16 stored in the industrial data storage unit 31 and calculates an actual measurement value corresponding to the failure mode value based on the read industrial data 16. When the failure mode value is the failure frequency, the actual measurement value calculation unit 51 calculates an actual measurement frequency that is an actual measurement value. The actual measurement frequency is a frequency of vibration generated in the part, and is calculated based on the industrial data 16 acquired by the devices 4B and 4C. In the case where the target part is a bearing, the actual measurement value calculation unit 51 calculates an actual measurement frequency based on the motor current value acquired by the device 4B. The actual measurement value calculation unit 51 calculates an actual measurement frequency by extracting frequency components by fast Fourier transform (FFT) of the motor current value. The industrial data storage unit 31 stores data obtained by the FFT as the industrial data 16.

When the vibration phenomenon is not observed in the device 4B, the actual measurement value calculation unit 51 may calculate the actual measurement frequency based on the industrial data 16 acquired by the device 4C. For calculation of the actual measurement frequency, vibration acceleration acquired by a vibration sensor which is a device 4C attached to a bearing may be used. The sound pressure level acquired by the sound pressure sensor which is the device 4C attached to a bearing may be used for calculation of measurement frequency. The actual measurement value calculation unit 51 may calculate the actual measurement frequency by extracting frequency components by FFT of vibration acceleration or sound pressure level.

The failure mode identification unit 54 compares the failure mode value calculated by the failure mode calculation unit 52 with the actual measurement value calculated by the actual value calculation unit 51 to specify the failure mode of the target part. When the target component is a bearing, the failure mode identifying unit 54 determines a failure frequency that matches the measured frequency among the failure frequencies of the first to fifth failure modes. If the measured frequency matches the above-mentioned rotational frequency f _m , the failure mode identifying unit 54 identifies the failure mode of the bearing as the target component as the first failure mode.

The failure mode identification unit 54 can apply various methods to the method of determining whether the failure mode value and the actual value match or not. The failure mode identification unit 54 may determine whether or not the failure mode value and the actual measurement value match, based on a predetermined error range. When the difference between the failure mode value and the actual measurement value is within the error range, the failure mode identification unit 54 determines that the failure mode value and the actual measurement value match. The failure mode identification unit 54 sends information indicating the identified failure mode to the representative life curve selection unit 53 and the life prediction curve generation unit 55.

The failure mode identification unit 54 may calculate an identification score indicating the probability that the phenomenon observed as the actual measurement value corresponding to the failure mode value is a phenomenon due to the failure of the identified failure mode. The failure mode identification unit 54 calculates an identification score based on the difference between the failure mode value and the actual measurement value. The identification score is sent to the life prediction unit 56 through the life prediction curve generation unit 55. The user of the device 4B can determine the reliability of the failure determination of the target part by referring to the identification score.

The representative life curve selection unit 53 selects a representative life curve and a failure threshold based on the failure mode specified by the failure mode specification unit 54. The representative life curve, which is the first curve, is a curve that approximates the data obtained by the accelerated life test of the part, and corresponds to the measured value and time corresponding to the failure mode value for the phenomenon that occurred in the test. Represents a relationship. The accelerated life test is a test for intentionally advancing the deterioration of a part to be tested to verify the life of the part. The failure threshold value is an actual measurement value when a component fails in a test.

FIG. 7 is a view showing a representative life curve C1 and a failure threshold T selected by the representative life curve selection unit 53 shown in FIG. When the failure mode value is the failure frequency, the representative life curve C1 represents the relationship between the amplitude of the vibration generated in the test and the time. The failure threshold T is the vibration amplitude when the part fails in the test. That is, when the failure mode value is the failure frequency, vibration amplitudes corresponding to the failure frequency are plotted in time series. In FIG. 7, the vertical axis represents vibration amplitude, and the horizontal axis represents time. In the following description, the vertical axis representing vibration amplitude may be referred to as Y axis, and the horizontal axis representing time may be referred to as X axis. Even if the parts are manufactured by the same manufacturer and are parts of the same type, data obtained by the accelerated life test may vary. The representative life curve C1 is a life curve representing a life curve manufactured by the same manufacturer and obtained by parts of the same type.

The representative life curve storage unit 34 which is a curve storage unit stores, for each part of the device 4B, a representative life curve and a failure threshold for each failure mode. The representative life curve selection unit 53 uses the representative life curve C1 and the failure threshold T corresponding to the target part and the specified failure mode from the representative life curve and the failure threshold stored in the representative life curve storage unit 34. select. The time L1 is a time when the vibration amplitude reaches the failure threshold T in the representative life curve C1.

The representative life curve selection unit 53 sends the selection result of the representative life curve C1 and the failure threshold T to the life prediction curve generation unit 55. In addition to the vibration amplitude, the vertical axis when the representative life curve C1 is expressed may represent temperature, frictional force or the like which is a parameter according to the failure mode value. Further, the horizontal axis may represent, besides time, an integrated temperature or the like which is a parameter indicating the progress of deterioration of the component.

The life prediction curve generation unit 55 generates a life prediction curve based on the representative life curve C1 selected by the representative life curve selection unit 53. The life prediction curve represents the prediction of the time series change of the measured value after the execution of the life prediction. When the failure mode value is the failure frequency, the life prediction curve represents the relationship between vibration amplitude and time after the execution of the life prediction.

FIG. 8 is a first diagram for explaining generation of a life prediction curve by the life prediction curve generation unit 55 shown in FIG. FIG. 9 is a second diagram for explaining generation of a life prediction curve by the life prediction curve generation unit 55 shown in FIG. The life prediction curve generation unit 55 reads the representative life curve C1 and the failure threshold T from the representative life curve storage unit 34 according to the selection result by the representative life curve selection unit 53.

The length of the time axis differs between the representative life curve C1 obtained by the life acceleration test and the plot of the measured value. Life prediction curve generation unit 55 represents the time axis up to time L1 of representative life curve C1 to the time axis up to time L2, which is the rated life according to the actual usage condition of the target part, in the horizontal axis direction The life curve C1 is elongated. The rated life is the life at which a standard product is used.

To give a specific example, the rated life when the ball bearing which is the target part is a ball bearing is expressed as (C / P) ³ × 16667 / n. The rated life when the ball bearing which is the target part is a roller bearing is expressed as (C / P) ^10/3 × 16667 / n. Here, “C” is a basic dynamic load rating, “P” is a dynamic equivalent load, and “n” is a rotational speed. The unit of “C” and “P” is Newton, and the unit of “n” is revolution per minute (rpm). The unit of rated life is hours.

A dynamic equivalent load on _{"P", P = X r × Fr +} Y a × Fa is established. Here, “X _r ” is a radial coefficient, “Fr” is a radial load, “Y _a ” is an axial coefficient, and “Fa” is an axial load. The unit of "Fr" and "Fa" is Newton. The life prediction curve generation unit 55 acquires each value of “C”, “n”, “X _r ” and “Y _a ” from the specification parameter 15 read from the specification parameter storage unit 32. The life prediction curve generation unit 55 acquires each value of “Fr” and “Fa” from the setting profile. The setting profile is a file defining information specific to the device 4B and the use environment of the device 4B. Further, the life prediction curve generation unit 55 determines which one of the ball bearing and the bearing the ball bearing is, based on the set profile. The life prediction curve generation unit 55 may calculate the time L2 which is the rated life based on the specification parameter 15 and the setting profile.

The life prediction curve generation unit 55 generates a rated curve C2 which is a second curve by deforming the representative life curve C1 based on the failure threshold T and the time L2 which is the rated life. The constants “a”, “b” and “c” of the exponential function Y = a × b ^x + c represented by the rated curve C2 are representative lifetimes in the X-axis direction until the vibration amplitude at time L2 matches the failure threshold T It is obtained by stretching the curve C1. The life prediction curve generation unit 55 stretches the representative life curve C1 by scaling the X axis among the X axis and the Y axis.

FIG. 10 is a third diagram illustrating generation of a life prediction curve by the life prediction curve generation unit 55 shown in FIG. The life prediction curve generation unit 55 reads out from the industrial data storage unit 31 the actual measurement values of the vibration amplitude up to the present time among the industrial data 16, and plots the read actual measurement values on the time axis so far. When the target part is a bearing, the vibration amplitude can be extracted from data obtained by FFT of the motor current value acquired by the device 4B by the actual measurement value calculation unit 51. The life prediction curve generation unit 55 generates an actual measurement curve C3 representing an exponential function Y = a '× b ^x + c' by approximation of the actual measurement values. The life prediction curve generation unit 55 generates an actual measurement curve C3 which is a third curve representing a relationship between an actual measurement value of vibration amplitude which is an actual measurement value corresponding to the failure mode value and time. The constant "b" of the measured curve C3 is made to coincide with the constant "b" of the rated curve C2. The time L3 is a time when the vibration amplitude reaches the failure threshold T in the actually measured curve C3.

When the actual measured value from the present to the latest is smaller than the actual measured value before the latest, the constant “a ′” may have a negative value. In this case, the life prediction curve generation unit 55 may use the constant “a ′” calculated in the previous life prediction to generate the measurement curve C3. Alternatively, when the previous constant "a '" does not exist, the life prediction curve generation unit 55 may use the constant "a" of the representative life curve C1 to generate the actual measurement curve C3.

FIG. 11 is a fourth diagram for explaining the generation of the life prediction curve C4 by the life prediction curve generation unit 55 shown in FIG. The life prediction curve generation unit 55 generates a life prediction curve C4 by mixing the rated curve C2 and the actual measurement curve C3. Thereby, the preventive maintenance processing unit 25 obtains the life prediction curve C4 generated based on the rated curve C2 and the actual measurement curve C3. The life prediction curve generation unit 55 generates the life prediction curve C4 by weighting the measurement curve C3 that represents the degree of control of the measurement curve C3 in the life prediction curve C4. Thus, the life prediction curve generation unit 55 changes the ratio of the actual measurement curve C3 included in the life prediction curve C4.

The life prediction curve generation unit 55 changes the weighting ratio p used to generate the life prediction curve C4 with the lower limit being 0% and the upper limit being 100%. When the weighting ratio p is 0%, the life prediction curve C4 matches the rated curve C2. When the weighting ratio p is 100%, the life prediction curve C4 matches the actual measurement curve C3. The time L4 is a time when the vibration amplitude reaches the failure threshold T in the life prediction curve C4.

Here, a setting example of the weighting ratio p will be described. The weighting ratio p is determined based on the condition of vibration amplitude which is the vertical axis shown in FIG. 11 and the condition of time which is the horizontal axis. The condition of vibration amplitude is Y-axis condition, and the condition of time is X-axis condition.

In the first life prediction after starting the operation of the device 4B, the weighting ratio p is 0% based on the X-axis condition that it is the first time. In addition, when the time from the start of the operation of the device 4B to the present time exceeds the time L2 which is the rated life of the target part, the weighting ratio p is 100% based on the X-axis condition that the time L2 is exceeded. I assume.

The weighting ratio p is determined in the previous life prediction based on the Y-axis condition that the vibration amplitude is constant if the current measurement value of the vibration amplitude is constant with the actual measurement in the previous life prediction. And the same as the weighting ratio p. Here, that two measured values are constant means that the difference between the two measured values is within a preset percentage range.

Based on the Y-axis condition that the vibration amplitude has increased if the current measured value of the vibration amplitude is not constant with the measured value at the previous life prediction and the current measured value is increased from the previous measured value. The weighting ratio p is increased from the previous time. In addition to these conditions, when the X-axis condition that the time from the start of the operation of the device 4B to the present is less than 70% of the time L2 holds, the weighting ratio p is increased by 10% from the previous time. When the X-axis condition that the time from the start of the operation of the device 4B to the present time is 70% or more and less than 80% of the time L2 is satisfied, the weighting ratio p is increased by 20% from the previous time. When the X-axis condition that the time from the start of the operation of the device 4B to the present time is 80% or more and less than 90% of the time L2 holds, the weighting ratio p is increased by 30% from the previous time. When the X-axis condition that the time from the start of the operation of the device 4B to the present time is 90% or more and less than 100% of the time L2 is satisfied, the weighting ratio p is increased by 40% from the previous time.

Based on the Y-axis condition that the vibration amplitude decreases if the current measured value of the vibration amplitude is not constant with the measured value at the previous life prediction, and the current measured value is reduced from the previous measured value. The weighting ratio p is set to be smaller than the previous one or equal to the previous one. In addition to this condition, when the X-axis condition that the time from the start of the operation of the device 4B to the present time is less than 70% of the time L2 holds, the weighting ratio p is reduced by 10% from the previous time. When the X-axis condition that the time from the start of the operation of the device 4B to the present time is 70% or more and less than 100% of the time L2 holds, the weighting ratio p is the same as the previous one.

Thus, by setting the weighting ratio p based on the X-axis condition, the life prediction curve generation unit 55 is at an early stage after starting the operation of the device 4B, and the time when accumulation of measured values is small Then, the life prediction curve C4 weighted so that the rated curve C2 is dominant compared to the actual measurement curve C3 is generated. Thereby, the preventive maintenance processing unit 25 can perform life prediction with a weight given to the rated life at a time when accumulation of measured values is small. Further, the life prediction curve generation unit 55 performs weighting such that the control degree of the measurement curve C3 becomes high as time passes. The life prediction curve generation unit 55 changes the life prediction curve C4 so as to be close to the measurement curve C3 as accumulation of the measured value increases with the passage of time. Thus, the preventive maintenance processing unit 25 can perform life prediction with a weight given to the accumulated actual measurement value as the accumulation of the actual measurement value increases. Further, by setting the weighting ratio p based on the Y-axis condition, the life prediction curve generation unit 55 performs weighting such that the control degree of the actual measurement curve C3 becomes high as the actual measurement value of the vibration amplitude increases. The life prediction curve generation unit 55 changes the life prediction curve C4 so as to be closer to the actual measurement curve C3 as the vibration amplitude increases. Thereby, the preventive maintenance processing unit 25 can perform the life prediction according to the situation where the measured value is increasing.

The life prediction unit 56 obtains the time L4 by substituting the failure threshold T into the exponential function represented by the life prediction curve C4 generated by the life prediction curve generation unit 55. The life prediction unit 56 calculates the remaining life which is the time from the present to the time L4. The life prediction unit 56 determines the failure mode specified by the failure mode specification unit 54, the remaining life which is the result 17 of the life prediction calculated by the life prediction unit 56, and the identification score calculated by the failure mode specification unit 54. Is sent to the life data storage unit 33. The life data storage unit 33 stores the failure mode, the remaining life, and the identification score. The display device 5 displays the failure mode read from the life data storage unit 33, the remaining life, and the identification score.

FIG. 12 is a flowchart showing the procedure of processing by the data processing device 2 after the preventive maintenance algorithm 12 shown in FIG. 2 is selected. In step S11, the failure mode calculation unit 52 calculates the failure frequency of each failure mode of the target part. In step S12, the measured value calculation unit 51 calculates the measured value of the frequency of the vibration generated in the component.

In step S13, the failure mode identifying unit 54 identifies the failure mode of the target component by comparing the actually measured frequency, which is the measured value, with the failure frequency. In step S14, the failure mode identifying unit 54 calculates an identification score of the identified failure mode. The representative life curve selection unit 53 selects a representative life curve and a failure threshold based on the specified failure mode.

In step S15, the life prediction curve generation unit 55 reads the representative life curve C1 and the failure threshold T selected by the representative life curve selection unit 53 from the representative life curve storage unit 34. In step S16, the life prediction curve generation unit 55 generates a life prediction curve C4 based on the read representative life curve C1.

FIG. 13 is a flowchart showing a procedure of processing of generating the life prediction curve C4 by the life prediction curve generation unit 55 shown in FIG. In step S21, the life prediction curve generation unit 55 obtains the rated curve C2 by extending the time axis of the representative life curve C1 based on the rated life and the failure threshold T.

In step S22, the life prediction curve generation unit 55 obtains the actual measurement curve C3 based on the actual measurement values of the vibration amplitude up to the present time. In step S23, the life prediction curve generation unit 55 obtains the life prediction curve C4 according to the weighting by applying weighting on the rated curve C2 to be close to the actual measurement curve C3. Thus, the process of generating the life prediction curve C4 by the life prediction curve generation unit 55 ends.

In step S17 shown in FIG. 12, the life prediction unit 56 calculates the remaining life of the target component based on the life prediction curve C4 generated by the life prediction curve generation unit 55. In step S18, life data storage unit 33 stores the failure mode specified in step S13, the remaining life calculated in step S17, and the identification score calculated in step S14. In step S19, the display device 5 displays the failure mode read from the life data storage unit 33, the remaining life, and the identification score. Thus, the data processing device 2 ends the process shown in FIG.

A part or all of the processing by the function of the data processing device 2 of the first embodiment may be performed in the cloud server 3. The cloud server 3 may hold a failure model which is a calculation formula of failure mode value, and perform calculation of failure mode value and identification of failure mode.

According to the first embodiment, the data processing device 2 includes the algorithm selection unit 26 which selects the preventive maintenance algorithm 12 corresponding to the target part. Compared with the case where an algorithm built specifically for the configuration of the production apparatus is installed in the preventive maintenance application 10, the load required for constructing the preventive maintenance application 10 can be reduced. As a result, it is possible to reduce the load required to build an application used for data processing for preventive maintenance of the device.

Second Embodiment
The data processing apparatus 2 according to the second embodiment of the present invention changes the execution cycle of the life prediction for each part according to the elapsed time from the start of use of the part for each part constituting the device 4B. . The data processing device 2 according to the second embodiment has the same configuration as the data processing device 2 according to the first embodiment. The preventive maintenance manager 24, which is an execution cycle manager, changes the execution cycle according to the elapsed time since the start of use of the part.

The increase in the measured value of the vibration amplitude becomes earlier as the life of the part approaches. In the second embodiment, the preventive maintenance managing unit 24 shortens the execution period of the life prediction and increases the execution frequency of the life prediction process as the elapsed time from the start of the use of the parts becomes longer. The preventive maintenance manager 24 may change the execution cycle of the life prediction based on the weighting ratio p in the first embodiment. As a result, the preventive maintenance management unit 24 increases the execution frequency of the life prediction process as the measured value of the vibration amplitude increases and as time passes.

According to the second embodiment, the data processing device 2 changes the execution cycle of the life prediction for each part according to the elapsed time since the start of use of the part, according to the increase degree of the actual measurement value. The execution frequency of the life prediction process can be changed. Thus, the data processing device 2 can increase the prediction accuracy of the remaining life.

The configuration shown in the above embodiment shows an example of the contents of the present invention, and can be combined with another known technique, and one of the configurations is possible within the scope of the present invention. Parts can be omitted or changed.

DESCRIPTION OF SYMBOLS 1 data processing system, 2 data processing apparatus, 3 cloud server, 4A, 4B, 4C device, 5 display apparatus, 10 preventive maintenance application, 11 task handler, 12 preventive maintenance algorithm, 13 threads, 14 setting information, 15 specification parameter , 16 industrial data, 20 control unit, 21 storage unit, 22 communication unit, 23 input unit, 24 preventive maintenance management unit, 25 preventive maintenance processing unit, 26 algorithm selection unit, 30 algorithm storage unit, 31 industry data storage unit, 32 Specifications parameter storage unit, 33 life data storage unit, 34 representative life curve storage unit, 35 setting information storage unit, 40 CPU, 41 RAM, 42 HDD, 43 communication circuit, 44 input device, 45 bus, 51 actual value calculation unit , 52 failure modes Out portion, 53 representative life curve selecting unit 54 failure mode identification unit 55 life prediction curve generation unit, 56 life prediction unit.

Claims

An algorithm selection unit for selecting an algorithm corresponding to a target part to be a target of life prediction among the parts constituting the apparatus from an algorithm storage part in which an algorithm for performing life prediction of parts constituting the apparatus is stored; ,
A life prediction processing unit that executes processing of life prediction of the target component based on the algorithm selected by the algorithm selection unit;
A data processing apparatus comprising:
The data processing apparatus according to claim 1, wherein the algorithm selection unit selects an algorithm based on identification information for identifying an algorithm for each part.
The said algorithm selection part selects the algorithm corresponding to the said target component for which the said execution period came based on the information of the execution period of the said lifetime prediction for every said component, The said Claim 1 or 2 characterized by the above-mentioned. Data processing equipment.
An execution cycle management unit that manages the execution cycle of each part;
4. The data processing apparatus according to claim 3, wherein the execution cycle management unit changes the execution cycle according to an elapsed time after the start of use of the component.
The algorithm selection unit inputs specification parameters, which are information specific to the target part, to the life prediction processing unit.
The said life prediction process part performs the process of the said life prediction using the algorithm and the said specification parameter which were selected by the said algorithm selection part in any one of the Claims 1 to 4 characterized by the above-mentioned. Data processor as described.
The life prediction processing unit
A failure mode calculation unit that calculates a failure mode value that is a numerical value used to identify a failure mode that indicates a failure cause of the part;
A failure mode identification unit that identifies a failure mode of the target component based on the failure mode value;
The data processing apparatus according to any one of claims 1 to 5, comprising:
A first curve representing a relationship between an actual measurement value obtained in a test for verifying the life of the part and time, and a failure threshold which is the actual measurement value when the part fails in the test; It has a storage unit to store
The life prediction processing unit reads the first curve and the failure threshold for the target component from the storage unit, and deforms the first curve based on the failure threshold and the rated life of the target component. Generating a second curve by causing the second curve to generate a life prediction curve to be used for calculating the predicted life of the target part based on the second curve. The data processing apparatus according to claim 6.
The life prediction curve generation unit generates a third curve representing a relationship between an actual measurement value corresponding to the failure mode value and time, and the life prediction by mixing the second curve and the third curve. The data processing apparatus according to claim 7, wherein a curve is generated.
9. The life prediction curve generation unit according to claim 8, wherein the life prediction curve is generated by applying a weighting to the third curve to represent the degree of control of the third curve in the life prediction curve. Data processor as described.
The data processing apparatus according to claim 9, wherein the life prediction curve generation unit performs the weighting such that the control degree of the third curve becomes higher as time passes.
The data processing apparatus according to claim 9, wherein the life prediction curve generation unit performs the weighting such that the degree of control of the third curve increases as the actual measurement value increases.
The failure mode identification unit is characterized by calculating an identification score indicating the probability that the phenomenon observed as the actual measurement value corresponding to the failure mode value is a phenomenon due to a failure of the identified failure mode. The data processing apparatus according to claim 6.
An algorithm selection unit for selecting an algorithm corresponding to a target part to be a target of life prediction among the parts constituting the apparatus from an algorithm storage part in which an algorithm for performing life prediction of parts constituting the apparatus is stored; ,
A life prediction processing unit that executes processing of life prediction of the target component based on the algorithm selected by the algorithm selection unit;
A data processing system comprising:
The data processor
Selecting an algorithm corresponding to a target part to be a target of the life prediction among the parts constituting the device, from an algorithm for performing the life prediction of the parts constituting the device;
Executing a process of life prediction of the target part based on the selected algorithm;
A data processing method comprising:
A data processing program that causes a computer to function as a data processing apparatus that performs processing of life prediction of parts constituting the apparatus,
Selecting an algorithm corresponding to a target part to be a target of the life prediction among the parts constituting the apparatus from the algorithm for performing the life prediction of the part;
Executing a process of life prediction of the target part based on the selected algorithm;
A data processing program causing the computer to execute the program.
A storage medium storing the data processing program according to claim 15 and readable by a computer.