CN117441291A

CN117441291A - Life prediction device, life prediction system, learning device, estimation device, and life prediction program

Info

Publication number: CN117441291A
Application number: CN202180098747.3A
Authority: CN
Inventors: 辻川孝辅
Original assignee: Mitsubishi Electric Corp
Current assignee: Mitsubishi Electric Corp
Priority date: 2021-12-01
Filing date: 2021-12-01
Publication date: 2024-01-23
Also published as: JPWO2023100300A1; WO2023100300A1; DE112021007632T5; JP7090832B1

Abstract

The lifetime prediction device (4) comprises: a lifetime storage unit (44) for storing, as lifetime data, a basic lifetime of an object, which is a device or a component constituting an industrial machine, and a lifetime which is a lifetime consumed from the basic lifetime; a device stress estimating unit (41) that estimates a device stress on the basis of device state information (8), which is information on the use state of the object; a lifetime consumption estimating unit (42) that estimates, based on the device stress, the lifetime consumption, which is the lifetime of the object consumed in each specified cycle; and a lifetime predicting unit (43) that predicts the predicted device lifetime of the object based on the lifetime data and lifetime consumption held by the lifetime storing unit (44), and outputs the device stress estimated by the device stress estimating unit (41) and the predicted device lifetime predicted by the lifetime predicting unit (43) so as to be notified to the outside.

Description

Life prediction device, life prediction system, learning device, estimation device, and life prediction program

Technical Field

The present invention relates to a life prediction device, a life prediction system, a learning device, an estimation device, and a life prediction program for predicting the life of a structural device or a component of an industrial machine, such as a work machine.

Background

In the field of industrial machines typified by work machines, maintenance such as replacement of components is performed before the life of components is reached in order to predict the life of components by preventing the occurrence of an accident due to failure of the components or devices constituting the machine. Therefore, it is necessary to accurately grasp the product life of each structural member in maintenance.

However, it is known that the lifetime of the constituent devices varies considerably depending on the operating environment.

For example, a motor control device, which is one of the above-described mechanical structure devices, is composed of a plurality of life components (capacitors, semiconductor elements, etc.). Since the operating conditions of these components greatly affect the life, the operating environment conditions (operating temperature, altitude, humidity) are defined, and the rated values are defined in accordance with the operating environment conditions, thereby ensuring a predetermined product life.

As a representative life member, there is an aluminum electrolytic capacitor mainly mounted in a converter unit of a motor control device. It is known that if the thermal stress increases with an increase in the temperature of the use environment, the lifetime of the aluminum electrolytic capacitor deteriorates according to the arrhenius law.

As described above, the life of the device and the product may be affected by the operating environment of the machine, and thus the life prediction value and the actual life may deviate. When the actual operating environment is mild, the stress applied to the device (hereinafter referred to as device stress) as a factor affecting the product lifetime is low, and thus the device lifetime is long. Conversely, in the case where the operating environment is severe, the device stress becomes excessive, and the device lifetime may be significantly deteriorated. In particular, in the latter case, maintenance costs increase such as machine stoppage at the time of failure and occurrence of a failure again without improvement in the operation environment.

For the above-described problems, a technique for estimating the lifetime of a device using operation data of the device is described in patent literature 1, for example.

Patent document 1 discloses a technique of estimating an ambient temperature from measurement data of an internal temperature of a power supply device, calculating a load factor from a current value flowing when operating in a state where a load is connected, and predicting a lifetime of the power supply device from the ambient temperature and the load factor of the power supply device.

Patent document 1: japanese patent laid-open No. 2020-058167

Disclosure of Invention

According to the technique described in patent document 1, the lifetime can be predicted using the internal temperature of the device and the value of the current flowing through the device, but the user cannot be made aware of the factors that affect the lifetime. For example, even when the lifetime of the device can be extended by changing the use condition of the device, the user cannot know which condition to change.

The present invention has been made in view of the above circumstances, and an object of the present invention is to obtain a life predicting device capable of providing a life predicting result of a device or a component constituting an industrial machine and a factor affecting the life to a user.

In order to solve the above problems and achieve the object, a lifetime prediction device according to the present invention includes: a lifetime storage unit that holds, as lifetime data, a lifetime which is a lifetime consumed from a basic lifetime of an object which is a device or a component constituting an industrial machine; a device stress estimating unit that estimates a device stress based on device state information, which is information on a use state of an object; a lifetime consumption estimating unit that estimates lifetime consumption, which is the lifetime of the object consumed in each specified cycle, based on the device stress; and a lifetime predicting unit that predicts a lifetime of the predicting device for the object based on the lifetime data and the lifetime consumption held by the lifetime storing unit. The lifetime predicting device outputs the device stress estimated by the device stress estimating unit and the predicted lifetime of the device predicted by the lifetime predicting unit so as to notify the outside.

ADVANTAGEOUS EFFECTS OF INVENTION

The lifetime prediction device according to the present invention has an effect that a lifetime prediction result of a device or a component constituting an industrial machine and a factor affecting lifetime can be provided to a user.

Drawings

Fig. 1 is a diagram showing a configuration example of a lifetime prediction system realized by applying the lifetime prediction device according to embodiment 1.

Fig. 2 is a diagram showing an example of a relationship between device stress and predicted device lifetime.

Fig. 3 is a flowchart showing an example of the operation of the lifetime prediction device according to embodiment 1.

Fig. 4 is a diagram showing an example of hardware for realizing the lifetime prediction device.

Fig. 5 is a diagram showing another configuration example of the lifetime prediction device according to embodiment 1.

Fig. 6 is a diagram showing a configuration example of a lifetime prediction system realized by applying the lifetime prediction device according to embodiment 2.

Fig. 7 is a diagram showing an example of the content of the life predicting device according to embodiment 2 that causes the display unit to display.

Fig. 8 is a diagram showing a configuration example of a lifetime prediction device in accordance with embodiment 3.

Fig. 9 is a diagram showing an example of the content of the life predicting device according to embodiment 3 that causes the display unit to display.

Fig. 10 is a diagram showing another configuration example of the lifetime prediction device in accordance with embodiment 3.

Fig. 11 is a diagram showing a configuration example of a lifetime prediction device in accordance with embodiment 4.

Fig. 12 is a diagram showing a configuration example of a learning device that performs machine learning for realizing a device stress estimating unit of the lifetime prediction device.

Fig. 13 is a diagram for explaining a neural network.

Fig. 14 is a flowchart showing an example of the operation of the learning device.

Fig. 15 is a diagram showing a configuration example of an estimating device that implements a device stress estimating unit of the lifetime estimating device.

Fig. 16 is a flowchart showing an example of an operation of the lifetime prediction device according to embodiment 5 for estimating the lifetime of the object by the estimating device.

Fig. 17 is a diagram showing a configuration example of a learning device that performs machine learning for realizing a lifetime consumption estimating unit of the lifetime prediction device.

Fig. 18 is a diagram showing a configuration example of an estimating device that implements a lifetime consumption estimating unit of the lifetime predicting device.

Fig. 19 is a flowchart showing an example of an operation of the lifetime prediction device according to embodiment 6 for estimating the lifetime of an object by using the estimating device.

Detailed Description

The life predicting device, the life predicting system, the learning device, the estimating device, and the life predicting program according to the embodiment of the present invention will be described in detail below with reference to the drawings.

Embodiment 1

Fig. 1 is a diagram showing a configuration example of a lifetime prediction system realized by applying the lifetime prediction device according to embodiment 1. The life prediction system shown in fig. 1 includes a life prediction device 4, a motor system 100 to which the life prediction device 4 predicts the life, and a display unit 9 for displaying a result of the life prediction obtained by the life prediction device 4, and the like. The device state information 8 shown in fig. 1 is used when predicting the lifetime of the object by the lifetime prediction device 4.

The motor system 100 includes a motor control device 1, a motor 2, and a mechanism unit 3, and is driven based on a command from a higher-level controller, not shown.

The motor control device 1 mainly includes a converter circuit 11, a smoothing capacitor 12, and an inverter circuit 13.

The converter circuit 11 is configured by a plurality of rectifying elements and switching elements, and mainly converts the input three-phase ac voltage of the system power supply into a dc voltage by three-phase full-wave rectification, and outputs the dc voltage between the positive bus P and the negative bus N.

The smoothing capacitor 12 is disposed between the positive electrode bus line P and the negative electrode bus line N. The smoothing capacitor 12 removes a ripple component included in the dc voltage rectified and output between the positive electrode bus P and the negative electrode bus N by the converter circuit 11, and generates a stable bus voltage.

A 1 st current sensor 51 is provided between the positive electrode bus bar P and the smoothing capacitor 12. A voltage sensor 52 is provided between the positive electrode bus P and the negative electrode bus N.

The inverter circuit 13 is connected in parallel with the smoothing capacitor 12, and is composed of a plurality of rectifying elements and switching elements. The inverter circuit 13 controls the motor 2 by turning ON/OFF a switching element based ON a drive command from an inverter control unit, not shown.

The motor 2 is connected to the inverter circuit 13 via a power cable. A 2 nd current sensor 53 is provided in a power cable for the motor 2. The motor 2 is provided with an angle sensor 54. The motor 2 is coupled to the mechanism 3 via a coupler.

The mechanism 3 is composed of a ball screw 31, a nut 32, a table 33, a bearing 34 for supporting the ball screw 31, and the like. A position sensor 55 is provided to detect the position of the table 33.

The information detected by the various sensors provided in the motor system 100 is sensor information 5. Examples of the sensor information 5 include a bus current detected by the 1 st current sensor 51, a bus voltage detected by the voltage sensor 52, a motor current detected by the 2 nd current sensor 53, a motor speed detected by the angle sensor 54, a mechanical speed and a mechanical position detected by the position sensor 55, a temperature detected by a temperature sensor not shown, and a humidity detected by a humidity sensor.

The setting information 6 is information related to the operation of a device or a component preset in the motor system 100. Specifically, the setting information 6 is a carrier frequency used when the inverter control unit, not shown, of the inverter circuit 13 generates a drive command, a cooling state of the devices constituting the motor system 100, an altitude of a place where the devices constituting the motor system 100 are used, and the like. The cooling state of the device is, for example, an operation state of a cooling fan provided to the motor 2.

The calculation information 7 is information calculated using a part or all of the operating state of the mechanism unit 3 and the sensor information 5, and is the friction amount, the load inertia, the operating time, the machining condition, and the like of the mechanism unit 3 when the table 33 is moved.

The information obtained by integrating the sensor information 5, the setting information 6, and the calculation information 7 is device state information 8, which is information on the use states of the respective devices constituting the motor system 100. That is, the device state information 8 is composed of the sensor information 5, the setting information 6, and the calculation information 7. For convenience of explanation, the device state information 8 is described in this specification, but the device state information 8 also includes information on the use states of the components constituting the motor system 100. In addition, when the object to be subjected to life prediction is a component, the term "component" may be also referred to as "device". For example, "device life" sometimes refers to the life of a device, and sometimes refers to the life of a component.

In the configuration shown in fig. 1, the calculation information 7 may be generated outside the lifetime prediction device 4, but the calculation information 7 may be generated inside the lifetime prediction device 4 based on the sensor information 5.

The lifetime prediction device 4 has 1 or more lifetime calculation units 40-1, 40-2, … each including a device stress estimation unit 41, a lifetime consumption estimation unit 42, a lifetime prediction unit 43, and a lifetime storage unit 44. In the following description, when common matters are described without distinguishing between the life calculation units 40-1, 40-2, and …, they are collectively referred to as the life calculation unit 40. The lifetime calculation unit 40 is provided in the same number as the devices and components that are the objects of lifetime calculation among the devices and components that constitute the motor system 100. For example, when the number of parts to be subjected to life calculation is 5 among the devices and components constituting the motor system 100, the life predicting device 4 has a configuration including 5 life calculating units 40-1 to 40-5. The lifetime calculation unit 40 may be provided individually for each of the motor control device 1, the motor 2, and the mechanism unit 3, for example, in order to predict the lifetime of each of them, and may be provided individually for each of lifetime components among components constituting the motor system 100, for example, the smoothing capacitor 12, the semiconductor element, and the like of the motor control device 1.

The life storage unit 44 stores, as life data, a basic life, which is an initial life time at a device operation start time under a standard use condition, and a life consumed from the basic life, which is a consumed life. The basic lifetime is a lifetime in the case of using the device (or the component) under the conditions recommended by the manufacturer, for example, a lifetime of the device in the case of using the device under the recommended conditions described in the catalog of the device.

The device stress estimating unit 41 receives as input device state information 8 generated based on information acquired from the motor system 100 and the like, and estimates device stress based on the input device state information 8. The device stress estimating unit 41 outputs the estimated device stress to the lifetime consumption estimating unit 42 and the display unit 9.

The lifetime consumption estimating unit 42 takes the device stress as an input, and estimates the lifetime consumption from the input device stress. The lifetime consumption estimating unit 42 outputs the estimated lifetime consumption to the lifetime predicting unit 43.

Here, the lifetime consumption is lifetime consumed in each calculation cycle of the lifetime consumption estimating unit 42. The lifetime consumption amount from the start of the operation of the device is integrated to obtain the lifetime consumption included in the lifetime data. In the following description, the lifetime consumption is denoted by Δln, the calculation cycle of the lifetime consumption estimating unit 42 is denoted by Δt, and the lifetime consumption is denoted by Ln.

The lifetime predicting unit 43 predicts the lifetime of the object (device or component) whose lifetime is calculated by the lifetime calculating unit 40 based on the lifetime consumption Δln input from the lifetime consumption estimating unit 42 and lifetime data stored in the lifetime storing unit 44.

Specifically, the lifetime predicting unit 43 calculates the lifetime of the object, that is, the predicted device lifetime L, according to equation (1) based on the remaining lifetime Lr obtained by subtracting the consumed lifetime Ln from the basic lifetime, the lifetime consumption Δln, and the calculation cycle Δt.

L＝Lr/ΔLn×Δt…(1)

The lifetime predicting unit 43 outputs the predicted device lifetime L predicted by calculation to the display unit 9 so as to be notified to the outside. The lifetime prediction unit 43 updates the lifetime consumption Ln held by the lifetime storage unit 44 based on the lifetime consumption Δln. That is, the lifetime predicting unit 43 updates the consumed lifetime Ln by adding the lifetime consumption Δln to the consumed lifetime Ln held by the lifetime storing unit 44.

As an example, description will be given of an operation of the lifetime calculating unit 40 corresponding to the smoothing capacitor 12 of the motor control device 1, that is, an operation of the lifetime calculating unit 40 predicting the lifetime with respect to the smoothing capacitor 12.

The heat generated in the smoothing capacitor 12 is expressed by equation (2) with the internal temperature of the capacitor being Tx.

Tx＝Ta+ΔTx＝Ta+IR＾2×ESR×β…(2)

In equation (2), ta is the ambient temperature of the smoothing capacitor 12. Δtx is the temperature rise from ambient temperature. IR is a pulsating current flowing through the smoothing capacitor 12. ESR is the equivalent series resistance (Equivalent Series Resistance) of the smoothing capacitor 12. And b is the heat dissipation coefficient of the smoothing capacitor 12.

The aluminum electrolytic capacitor used as the smoothing capacitor 12 is degraded in lifetime according to the arrhenius law due to thermal stress accompanying the temperature rise. The lifetime L of the aluminum electrolytic capacitor is represented by formula (3).

L＝Lbc×[2＾{(To－Tx)/10}]×(Vo/Vx)＾K…(3)

In equation (3), lbc is a basic lifetime of rated voltage application and rated ripple voltage application at the upper limit temperature of use of the smoothing capacitor 12. To is the allowable internal temperature of the smoothing capacitor 12. Vo is the rated voltage of the smoothing capacitor 12. Vx is the direct voltage applied to the smoothing capacitor 12 in use. The power K is the applied voltage reduction rate of the smoothing capacitor 12.

If the above fit is made to the lifetime prediction device 4, the device stress estimating unit 41 can estimate by applying the equation (2), and the device state information 8 to be input can be set to the ambient temperature and the bus current, and the device stress to be output can be set to the internal temperature Tx of the capacitor. Here, the bus current corresponds to the ripple current IR. The equivalent series resistance ESR and the heat dissipation coefficient β are used as fixed values in the calculation.

Similarly, the lifetime consumption estimating unit 42 can estimate by applying expression (3). The lifetime consumption estimating unit 42 sets the input device stress as the internal temperature Tx of the capacitor, and outputs the lifetime L obtained from the internal temperature Tx and equation (3) as the lifetime consumption Δln. The lifetime predicting unit 43 calculates the predicted device lifetime L from the lifetime consumption Δln by the above equation (1).

The lifetime prediction device 4 outputs the predicted device lifetime L calculated by the lifetime calculation unit 40 as described above to the display unit 9 so as to be notified to the outside. The lifetime prediction device 4 outputs the acquired device state information 8 and the device stress estimated by the device stress estimating unit 41 to the display unit 9 so as to be notified to the outside.

The display unit 9 displays the predicted device lifetime L, the device state information 8, and the device stress outputted from the lifetime prediction device 4.

In fig. 1, the display 9 is provided on a display device such as a monitor, for example, which is external to the lifetime prediction device 4, but the display 9 may be provided inside the lifetime prediction device 4. That is, the lifetime prediction device 4 may output the predicted device lifetime L, the device state information 8, and the device stress to the outside of the lifetime prediction device 4.

Fig. 2 is a diagram showing an example of a relationship between device stress and predicted device lifetime. In FIG. 2, S _min Is the minimum value of the stress of the device, S _max Is the maximum of the device stress. L (L) _min Is the predicted device life at which device stress is minimal, L _max Is the predicted device life at which device stress is greatest. S is S _x L and L _n Respectively representing the device stress and predicted device lifetime under the current operating conditions. As described above, the relationship between the current operating condition, the device stress, and the lifetime can be easily grasped. In addition, in the present embodiment, since the predicted device lifetime is calculated based on the lifetime consumption amount, the predicted device lifetime when the device stress is changed can be easily calculated.

The operation of the lifetime prediction device 4 to predict the lifetime of the device or the component can be represented by a flowchart shown in fig. 3. Fig. 3 is a flowchart showing an example of the operation of the lifetime prediction device 4 according to embodiment 1.

As shown in fig. 3, the lifetime prediction device 4 first collects information used for lifetime prediction (step S1). That is, the lifetime prediction device 4 obtains the device state information 8 from the outside.

The lifetime prediction device 4 then estimates the device stress (step S2). That is, the device stress estimating unit 41 of the lifetime calculating unit 40 constituting the lifetime predicting device 4 estimates the device stress of the lifetime predicting object based on the device state information 8.

The lifetime prediction device 4 then estimates the lifetime consumption amount (step S3). That is, the life consumption estimating unit 42 of the life calculating unit 40 constituting the life predicting device 4 estimates the life consumption based on the device stress of the life predicting object.

The lifetime prediction device 4 then calculates a predicted device lifetime (step S4). That is, the life predicting unit 43 of the life calculating unit 40 constituting the life predicting device 4 calculates the predicted device life based on the life consumption amount of the life predicting object.

The lifetime prediction device 4 then notifies the collected information, device stress, and predicted device lifetime (step S5). That is, the lifetime predicting device 4 outputs the information collected in step S1, the device stress estimated in step S2, and the predicted device lifetime calculated in step S4 to the display unit 9, and causes the display unit 9 to display the information, thereby notifying the user of the information.

Next, hardware for realizing the lifetime prediction device 4 according to embodiment 1 will be described. The lifetime prediction device 4 can be realized by, for example, the hardware shown in fig. 4, that is, the processor 101 and the memory 102. Fig. 4 is a diagram showing an example of hardware for realizing the lifetime prediction device 4.

Examples of the processor 101 are a CPU (also referred to as Central Processing Unit, central processing unit, arithmetic unit, microprocessor, microcomputer, DSP (Digital Signal Processor)) or a system LSI (Large Scale Integration). Examples of the memory 102 are nonvolatile or volatile semiconductor memories such as RAM (Random Access Memory) and ROM (Read Only Memory) and flash memories, magnetic disks, and the like.

The device stress estimating unit 41, the lifetime consumption estimating unit 42, and the lifetime predicting unit 43 constituting each lifetime calculating unit 40 of the lifetime predicting device 4 are realized by executing a lifetime predicting program for operating each unit as the processor 101. A life prediction program for operating as the device stress estimating unit 41, the life consumption estimating unit 42, and the life predicting unit 43 is stored in advance in the memory 102. The processor 101 reads out and executes the lifetime prediction program from the memory 102, thereby operating as the device stress estimating unit 41, the lifetime consumption estimating unit 42, and the lifetime predicting unit 43.

The lifetime storage unit 44 of each lifetime calculation unit 40 constituting the lifetime prediction device 4 is realized by a memory 102.

The processor 101 and the memory 102 shown in fig. 4 may be hardware constituting an electronic computer. That is, the lifetime prediction device 4 may be realized by an electronic computer having a processor 101 and a memory 102, and a lifetime prediction program executed by the processor 101 of the electronic computer. The functions of the lifetime prediction device 4 may be realized by a plurality of electronic computers operating in cooperation with each other.

The life prediction program for operating as the device stress estimating unit 41, the life consumption estimating unit 42, and the life predicting unit 43 is stored in the memory 102 in advance, but the present invention is not limited thereto. The lifetime prediction program may be provided to a user in a state of being written to a recording medium such as CD (Compact Disc) -ROM or DVD (Digital Versatile Disc) -ROM, and may be installed in the memory 102 by the user. The lifetime prediction program may be provided to a user via a network such as the internet.

The hardware for realizing the lifetime prediction device 4 is described, but other lifetime prediction devices described later can be realized by the same hardware.

As described above, according to the lifetime prediction device 4 of the present embodiment, a plurality of factors affecting lifetime corresponding to the operation environment can be simply calculated as 1 device stress and provided to the user together with the predicted device lifetime, and therefore the user can easily grasp the operation environment of the machine and determine the maintenance condition.

Further, the lifetime prediction device may be configured as shown in fig. 5. Fig. 5 is a diagram showing another configuration example of the lifetime prediction device according to embodiment 1. As shown in fig. 5, the lifetime predicting device 4a includes lifetime calculating parts 40a-1, 40a-2, … instead of the lifetime calculating parts 40-1, 40-2, … of the lifetime predicting device 4. The lifetime calculating unit 40a (lifetime calculating units 40a-1, 40a-2, …) is configured by replacing the lifetime consumption estimating unit 42 and the lifetime storing unit 44 of the lifetime calculating unit 40 with the lifetime consumption estimating unit 42a and the lifetime storing unit 44 a. In fig. 5, the same components as those of the lifetime prediction device 4 shown in fig. 1 are denoted by the same reference numerals as in fig. 1. The same reference numerals as those in fig. 1 are given to the components, and the description thereof is omitted.

In addition to the device stress output from the device stress estimating unit 41, a degradation coefficient calculated from the basic lifetime and the consumed lifetime is input from the lifetime storing unit 44a to the lifetime consumption estimating unit 42a. The degradation coefficient may be calculated by the lifetime storage unit 44a or by the lifetime prediction unit 43. In addition, the degradation coefficient may be calculated by the lifetime consumption estimating unit 42a, instead of being input to the lifetime consumption estimating unit 42a, by inputting information used for calculating the degradation coefficient to the lifetime consumption estimating unit 42a.

The degradation coefficient Lk input to the lifetime consumption estimating unit 42a is a result of dividing the remaining lifetime Lr, which is a value obtained by subtracting the consumed lifetime from the basic lifetime, by the basic lifetime. In the case of the above-described configuration, the life consumption amount estimated by the life consumption amount estimating unit 42a can be corrected using the coefficient Lk associated with degradation as an influence amount associated with degradation of the prediction target. For example, in an aluminum electrolytic capacitor used as the smoothing capacitor 12, if degradation progresses, the electrostatic capacity decreases or the equivalent series resistance increases, and thus, as degradation progresses, an error in the estimated result of the lifetime increases. In the case of predicting the life of the component whose life consumption is affected by the deterioration as described above, the life predicting device 4a having the structure shown in fig. 5 is suitable, and the prediction accuracy can be improved.

The correction of the estimated lifetime consumption by the lifetime consumption estimating unit 42a is performed by multiplying the lifetime consumption estimated by the same method as that of the lifetime consumption estimating unit 42 by the degradation coefficient Lk, for example. The degradation coefficient may be included in a calculation formula used for estimating the lifetime consumption, and the lifetime prediction unit 43 may correct the degradation coefficient Lk so as to be updated each time the lifetime prediction unit calculates the lifetime of the device.

Embodiment 2

Fig. 6 is a diagram showing a configuration example of a lifetime prediction system realized by applying the lifetime prediction device according to embodiment 2. The life prediction system according to embodiment 2 includes the same life prediction device 4, information input device 300, and motion estimation unit 200 as in embodiment 1.

In embodiment 1, sensor information 5 and calculation information 7 obtained by actually operating the motor system 100 are input to the lifetime prediction device 4 together with the setting information 6. In contrast, in embodiment 2, the operation estimating unit 200 simulates the operation of the motor system 100 under the condition input via the information input device 300, and generates the sensor information 5 and the calculation information 7.

The information input device 300 is, for example, an electronic computer executing software for controlling the machine tool CNC (Computerized Numerically Controlled). The input information to the information input device 300 includes an operation condition (an operation program, an operation mode, parameters, and the like), mechanical information (a mechanical structure, and the like), and an operation environment (an ambient temperature, a power supply environment, and the like). The information input device 300 determines a movement command, a simulation condition (operation time, a mechanical prediction model to be a simulation target), and the like based on the input information.

The operation estimation unit 200 is a mechanical prediction model including the motor system prediction model selected by the information input device 300, and performs simulation according to the operation condition related to the input from the information input device 300, and outputs the sensor information 5. The motion estimation unit 200 can be realized by, for example, an electronic computer and a program for causing the electronic computer to execute simulation.

The lifetime prediction device 4 according to the present embodiment is the same as the lifetime prediction device 4 shown in fig. 1 described in embodiment 1. The lifetime prediction device 4 shown in fig. 6 may be replaced with a lifetime prediction device 4a shown in fig. 5.

Fig. 7 is a diagram showing an example of the content displayed on the display unit 9 by the lifetime prediction device 4 according to embodiment 2. As shown in fig. 7, the lifetime prediction device 4 causes the display unit 9 to display the input conditions and the output results.

In the example shown in fig. 7, the display unit 9 displays mechanical information, operation conditions, and operation environments as input conditions. The display unit 9 displays the device stress and the predicted device lifetime as well as the constituent devices and constituent components (motor #1, motor #2, inverter unit, cooling fan, etc.) determined by the mechanical structure as the output result. The device stress was set to 100% when the device had reached the basic lifetime.

By adopting the above-described mode, the user can easily grasp the current device stress and the predicted device lifetime under the selected conditions. Further, since the operation condition and the selected machine structure can be changed in accordance with the obtained predicted device life and device stress, the predicted device life and device stress can be used in selecting the structural device and the component of the machine.

As described above, according to the life prediction system of embodiment 2, the device stress and the predicted device life under the specified operating conditions can be grasped without operating the industrial machine such as the motor system 100, and therefore the operating conditions and the machine configuration can be determined in accordance with the life desired by the user.

Embodiment 3

Fig. 8 is a diagram showing a configuration example of the lifetime prediction device 4b according to embodiment 3. The device state information 8 shown in fig. 8 is generated by the method shown in embodiment 1 or embodiment 2, similarly to the device state information 8 described in embodiment 1.

The lifetime prediction device 4b according to embodiment 3 is configured such that a modified lifetime calculation unit 400 is added to the lifetime prediction device 4 according to embodiment 1. The lifetime prediction device 4b shown in fig. 8 is configured to have 1 set of lifetime calculation units 40 and modified lifetime calculation units 400, but may have 2 or more sets.

The modified lifetime calculation unit 400 includes a lifetime comparison unit 401, an operation condition modification unit 402, a modified device stress estimation unit 403, a modified lifetime consumption estimation unit 404, and a modified lifetime prediction unit 405.

The lifetime comparing unit 401 compares the predicted device lifetime calculated by the lifetime predicting unit 43 with the target device lifetime set in advance by the user. The target device lifetime is stored in the lifetime storage unit 44. When the result of the comparison is "predicted device lifetime < target device lifetime", the operation condition changing unit 402 changes the operation condition.

The operation condition changing unit 402 holds in advance the operation condition associated with the device state information 8 corresponding to the device to be lifetime prediction.

As an example, the operation of the lifetime prediction device 4b in the case where the lifetime prediction target is the motor control device 1 will be described.

When the comparison result obtained by the lifetime comparison unit 401 is "predicted device lifetime < target device lifetime", the operating condition changing unit 402 sets motor current, motor rotation speed, mechanical speed, and the like, which are information that can be controlled by the motor control device 1, among the information included in the device state information 8, as targets, and changes the operating condition so that the device stress is reduced. The operating conditions herein are time constant (acceleration) at the time of acceleration and deceleration, operating speed, machining load, and the like.

Based on the changed device state information 8 obtained by changing the operation condition by the operation condition changing unit 402, the changed device stress estimating unit 403, the changed lifetime consumption estimating unit 404, and the changed lifetime predicting unit 405 execute processing, respectively, to calculate the predicted device lifetime. Here, the modified device stress estimating unit 403 performs the same processing as the device stress estimating unit 41 described above to estimate the device stress. The modified lifetime consumption estimating unit 404 performs the same processing as the lifetime consumption estimating unit 42 described above to estimate the lifetime consumption. The modified lifetime prediction unit 405 performs the same processing as the lifetime prediction unit 43 to obtain a predicted device lifetime, and outputs the predicted device lifetime as a modified predicted device lifetime.

The change predicted device lifetime is input to the lifetime comparing unit 401, and compared with the target device lifetime.

When the comparison result obtained by the lifetime comparison unit 401 is "the predicted device lifetime is greater than or equal to the target device lifetime", the lifetime change calculation unit 400 displays the operation condition to be changed, the device state information 8 (the device state information 8 after the change) obtained by changing the operation condition, and the device stress and the predicted device lifetime obtained by using the device state information 8 after the change on the display unit 9.

On the other hand, when the comparison result obtained by the lifetime comparison unit 401 is "the predicted device lifetime is less than the target device lifetime", the operation condition is changed by the operation condition changing unit 402 so that the device stress is further reduced, and the predicted device lifetime is calculated again by performing processing by each of the changed device stress estimating unit 403, the changed lifetime consumption estimating unit 404, and the changed lifetime predicting unit 405 based on the changed device state information 8 obtained by the change.

The same process is repeated until the comparison result obtained by the lifetime comparison unit 401 becomes "the predicted device lifetime is equal to or longer than the target device lifetime".

The target device lifetime stored in the lifetime storage unit 44 is updated by subtracting the lifetime consumption estimated by the lifetime consumption estimating unit 42 for each calculation cycle of the lifetime predicting unit 43.

Fig. 9 is a diagram showing an example of the content of the lifetime prediction device 4b according to embodiment 3 that causes the display unit 9 to display. As shown in fig. 9, the lifetime prediction device 4b causes the display unit 9 to display the input conditions and the output results.

As shown in fig. 9, the lifetime prediction device 4b causes the display 9 to display the input conditions, the output results, the content of the change in the operation conditions, and the device state (device stress, predicted device lifetime) after the change in the operation conditions.

The "output result" shown in fig. 9 is added with the target operation time to the "output result" shown in fig. 7. The "input conditions" shown in fig. 9 are added with the operation conditions (target operation time) to the "input conditions" shown in fig. 7. Here, the operating conditions to be changed may be plural. Fig. 9 shows an example of changing 2 operation conditions of the time constant and the stop time.

By adopting the above-described mode, the user can grasp whether or not the target device lifetime is reached and the operating conditions for satisfying the target device lifetime. Fig. 9 shows an example in which the change content of the operation condition, the device stress, and the predicted device lifetime are displayed on the display unit 9, but the device stress is not essential. At least the change content of the operation condition and the predicted lifetime of the device are displayed, information required by the user can be provided.

As shown in fig. 10, the life predicting device 4c may be configured to notify the control device 10 of the changed operation condition and control the device so as to reduce the stress. Fig. 10 is a diagram showing another configuration example of the lifetime prediction device in accordance with embodiment 3.

When the life predicting system is implemented by using the life predicting device 4c shown in fig. 10, the actual operation of the object to be life-predicted is automatically controlled so as to satisfy the target device life. Here, when the target object for life prediction is the motor system 100, the motor control device 1 becomes the control device 10 of fig. 10.

As described above, according to the lifetime prediction device 4b of embodiment 3, the user can grasp the operation condition for satisfying the target device lifetime required by the user. Further, according to the lifetime prediction device 4c of embodiment 3, the operation condition can be automatically changed so as to satisfy the target lifetime of the device.

The lifetime calculation unit 40 of the lifetime prediction device 4b can be replaced with the lifetime calculation unit 40a of the lifetime prediction device 4a described in embodiment 1. Similarly, the lifetime calculating unit 40 of the lifetime predicting device 4c may be replaced with the lifetime calculating unit 40a.

Embodiment 4

Fig. 11 is a diagram showing a configuration example of a lifetime prediction device 4d according to embodiment 4. The device state information 8 shown in fig. 11 is generated by the method shown in embodiment 1 or embodiment 2, similarly to the device state information 8 described in embodiment 1.

The basic configuration of the lifetime prediction device 4d according to the present embodiment is the same as that of the lifetime prediction device 4 according to embodiment 1, but differs from that of the lifetime prediction device 4 in that the consumed lifetime held in the lifetime storage unit 44 can be updated by the lifetime consumption calculated by the external lifetime prediction device 450. In the following description, the lifetime consumption calculated by the lifetime consumption estimating unit 42 of the lifetime prediction device 4d is referred to as the 1 st lifetime consumption, and the lifetime consumption calculated by the external lifetime prediction device 450 is referred to as the 2 nd lifetime consumption.

The external lifetime prediction device 450 estimates the 2 nd lifetime consumption amount of the object by offline estimation.

The offline estimation here is an estimation performed by a dedicated operation in order to estimate the life of the object consumed during periodic maintenance or the like. In the offline estimation, the lifetime consumption can be calculated with high accuracy by performing a dedicated operation for estimation, but there is a problem in that the machine cannot be operated during this period.

For example, as an offline estimation of the life consumption of the smoothing capacitor 12, a method is known in which, when the main power supply of the motor control device 1 is turned OFF, a constant current I is caused to flow for a constant period Δτ, and the voltage drop Δv of the smoothing capacitor 12 during the period is measured. Then, the remaining capacity C of the smoothing capacitor 12 is calculated by applying the formula "c= (i×Δτ)/Δv", and the calculated remaining capacity C is compared with the allowable value of the capacitor capacity to diagnose the life deterioration. With this method, the external lifetime prediction device 450 diagnoses lifetime degradation of the object, and sets the lifetime degradation amount obtained by the diagnosis as the 2 nd lifetime consumption amount.

The OFF-line estimation allows the remaining life of the smoothing capacitor 12 to be calculated with high accuracy without calculating the thermal stress amount, but the main power supply of the motor control device 1 needs to be turned ON/OFF during maintenance, and the mechanical operation needs to be stopped.

Here, the online estimation is a real-time estimation that does not perform a dedicated operation for estimation. Since no dedicated operation is required for online estimation, estimation of the lifetime consumption can be performed while continuing the mechanical operation.

The lifetime prediction device 4d according to the present embodiment performs the 1 st lifetime consumption estimation and the update of the lifetime consumption held by the lifetime storage unit 44 by the online estimation during the operation of the industrial machine at the time of the steady operation, and performs the 2 nd lifetime consumption estimation and the update of the lifetime consumption held by the lifetime storage unit 44 by the offline estimation during the maintenance of the industrial machine, thereby obtaining a more accurate lifetime estimation result during the operation of the machine at the time of the steady operation.

Further, the off-line estimation is performed in the external lifetime prediction device 450, which is the outside of the lifetime prediction device 4d, but may be performed in the lifetime prediction device 4 d. For example, the lifetime consumption estimating unit that performs offline estimation may be provided in the lifetime predicting device 4 d.

According to the lifetime prediction device 4d of embodiment 4, the estimation accuracy of the lifetime of the prediction device can be improved.

Embodiment 5

The device stress estimating unit 41 of the lifetime prediction device 4 according to embodiment 1 can be realized by applying machine learning. Therefore, in the present embodiment, a method of implementing the device stress estimating unit 41 by applying machine learning, specifically, a learning device that generates a trained model for implementing the device stress estimating unit 41 and an estimating device that estimates the device stress using the trained model will be described. Further, as an example, the case where the machine learning is applied to the device stress estimating unit 41 of the lifetime predicting device 4 is described, but the machine learning may be applied to the device stress estimating units 41 of the lifetime predicting devices 4a to 4 d.

Next, a description will be given of a learning stage in which a trained model for realizing the device stress estimating unit 41 is generated by the learning device, and a stage in which the device stress is estimated by the estimating device using the trained model, thereby realizing effective use of the device stress estimating unit 41.

< learning phase >)

Fig. 12 is a diagram showing a configuration example of a learning device 61 that performs machine learning for realizing the device stress estimating unit 41 of the lifetime prediction device 4. The learning device 61 includes a data acquisition unit 62 and a model generation unit 63. The learning device 61 may be provided inside the lifetime prediction device 4 or may be provided outside the lifetime prediction device 4. The trained model storage unit 71 shown in fig. 12 may be provided outside the learning device 61 or inside the learning device 61.

The data acquisition unit 62 acquires, as learning data, device state information 8 including at least one of ambient temperature, bus voltage, bus current, motor current, carrier frequency, motor applied voltage, cooling state of the device, rotation speed of the motor, friction amount, transport mass, load inertia, feed speed, mechanical position, humidity, operation time, and processing condition, and device stress. The device stress corresponds to machine-learned teacher data.

Here, the device stress acquired by the data acquisition unit 62 varies according to the generated trained model.

For example, when a trained model of a capacitor connected to the motor control device 1 or built in the motor control device 1 is generated, the device stress is set to the capacitor temperature.

In addition, for example, when a trained model of the semiconductor element incorporated in the motor control device 1 is generated, the device stress is set to the junction temperature. In the semiconductor element, it is known that deterioration inside the semiconductor element progresses according to junction temperature.

In addition, for example, when a trained model of the motor 2 is generated, the device stress is set to the coil temperature of the motor 2. In a general motor including the motor 2, it is known that deterioration of an insulating material of a coil progresses according to temperature.

In addition, for example, when a trained model of the mechanism section 3 is generated, the device stress is set to an axial load. Among the structural members (ball screw 31, bearings 34, etc.) of the mechanism section 3, it is known that axial load affects the life time.

The model generating unit 63 learns the relationship between the device state information 8 and the device stress based on learning data created by the combination of the device state information 8 and the device stress output from the data acquiring unit 62. That is, the model generating unit 63 generates a trained model that outputs the optimal device stress if the device state information 8 of the lifetime prediction target is input. Here, the learning data is data that correlates the device state information 8 and the device stress with each other.

The learning device 61 is configured to learn the device stress corresponding to the device state information 8, and may be connected to the lifetime prediction device 4 via a network, for example, and may operate as a device separate from the lifetime prediction device 4. The learning device 61 may be incorporated in the lifetime prediction device 4. The learning device 61 may be present on the cloud server.

The learning algorithm used by the model generating unit 63 may be a known algorithm such as teacher learning, non-teacher learning, or reinforcement learning. As an example, a case where a neural network is applied will be described.

The model generating unit 63 learns the device stress corresponding to the device state information 8 by so-called teacher learning, for example, in accordance with a neural network model. Here, teacher learning means a method of giving a data set of input and result (label) to the learning device 61, thereby learning the features existing in the learning data, and estimating the result from the input.

The neural network is composed of an input layer composed of a plurality of neurons, an intermediate layer (hidden layer) composed of a plurality of neurons, and an output layer composed of a plurality of neurons. The intermediate layer may be 1 layer or greater than or equal to 2 layers.

Fig. 13 is a diagram for explaining a neural network. For example, in the case of the 3-layer neural network shown in fig. 13, after a plurality of inputs are input to the input layers X (X1 to X3), the values are given weights W1 (W11 to W16) and input to the intermediate layers Y (Y1 to Y2), and further, the results are given weights W2 (W21 to W26) and output from the output layers Z (Z1 to Z3). The output result varies according to the values of the weights W1 and W2.

The neural network used by the model generation unit 63 of the learning device 61 according to the present embodiment learns the device stress corresponding to the device state information 8 by so-called teacher learning, based on the learning data created based on the combination of the device state information 8 and the device stress acquired by the data acquisition unit 62.

That is, the neural network learns by adjusting the weights W1 and W2 so that the device stress approaches the result output from the output layer by inputting the device state information 8 to the input layer.

The model generation unit 63 performs the above learning to generate and output a trained model.

The trained model storage unit 71 stores the trained model output from the model generation unit 63.

Next, the operation of generating a trained model by the learning device 61 will be described with reference to fig. 14. Fig. 14 is a flowchart showing an example of the operation of the learning device 61.

The learning device 61 first acquires learning data (step S11). Specifically, the data acquisition unit 62 acquires the device state information 8 and the device stress, and correlates them to create learning data. The device state information 8 and the device stress are acquired at the same time, but the device state information 8 and the device stress may be acquired at different timings as long as the device state information 8 and the device stress can be acquired in association with each other.

The learning device 61 next performs learning processing (step S12). Specifically, the model generating unit 63 generates a trained model by learning the device stress corresponding to the device state information 8 through so-called teacher learning, based on the learning data created based on the combination of the device state information 8 and the device stress acquired by the data acquiring unit 62.

The learning device 61 then causes the trained model to be stored in the trained model storage unit 71 (step S13). Specifically, the model generation unit 63 outputs the trained model generated in step S12, and the trained model storage unit 71 stores the trained model.

< valid use phase >

Next, a description will be given of an effective use stage in which the device stress is estimated by using the trained model.

Fig. 15 is a diagram showing a configuration example of an estimating device 65 of the device stress estimating unit 41 that realizes the lifetime estimating device 4. The estimating device 65 includes a data acquiring unit 66 and an estimating unit 67. The estimation device 65 may be provided inside the lifetime prediction device 4 or may be provided outside the lifetime prediction device 4. The trained model storage unit 71 shown in fig. 15 corresponds to the trained model storage unit 71 shown in fig. 12, and stores the trained model generated by the learning device 61.

The data acquisition unit 66 acquires the device state information 8.

The estimating unit 67 estimates the device stress corresponding to the device state information 8 using the trained model stored in the trained model storage unit 71. That is, the estimating unit 67 inputs the trained model stored in the trained model storage unit 71 into the device state information 8 acquired by the data acquiring unit 66, performs estimation, and acquires the device stress as an estimation result.

The estimating device 65 is used for estimating the device stress corresponding to the device state information 8, but may be connected to the lifetime predicting device 4 via a network, for example, and may operate as a device separate from the lifetime predicting device 4. The estimating device 65 may be incorporated in the lifetime predicting device 4. The estimation device 65 may be present on the cloud server.

In the present embodiment, the device stress is estimated by the estimating device 65 using the trained model generated by the learning device 61 constituting the lifetime predicting device 4, but the estimating device 65 may acquire the trained model generated outside the lifetime predicting device 4 and estimate the device stress using the trained model.

Next, an operation of estimating the device stress by the estimating device 65 and an operation of estimating the lifetime of the object from the estimated device stress will be described with reference to fig. 16. Fig. 16 is a flowchart showing an example of an operation of estimating the lifetime of the object by the estimating device 65 in the lifetime estimating device according to embodiment 5.

The estimating device 65 first acquires data used for estimating the device stress (step S21). Specifically, the data acquisition unit 66 acquires the device state information 8.

The estimation device 65 then inputs the data acquired in step S21 into the trained model stored in the trained model storage unit 71 (step S22), and obtains an estimation result. That is, the estimating unit 67 inputs the device state information 8 acquired by the data acquiring unit 66 into the trained model stored in the trained model storage unit 71, and acquires the device stress output from the trained model in association with the input.

The estimating device 65 then outputs the device stress as the result of the estimation (step S23), and the lifetime consumption estimating unit 42 and the lifetime predicting unit 43 estimate the lifetime of the device based on the device stress (step S24). The estimated device lifetime is outputted from the lifetime predicting unit 43 to the display unit 9 as a predicted device lifetime.

As a result, the factors of the change in the operating environment can be estimated as the device stress from the device state information 8 including the information that can be acquired by the motor system 100, and the lifetime can be predicted easily and with high accuracy.

As described above, the device stress estimating unit 41 of the lifetime prediction device 4 can be realized by machine learning.

In the present embodiment, the case where teacher learning is applied to the learning algorithm used by the model generating unit 63 has been described, but the present invention is not limited to this. As for the learning algorithm, reinforcement learning, non-teacher learning, half-teacher learning, or the like can be applied in addition to teacher learning.

The model generating unit 63 may learn the device stress according to learning data created for a plurality of industrial machines. The model generating unit 63 may acquire learning data from a plurality of industrial machines used in the same area, or learn the device stress by using learning data collected from a plurality of industrial machines operating independently in different areas. In addition, the industrial machine that collects learning data may be added to or removed from the subject in the middle of the process. The learning device that learns the device stress with respect to one industrial machine may be applied to another industrial machine, and the device stress may be relearned and updated with respect to the other industrial machine.

As a Learning algorithm used in the model generating unit 63, deep Learning (Deep Learning) for Learning the extraction of the feature quantity itself may be used, and machine Learning may be performed according to other known methods, such as genetic programming, functional logic programming, and support vector machine.

Embodiment 6

In embodiment 5, the case where the device stress estimating unit 41 of the lifetime predicting device 4 is realized by applying machine learning has been described, but the lifetime consumption estimating unit 42 may be realized by applying machine learning. Therefore, in the present embodiment, a method of implementing the lifetime consumption estimating unit 42 by applying machine learning, specifically, a learning device that generates a trained model for implementing the lifetime consumption estimating unit 42 and an estimating device that estimates lifetime consumption using the trained model will be described. Further, although the case where the machine learning is applied to the life consumption amount estimation unit 42 of the life prediction device 4 is described as an example, the machine learning may be applied to the life consumption amount estimation unit 42a of the life prediction device 4a, or the machine learning may be applied to the life consumption amount estimation units 42 of the life prediction devices 4b to 4 d.

Next, description will be made of a learning stage in which a trained model for realizing the lifetime consumption estimating unit 42 is generated by the learning device, and a stage in which the lifetime consumption is estimated by the estimating device using the trained model, thereby realizing effective use of the lifetime consumption estimating unit 42.

< learning phase >)

Fig. 17 is a diagram showing a configuration example of a learning device 81 that performs machine learning for realizing the lifetime consumption estimating unit 42 of the lifetime predicting device 4. The learning device 81 includes a data acquisition unit 82 and a model generation unit 83. The learning device 81 may be provided inside the lifetime prediction device 4 or may be provided outside the lifetime prediction device 4. The trained model storage unit 72 shown in fig. 17 may be provided outside the learning device 81 or inside the learning device 81.

The data acquisition unit 82 acquires the device stress and the lifetime consumption as learning data. The lifetime consumption corresponds to teacher data for machine learning. The lifetime consumption is calculated based on, for example, actual lifetime data, which is the time until the device is damaged when the device is operated under a constant device stress, and the calculation period Δt of the lifetime consumption estimating unit 42. The data acquisition unit 82 may acquire the degradation coefficient as learning data in addition to the device stress.

The model generating unit 83 learns the relationship between the device stress and the lifetime consumption amount based on learning data created according to the combination of the device stress and the lifetime consumption amount output from the data acquiring unit 82. That is, the model generating unit 83 generates a trained model that outputs the optimal lifetime consumption amount if the device stress of the target object for lifetime prediction is input. Here, the learning data is data that correlates the device stress and the lifetime consumption amount with each other.

The learning device 81 is used for learning the lifetime consumption amount according to the device stress, and may be connected to the lifetime prediction device 4 via a network, for example, and may operate as a device separate from the lifetime prediction device 4. The learning device 81 may be incorporated in the lifetime prediction device 4. The learning device 81 may be present on the cloud server.

The learning algorithm used by the model generating unit 83 can be a known algorithm such as teacher learning, non-teacher learning, reinforcement learning, and the like, as in the model generating unit 63 of the learning device 61.

The model generating unit 83 generates the training model in the same manner as the model generating unit 63 of the learning device 61 generates the training model, except that learning data used for the operation process of the training model is different. Therefore, the details of the operation of generating the trained model by the model generating unit 83 are omitted.

The model generation unit 83 outputs the generated trained model to the trained model storage unit 72.

The trained model storage unit 72 stores the trained model output from the model generation unit 83.

< valid use phase >

Fig. 18 is a diagram showing a configuration example of an estimating device 85 that implements the lifetime consumption estimating unit 42 of the lifetime predicting device 4. The estimating device 85 includes a data acquiring unit 86 and an estimating unit 87. The estimation device 85 may be provided inside the lifetime prediction device 4 or may be provided outside the lifetime prediction device 4. The trained model storage unit 72 shown in fig. 18 corresponds to the trained model storage unit 72 shown in fig. 17, and stores the trained model generated by the learning device 81.

The data acquisition unit 86 acquires the device stress.

The estimating unit 87 estimates the lifetime consumption amount corresponding to the device stress by using the trained model stored in the trained model storage unit 72. That is, the estimating unit 87 performs estimation by inputting the device stress acquired by the data acquiring unit 86 into the trained model stored in the trained model storage unit 72, and acquires the lifetime consumption as an estimation result.

The estimating device 85 is used to estimate the lifetime consumption amount corresponding to the device stress, but may be connected to the lifetime predicting device 4 via a network, and may operate as a device separate from the lifetime predicting device 4. The estimating device 85 may be incorporated in the lifetime predicting device 4. The estimation device 85 may be present on the cloud server.

In the present embodiment, the description has been made using the trained model generated by the learning device 81 constituting the lifetime prediction device 4, and the estimation device 85 estimates the lifetime consumption amount, but the estimation device 85 may acquire the trained model generated outside the lifetime prediction device 4 and estimate the lifetime consumption amount using the trained model.

Next, an operation of estimating the device stress by the estimating device 85 and an operation of estimating the lifetime of the object from the estimated device stress will be described with reference to fig. 19. Fig. 19 is a flowchart showing an example of an operation of estimating the lifetime of the object by the lifetime estimating device according to embodiment 6 using the estimating device 85.

The estimating device 85 first obtains data used for estimating the lifetime consumption amount (step S31). Specifically, the data acquisition unit 86 acquires the device stress.

The estimation device 85 then inputs the data acquired in step S31 to the trained model stored in the trained model storage unit 72 (step S32), and obtains an estimation result. That is, the estimating unit 87 inputs the device stress acquired by the data acquiring unit 86 to the trained model stored in the trained model storage unit 72, and acquires the lifetime consumption amount output from the trained model in association with the input.

The estimating device 85 then outputs the lifetime consumption amount as the estimation result (step S33), and the lifetime predicting unit 43 estimates the lifetime of the device based on the lifetime consumption amount (step S34). The estimated device lifetime is outputted from the lifetime prediction unit 43 to the display unit 9 via the predicted device lifetime.

As described above, the lifetime consumption estimating unit 42 of the lifetime prediction device 4 can be realized by machine learning.

The learning device 81 may include the degradation coefficient in the learning data used for machine learning, and the model generation unit 83 may generate a model trained by learning for estimating the life consumption amount based on the device stress and the degradation coefficient. In this case, the estimating device 85 obtains a degradation coefficient in addition to the device stress, and estimates the lifetime consumption amount based on the obtained device stress and degradation coefficient.

The configuration shown in the above embodiment is an example, and other known techniques may be combined, or the embodiments may be combined with each other, and a part of the configuration may be omitted or changed without departing from the scope of the present invention.

Description of the reference numerals

A motor control device, a motor, a 3-mechanism unit, a 4, 4a, 4b, 4c, 4d life prediction unit, 5 sensor information, 6 setting information, 7 calculation information, 8 device state information, 9 display unit, 10 control unit, 11 converter circuit, 12 smoothing capacitor, 13 inverter circuit, 31 ball screw, 32 nut, 33 table, 34 bearing, 40-1 to 40-5, 40a-1, 40a-2 life calculation unit, 41 device stress estimation unit, 42a life consumption amount estimation unit, 43 life prediction unit, 44a life storage unit, 51 current sensor (1 st current sensor), 52 voltage sensor, 53 current sensor (2 nd current sensor), 54 angle sensor, 55 position sensor, 61, 81 learning unit, 62, 66, 82, 86 data acquisition unit, 63, 83 model generation unit, 65, 85 estimation unit, 67, 87 estimation unit, 71, 72 trained model storage unit, 100 motor system, 200 estimation unit, 300 calculation unit, 401, 400 life calculation unit, 401, and 404 change motion condition comparison operation condition estimation unit, and 404 change operation condition prediction unit, and a change operation condition prediction unit, respectively.

Claims

1. A lifetime prediction device is characterized by comprising:

a lifetime storage unit that holds, as lifetime data, a basic lifetime of an object, which is a device or a component constituting an industrial machine, and a lifetime, which is a lifetime consumed from the basic lifetime;

a device stress estimating unit that estimates a device stress based on device state information, which is information on a use state of the object;

a lifetime consumption estimating unit that estimates lifetime consumption, which is lifetime of the object consumed for each specified period, based on the device stress; and

a lifetime predicting unit that predicts a lifetime of the predicting device of the object based on the lifetime data and the lifetime consumption held by the lifetime storing unit,

the device stress estimated by the device stress estimating unit and the predicted device lifetime predicted by the lifetime predicting unit are outputted so as to be notified to the outside.

2. The life predicting apparatus according to claim 1, wherein,

the lifetime predicting unit predicts the lifetime of the predicting device of the object each time the lifetime consumption is input from the lifetime consumption estimating unit, and updates the lifetime consumption included in the lifetime data held by the lifetime storing unit based on the input lifetime consumption.

3. The life predicting device according to claim 1 or 2, wherein,

the device state information includes sensor information detected by a sensor provided to the object and information related to an operation of the object.

4. A life predicting device according to any one of claim 1 to 3, wherein,

the lifetime consumption estimating unit corrects the lifetime consumption estimation result using a degradation coefficient calculated based on the basic lifetime and the consumed lifetime after estimating the lifetime consumption, and outputs the corrected estimation result.

5. The life predicting apparatus according to claim 4, wherein,

the degradation coefficient is a result of dividing a value obtained by subtracting the consumed lifetime from the basic lifetime by the basic lifetime.

6. The life predicting device according to any one of claims 1 to 5, wherein,

the device state information is generated based on information output from an operation estimating unit that performs simulation of the operation of the object in accordance with the specified operation condition.

7. The life predicting device according to any one of claims 1 to 6, wherein,

the device state information is output so as to be notified to the outside.

8. The life predicting device according to any one of claims 1 to 7, wherein,

the device state information is changed so that the device stress estimated by the device stress estimating unit is reduced when the predicted device lifetime, which is a result of lifetime prediction by the lifetime predicting unit, is shorter than a predetermined target device lifetime.

9. The life predicting apparatus according to claim 8, wherein,

when the operation condition changing unit changes the device state information,

the operation condition changing unit outputs, to the outside, a result of predicting the life of the device based on the change content of the device state information and the result of predicting the change of the object when the object is operated under the condition corresponding to the changed device state information.

10. The life predicting apparatus according to claim 8, wherein,

When the operation condition changing unit changes the device state information, the operation condition changing unit outputs the changed device state information to the object, and operates the object under the condition indicated by the changed device state information.

11. The life predicting device according to any one of claims 1 to 10, wherein,

comprises a life consumption estimating unit for obtaining the life consumption of the object by offline estimation,

the lifetime predicting unit predicts the lifetime of the predicting device by using the lifetime consumption calculated by the lifetime consumption estimating unit during the operation of the object,

the lifetime predicting unit predicts the lifetime of the predicting device using the lifetime consumption calculated by the lifetime consumption estimating unit during the object stop.

12. The life predicting device according to any one of claims 1 to 11, wherein,

the device state information includes at least one of an ambient temperature of the motor control device, a power supply voltage of the motor control device, a power supply frequency of the motor control device, a bus voltage of the motor control device, a bus current of the motor control device, a motor current of the motor controlled by the motor control device, a carrier frequency used by the motor control device when controlling the motor, a cooling state of the motor control device, a rotational speed of the motor controlled by the motor control device, a movement speed of a machine connected to the motor controlled by the motor control device, a friction amount of the machine connected to the motor controlled by the motor control device, a load inertia of the machine connected to the motor controlled by the motor control device, an axial load of the machine connected to the motor controlled by the motor control device, a vibration of the machine connected to the motor controlled by the motor control device, an altitude of a place where the motor control device is installed, a humidity of the place where the motor control device is installed, an operation time of the motor control device, and a machining condition using the motor control device.

13. The life predicting device according to any one of claims 1 to 12, wherein,

the device stress is set to a temperature of a capacitor connected to or incorporated in the motor control device.

14. The life predicting device according to any one of claims 1 to 12, wherein,

the device stress is set to a temperature of a switching element incorporated in the motor control device.

15. The life predicting device according to any one of claims 1 to 12, wherein,

the device stress is set to the temperature of the coil of the motor controlled by the motor control device.

16. The life predicting device according to any one of claims 1 to 12, wherein,

the device stress is set as an axial load of a machine connected to a motor controlled by a motor control device.

17. A life prediction system, comprising:

the lifetime prediction device according to any one of claims 1 to 16; and

and a motor system including the object.

18. A learning device, comprising:

a data acquisition unit that acquires device state information, which is information on a state of use of a device or a component constituting an industrial machine, that is, an object, and device stress used for predicting the lifetime of the object; and

And a model generation unit that generates a trained model for estimating the device stress from the device state information, using learning data created based on the device state information and the device stress.

19. An estimation device is characterized by comprising:

a data acquisition unit that acquires device state information, which is information on the use state of an object that is a device or a component constituting an industrial machine; and

an estimating unit that estimates the device stress from the device state information acquired by the data acquiring unit, using a trained model for estimating the device stress used for lifetime prediction of the object from the device state information.

20. A learning device, comprising:

a data acquisition unit that acquires device stress used for predicting the life of an object, which is a device or a component constituting an industrial machine, and a life consumption amount, which is the life of the object consumed in a specified cycle; and

and a model generation unit that generates a trained model for estimating the lifetime consumption amount from the device stress, using learning data created based on the device stress and the lifetime consumption amount.

21. The learning apparatus of claim 20, wherein the learning device,

the data acquisition unit acquires a degradation coefficient calculated based on a basic life of the object and a life consumed from the basic life, that is, a consumed life, in addition to the device stress and the life consumption,

the model generation unit generates a trained model for estimating the lifetime consumption amount from the device stress and the degradation coefficient.

22. An estimation device is characterized by comprising:

a data acquisition unit that acquires device stress of an object, which is a device or a component constituting an industrial machine; and

an estimating unit that estimates the lifetime consumption amount from the device stress acquired by the data acquiring unit, using a trained model for estimating lifetime consumption amount, which is lifetime of the object consumed for each cycle determined from the device stress.

23. The inference apparatus of claim 22, wherein,

the data acquisition unit acquires a degradation coefficient calculated based on a basic life of the object and a life consumed from the basic life, that is, a consumed life, in addition to the device stress,

The estimating unit estimates the lifetime consumption amount from the device stress and the degradation coefficient acquired by the data acquiring unit, using the trained model for estimating the lifetime consumption amount from the device stress and the degradation coefficient.

24. A lifetime prediction program, characterized by causing a computer to execute:

step 1, holding a life consumption, which is a life consumed from a basic life of an object, which is a device or a component constituting an industrial machine, as life data;

step 2, estimating device stress based on device state information, which is information on the use state of the object;

step 3, estimating a lifetime consumption amount, which is a lifetime of the object consumed for each specified period, based on the device stress; and

and 4. Predicting the lifetime of the object based on the lifetime data and the lifetime consumption held in the 1 st step.