Management system for work machine

The work machine management system uses machine learning to improve the accuracy of oil state estimation in working machines, facilitating timely maintenance decisions.

JP2024161797APending Publication Date: 2024-11-20SUMITOMO HEAVY IND LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2023076826
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-08
Publication Date
2024-11-20

AI Technical Summary

Technical Problem

Conventional methods for estimating the deterioration state of oil in working machines based on sensor detection are inaccurate.

Method used

A work machine management system that includes a management device storing operating information and using machine learning to estimate the state of hydraulic oil by inputting operation information into a model.

Benefits of technology

Enables accurate estimation of oil condition, allowing timely oil replacement decisions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024161797000001_ABST
    Figure 2024161797000001_ABST
Patent Text Reader

Abstract

To estimate oil condition with high accuracy.SOLUTION: A management system includes a work machine and a management device for the work machine. The management device has a model that learns relationship between operation information of the work machine and a condition of hydraulic oil of the work machine, an operation information acquisition unit that acquires the operation information from the work machine, and a state estimation unit that inputs the operation information acquired by the operation information acquisition unit to the model and acquires estimation result information obtained by estimating the condition of the hydraulic oil.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to a management system for a work machine. [Background technology]

[0002] 2. Description of the Related Art Conventionally, there is known a work machine that detects oil property values ​​such as temperature, kinetic viscosity, density, and dielectric constant using sensors and estimates the deterioration state of the oil based on the detected oil property values. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 7009408 Summary of the Invention [Problem to be solved by the invention]

[0004] In regular maintenance of a work machine, a serviceman takes an oil sample from the work machine and has the sample analyzed by an analytical laboratory, etc., to determine the deterioration state of the oil. In contrast, in the conventional technology described above, the deterioration state of the oil is estimated only based on the detection results by the sensor, and therefore the accuracy is insufficient.

[0005] The present disclosure aims to estimate the condition of oil with high accuracy. [Means for solving the problem]

[0006] A work machine management system according to an embodiment of the present invention is a management system including a work machine and a management device for the work machine, the management device having a model that has learned the relationship between operation information of the work machine and a state of hydraulic oil of the work machine, an operation information acquisition unit that acquires the operation information from the work machine, and a state estimation unit that inputs the operation information acquired by the operation information acquisition unit into the model and acquires estimation result information that estimates the state of the hydraulic oil. Effect of the Invention

[0007] The oil condition can be estimated with high accuracy. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram illustrating an example of a system configuration of a management system for a work machine. [Diagram 2] 2 is a diagram illustrating the hardware configuration of each device in the work machine management system. FIG. [Diagram 3] FIG. 2 is a diagram illustrating functions of a management device. [Figure 4] 11 is a first flowchart illustrating a process of a management device. [Diagram 5] 13 is a second flowchart illustrating the process of the management device. [Figure 6] FIG. 13 is a diagram illustrating extraction of learning data. [Figure 7] FIG. 11 is a diagram illustrating an example of estimation result information. [Figure 8] FIG. 13 is a diagram showing a display example of a display device of a shovel. [Figure 9] FIG. 13 is a diagram showing an example of a display of an estimation result of an oil state. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] The construction machine management system of this embodiment will be described below with reference to the drawings. Fig. 1 is a diagram showing an example of the system configuration of the construction machine management system. In this embodiment, a shovel 100 will be described as an example of a construction machine.

[0010] The shovel management system SYS of the present embodiment includes a shovel 100 and a management device 200. In the following description, the management system SYS of the shovel 100 will be simply referred to as the management system SYS.

[0011] In the management system SYS of this embodiment, the shovel 100 and the management device 200 are connected via a network or the like.

[0012] First, a configuration of a shovel 100 according to the present embodiment will be described. In Fig. 1, a side view of the shovel 100 is shown.

[0013] The excavator 100 has a lower traveling body 1, a slewing mechanism 2, and an upper rotating body 3. In the excavator 100, the upper rotating body 3 is rotatably mounted on the lower traveling body 1 via the slewing mechanism 2. A boom 4 is attached to the upper rotating body 3. An arm 5 is attached to the tip of the boom 4, and a bucket 6 is attached to the tip of the arm 5 as an end attachment.

[0014] The boom 4, arm 5, and bucket 6 constitute an excavation attachment, which is an example of an attachment. The boom 4 is driven by a boom cylinder 7, the arm 5 is driven by an arm cylinder 8, and the bucket 6 is driven by a bucket cylinder 9. A boom angle sensor S1 is attached to the boom 4, an arm angle sensor S2 is attached to the arm 5, and a bucket angle sensor S3 is attached to the bucket 6.

[0015] The boom angle sensor S1 is configured to detect the rotation angle of the boom 4. In this embodiment, the boom angle sensor S1 is an acceleration sensor, and can detect the rotation angle of the boom 4 relative to the upper rotating body 3 (hereinafter referred to as the "boom angle"). For example, the boom angle is at its minimum when the boom 4 is lowered to the lowest and increases as the boom 4 is raised.

[0016] The arm angle sensor S2 is configured to detect the rotation angle of the arm 5. In this embodiment, the arm angle sensor S2 is an acceleration sensor, and can detect the rotation angle of the arm 5 relative to the boom 4 (hereinafter referred to as the "arm angle"). For example, the arm angle is at its minimum when the arm 5 is fully closed, and increases as the arm 5 is opened.

[0017] The bucket angle sensor S3 is configured to detect the rotation angle of the bucket 6. In this embodiment, the bucket angle sensor S3 is an acceleration sensor, and can detect the rotation angle of the bucket 6 relative to the arm 5 (hereinafter referred to as the "bucket angle"). For example, the bucket angle is at its minimum when the bucket 6 is fully closed, and increases as the bucket 6 is opened.

[0018] The boom angle sensor S1, the arm angle sensor S2, and the bucket angle sensor S3 may each be a potentiometer using a variable resistor, a stroke sensor that detects the stroke amount of the corresponding hydraulic cylinder, a rotary encoder that detects the rotation angle around the connecting pin, a gyro sensor, or a combination of an acceleration sensor and a gyro sensor, etc.

[0019] A boom rod pressure sensor S7R and a boom bottom pressure sensor S7B are attached to the boom cylinder 7. An arm rod pressure sensor S8R and an arm bottom pressure sensor S8B are attached to the arm cylinder 8.

[0020] A bucket rod pressure sensor S9R and a bucket bottom pressure sensor S9B are attached to the bucket cylinder 9. The boom rod pressure sensor S7R, the boom bottom pressure sensor S7B, the arm rod pressure sensor S8R, the arm bottom pressure sensor S8B, the bucket rod pressure sensor S9R and the bucket bottom pressure sensor S9B are also collectively referred to as the "cylinder pressure sensors."

[0021] The boom rod pressure sensor S7R detects the pressure in the rod-side oil chamber of the boom cylinder 7 (hereinafter referred to as the "boom rod pressure"), and the boom bottom pressure sensor S7B detects the pressure in the bottom-side oil chamber of the boom cylinder 7 (hereinafter referred to as the "boom bottom pressure"). The arm rod pressure sensor S8R detects the pressure in the rod-side oil chamber of the arm cylinder 8 (hereinafter referred to as the "arm rod pressure"), and the arm bottom pressure sensor S8B detects the pressure in the bottom-side oil chamber of the arm cylinder 8 (hereinafter referred to as the "arm bottom pressure").

[0022] The bucket rod pressure sensor S9R detects the pressure in the rod side oil chamber of the bucket cylinder 9 (hereinafter referred to as the "bucket rod pressure"), and the bucket bottom pressure sensor S9B detects the pressure in the bottom side oil chamber of the bucket cylinder 9 (hereinafter referred to as the "bucket bottom pressure").

[0023] The upper rotating body 3 is provided with a cabin 10 serving as a driver's cab, and is equipped with a power source such as an engine 11. A sensor for detecting CO2 emissions may be provided near the exhaust mechanism of the engine 11. The upper rotating body 3 is also provided with a hydraulic oil tank 49 that stores oil (hydraulic oil) used as a lubricant and power transmission medium for key components of the engine system and hydraulic equipment system of the excavator 100.

[0024] Furthermore, on the upper rotating body 3, a controller 30, a display device 40, an input device 42, an audio output device 43, a storage device 47, a positioning device P1, an aircraft tilt sensor S4, a rotation angular velocity sensor S5, an imaging device S6 and a communication terminal T are attached.

[0025] The upper rotating body 3 may be equipped with a power storage unit that supplies electric power, and a motor generator that generates electric power using the rotational driving force of the engine 11. The power storage unit is, for example, a capacitor or a lithium ion battery. The motor generator may function as an electric motor to drive a mechanical load, or may function as a generator to supply electric power to an electric load.

[0026] The controller 30 functions as a main control unit that controls the drive of the shovel 100. In this embodiment, the controller 30 is configured by a computer including a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and the like. The various functions of the controller 30 are realized, for example, by the CPU executing a program stored in the ROM. The various functions may include, for example, at least one of a machine guidance function that guides (guides) the manual operation of the shovel 100 by the operator, and a machine control function that automatically assists the manual operation of the shovel 100 by the operator.

[0027] The display device 40 is configured to display various information. The display device 40 may be connected to the controller 30 via a communication network such as a CAN, or may be connected to the controller 30 via a dedicated line.

[0028] The input device 42 is configured to enable an operator to input various information to the controller 30. The input device 42 includes at least one of a touch panel, a knob switch, a membrane switch, and the like, which are installed in the cabin 10.

[0029] The audio output device 43 is configured to output audio. The audio output device 43 may be, for example, an in-vehicle speaker connected to the controller 30, or an alarm such as a buzzer. In this embodiment, the audio output device 43 is configured to output various information by audio in response to an audio output command from the controller 30.

[0030] The storage device 47 is configured to store various information. The storage device 47 is, for example, a non-volatile storage medium such as a semiconductor memory. The storage device 47 may store information output by various devices during operation of the shovel 100, or may store information acquired via various devices before the operation of the shovel 100 is started.

[0031] The storage device 47 may also store operation information of the shovel 100. The operation information of the shovel 100 includes values ​​detected by various sensors of the shovel 100. Specifically, the operation information may include the pilot pressure, engine speed, engine load factor, and the like of the shovel 100. The operation information may also include position information indicating the current position of the shovel, orientation information indicating the orientation of the shovel, attitude information indicating the attitude of the shovel, work content information indicating the work content, load rate information indicating the load rate, accumulated time information indicating the accumulated operation time, fuel information including the fuel injection amount, CO2 emission amount, work amount, operating state, and the like. The operation information may be periodically transmitted to the management device 200 by the communication terminal T.

[0032] The positioning device P1 is configured to measure the position of the upper rotating body 3. The positioning device P1 may be configured to measure the orientation of the upper rotating body 3. In this embodiment, the positioning device P1 is, for example, a GNSS compass, detects the position and orientation of the upper rotating body 3, and outputs the detected value to the controller 30. Therefore, the positioning device P1 can also function as an orientation detection device that detects the orientation of the upper rotating body 3. The orientation detection device may be a direction sensor attached to the upper rotating body 3.

[0033] The machine body inclination sensor S4 is configured to detect the inclination of the upper rotating body 3. In this embodiment, the machine body inclination sensor S4 is an acceleration sensor that detects the longitudinal inclination angle about the longitudinal axis and the lateral inclination angle about the lateral axis of the upper rotating body 3 with respect to a virtual horizontal plane. The longitudinal axis and the lateral axis of the upper rotating body 3 are perpendicular to each other at, for example, the shovel center point, which is a point on the rotation axis of the shovel 100.

[0034] The rotation angular velocity sensor S5 is configured to detect the rotation angular velocity of the upper rotating body 3. The rotation angular velocity sensor S5 may be configured to detect or calculate the rotation angle of the upper rotating body 3. In this embodiment, the rotation angular velocity sensor S5 is a gyro sensor. The rotation angular velocity sensor S5 may be a resolver, a rotary encoder, or the like.

[0035] The imaging device S6 is an example of a spatial recognition device, and is configured to acquire an image of the periphery of the shovel 100. In this embodiment, the imaging device S6 includes a front camera S6F that images the space in front of the shovel 100, a left camera S6L that images the space to the left of the shovel 100, a right camera S6R that images the space to the right of the shovel 100, and a rear camera S6B that images the space behind the shovel 100.

[0036] The imaging device S6 is, for example, a monocular camera having an imaging element such as a CCD or CMOS, and outputs a captured image to the display device 40. The imaging device S6 may be a stereo camera, a distance image camera, or the like. The imaging device S6 may be replaced with other spatial recognition devices such as a three-dimensional distance image sensor, an ultrasonic sensor, a millimeter wave radar, a LIDAR, or an infrared sensor, or may be replaced with a combination of another spatial recognition device and a camera.

[0037] The front camera S6F is attached, for example, to the ceiling of the cabin 10, i.e., inside the cabin 10. However, the front camera S6F may be attached to the outside of the cabin 10, such as the roof of the cabin 10 or the side of the boom 4. The left camera S6L is attached to the left end of the upper surface of the upper rotating body 3, the right camera S6R is attached to the right end of the upper surface of the upper rotating body 3, and the rear camera S6B is attached to the rear end of the upper surface of the upper rotating body 3.

[0038] The communication terminal T is configured to control communication with an external device outside the excavator 100. In this embodiment, the communication terminal T controls communication with the external device via a satellite communication network, a mobile phone communication network, an Internet network, or the like. The external device is, for example, a management device 200 such as a server installed in an external facility.

[0039] The management device 200 of this embodiment estimates the oil condition by using operation information transmitted from the shovel 100 and the analysis results of oil collected during regular maintenance of the shovel 100. Furthermore, the management device 200 of this embodiment determines whether or not the oil needs to be changed according to the estimation result of the oil condition.

[0040] The process of estimating the state of oil by the management device 200 of this embodiment will be outlined below.

[0041] Periodic maintenance is performed on the excavator 100. In this maintenance, a serviceman or the like samples oil from the hydraulic oil tank 49, and the sampled oil is analyzed. The analysis of the oil is performed by an analysis agency or the like.

[0042] The management device 200 of this embodiment acquires analysis result information indicating the analysis result of the oil from an analysis institute, etc. Then, the management device 200 learns the relationship between the oil analysis result information and the operation information acquired from the shovel 100 by machine learning, and generates an estimation model for estimating the state of the oil.

[0043] In addition, when the management device 200 of this embodiment generates an estimation model, it inputs operation information into the estimation model at any timing and obtains estimation result information that estimates the oil state at the time when the shovel 100 acquires the operation information.

[0044] Furthermore, the management device 200 of this embodiment determines whether or not an oil change is necessary from the estimation result information, and if it is determined that an oil change is necessary, notifies the manager of the shovel 100, etc., to that effect. The manager of the shovel 100 may be, for example, a serviceman who performs maintenance on the shovel 100, or a user who performs work using the shovel 100.

[0045] In this embodiment, an estimation model is generated using analysis result information of the oil collected from the shovel 100, and the state of the oil is estimated using the estimation model, thereby improving the accuracy of the estimation.

[0046] Here, the oil analysis result information will be described. The oil analysis result information in this embodiment includes information indicating the date and time when the oil was collected from the shovel 100, and information indicating the state of the oil at the time of collection. In the following description, the information indicating the date and time when the oil was collected from the shovel 100 may be expressed as collection date and time information. Furthermore, the information indicating the state of the oil may include values ​​of the oil properties.

[0047] The information indicating the oil condition may include, for example, the acid value, kinetic viscosity, and the weight of each metal element contained in the oil. The acid value and kinetic viscosity increase with deterioration of the oil and indicate the deterioration state of the oil. The weight of each metal element indicates the wear state of the excavator 100.

[0048] The information indicating the oil condition may include items other than those described above. The items included in the information indicating the oil condition may vary depending on the analysis institute that analyzes the collected oil.

[0049] 1, the management device 200 is realized by one information processing device, but is not limited to this. The management device 200 may be realized by multiple information processing devices. In other words, the functions realized by the management device 200 may be realized by multiple information processing devices.

[0050] Next, the hardware configuration of each device in the management system SYS will be described with reference to Fig. 2. Fig. 2 is a diagram illustrating the hardware configuration of each device in the management system for a work machine.

[0051] First, a description will be given of the hardware configuration of the excavator 100. In the cabin 10, a controller 30, a display device 40, a communication terminal T, etc. are installed.

[0052] The engine 11 is a diesel engine that employs isochronous control for maintaining a constant engine speed regardless of an increase or decrease in the engine load. The amount of fuel injected into the engine 11, the timing of fuel injection, and other parameters are controlled by an engine control unit 50.

[0053] The rotating shaft of the engine 11 is connected to the respective rotating shafts of a main pump 14 and a pilot pump 15 which serve as hydraulic pumps.

[0054] The regulator 13 is configured to be able to control the displacement volume of the main pump 14. In this embodiment, the regulator 13 controls the displacement volume of the main pump 14 by adjusting the swash plate tilt angle of the main pump 14 in response to a control command from the controller 30. For example, the controller 30 changes the displacement volume of the main pump 14 by outputting a control command to the regulator 13 in response to the output of the discharge pressure sensor 28, etc. The regulator 13 is configured to be able to transmit data representing the swash plate tilt angle to the controller 30.

[0055] The main pump 14 is configured to supply hydraulic oil via a hydraulic oil line to the control valve 17. In this embodiment, the main pump 14 is a swash plate type variable displacement hydraulic pump.

[0056] The oil temperature sensor 14c is configured to detect the temperature of the hydraulic oil flowing through the hydraulic oil line between the hydraulic oil tank 49 and the main pump 14. The oil temperature sensor 14c is configured to transmit the detected data to the controller 30.

[0057] The pilot pump 15 is configured to be able to supply hydraulic oil to various hydraulic control devices such as an operating device through a pilot line. In this embodiment, the pilot pump 15 is a fixed displacement hydraulic pump. However, the pilot pump 15 may be omitted. In this case, the function of the pilot pump 15 may be realized by the main pump 14. In this case, the main pump 14 may have a function of supplying hydraulic oil to an operating device or the like after reducing the pressure of the hydraulic oil by a throttle or the like, in addition to the function of supplying hydraulic oil to the control valve 17.

[0058] The control valve 17 is a hydraulic control device that controls the hydraulic system of the excavator 100. The control valve 17 is connected to hydraulic actuators such as the right-side traveling hydraulic motor, the left-side traveling hydraulic motor, the boom cylinder 7, the arm cylinder 8, the bucket cylinder 9, and the swing hydraulic motor.

[0059] The controller 30 is a control device that controls the shovel 100. In this embodiment, the controller 30 is configured with a microcomputer including a CPU, a volatile storage device, a nonvolatile storage device, etc. Various functions of the controller 30 are realized by the CPU executing programs stored in the nonvolatile storage device.

[0060] The display device 40 is attached between the right pillar and the driver's seat, and is arranged so that the operator seated in the driver's seat can see the screen. The display device 40 is configured to be able to display various information in response to commands from the controller 30. In this embodiment, the display device 40 is a liquid crystal display connected to the controller 30.

[0061] An image is displayed on the image display unit 41. The switch panel 45 is a panel including various hardware switches. The switch panel 45 may be a touch panel disposed on the image display unit 41.

[0062] In this embodiment, the display device 40 is configured to operate by receiving power from a storage battery 70. The storage battery 70 is charged with power generated by an alternator 11a (generator) of the engine 11. The power of the storage battery 70 is also supplied to the controller 30, the communication terminal T, etc. For example, the starter 11b of the engine 11 is configured to be driven by power from the storage battery 70 to start the engine 11.

[0063] The communication terminal T is configured to be able to control communication between the shovel 100 and an external device. In this embodiment, the communication terminal T controls wireless communication between the shovel 100 and at least one of the management devices 200 via at least one of a satellite communication line, a mobile phone communication line, a short-range wireless communication line, and the like.

[0064] The engine control unit 50 is configured to be able to control the engine 11. In addition, the engine control unit 50 is configured to be able to transmit data representing the state of the engine 11, such as the coolant temperature, to the controller 30.

[0065] The engine speed adjustment dial 75 is a dial for adjusting the speed of the engine 11. The engine speed adjustment dial 75 transmits data indicating the setting state of the engine speed to the controller 30. The engine speed adjustment dial 75 is configured to be able to switch the engine speed in four stages: SP mode, H mode, A mode, and idling state. The SP mode is a speed mode selected when a priority is given to the amount of work, and uses the highest engine speed. The H mode is a speed mode selected when a priority is given to both the amount of work and fuel efficiency, and uses the second highest engine speed. The A mode is a speed mode selected when a priority is given to fuel efficiency while operating the excavator 100 with low noise, and uses the third highest engine speed. The idling state is a speed mode selected when the engine 11 is to be in an idling state, and uses the lowest engine speed. The engine 11 is controlled to rotate at a constant engine speed corresponding to the speed mode set by the engine speed adjustment dial 75.

[0066] Next, a description will be given of the hardware configuration of the management device 200 of this embodiment. The management device 200 of this embodiment is a computer having a CPU 201, a storage device 202, a communication device 203, an input device 204, and a display device 205, which are all connected to each other via a bus.

[0067] The CPU 201 controls the overall operation of the management device 200. The storage device 202 stores programs executed by the CPU 201, various information related to the shovel 100, and the like. The communication device 203 communicates with the shovel 100 via a network.

[0068] The input device 204 is for inputting information to the management device 200, and is realized by, for example, a keyboard, a pointing device, etc. The display device 205 is for displaying various information output from the management device 200, and is realized by a display, etc.

[0069] Next, functions of the management device 200 of this embodiment will be described with reference to Fig. 3. Fig. 3 is a diagram for explaining the functions of the management device.

[0070] The management device 200 of this embodiment includes an analysis result acquisition unit 210 , an operation information acquisition unit 220 , a learning data extraction unit 230 , a learning unit 240 , a state estimation unit 250 , an output unit 260 , and a memory unit 270 .

[0071] The analysis result acquisition unit 210, the operation information acquisition unit 220, the learning data extraction unit 230, the learning unit 240, the state estimation unit 250, and the output unit 260 of the management device 200 may be realized by the CPU 201 reading and executing programs stored in the storage device 202 or the like. The storage unit 270 may be realized by the storage device 202 or the like.

[0072] The analysis result acquisition unit 210 acquires oil analysis result information 271 from an analysis institute or the like, and stores it in the storage unit 270. The analysis result information 271 may be attached to an email or the like sent from the analysis institute or the like, for example.

[0073] The operation information acquiring unit 220 acquires operation information 272 collected by the shovel 100 and stores the operation information 272 in the storage unit 270. The operation information acquiring unit 220 may receive the operation information 272 from the shovel 100 at a preset time during the day, for example.

[0074] Furthermore, in this embodiment, when the analysis result acquisition unit 210 acquires the analysis result information 271, the operation information 272 acquired at the date and time closest to the date and time indicated by the collection date and time information included in the analysis result information 271 may be stored in the storage unit 270 in association with the analysis result information 271. In other words, the analysis result information 271 may be stored in the storage unit 270 in association with the operation information 272 corresponding to the analysis result information 271.

[0075] The learning data extraction unit 230 specifies a range of learning data to be used for machine learning by the learning unit 240 from the information stored in the storage unit 270, and extracts learning data based on the specified range. Details of the processing by the learning data extraction unit 230 will be described later.

[0076] The learning unit 240 generates an estimation model 273 that estimates the state of oil, and stores it in the storage unit 270. The estimation model 273 is a trained model that is obtained by machine learning the training data extracted by the training data extraction unit 230.

[0077] Furthermore, when the analysis result acquisition unit 210 acquires new analysis result information 271, the learning unit 240 updates the estimation model 273. In other words, the learning unit 240 updates the estimation model 273 every time new analysis result information 271 is added.

[0078] The state estimation unit 250 estimates the state of the oil in the shovel 100 using the estimation model 273, and stores estimation result information 274 including the estimation result in the storage unit 270. Specifically, the state estimation unit 250 inputs the operation information 272 received from the shovel 100 to the estimation model 273, and acquires a predicted value of the oil properties output from the estimation model 273. The predicted value of the oil properties is, in other words, a result of estimating the oil state, and is a predicted value indicating the oil state.

[0079] Then, the state estimation unit 250 associates the predicted value of the oil property with the operation information 272 input to the estimation model 273 , and stores the result as estimation result information 274 in the storage unit 270 .

[0080] Furthermore, the state estimation unit 250 of this embodiment judges whether or not the oil in the shovel 100 needs to be changed, depending on the result of estimating the oil state, and when it is judged that the oil needs to be changed, notifies the manager or the like of the shovel 100 of that effect. Details of the notification when it is judged that the oil change is necessary will be described later.

[0081] The output unit 260 outputs various information. Specifically, the output unit 260 may output the estimation result information 274 and a predicted value of the oil property. Furthermore, the output unit 260 may output the determination result of the necessity of an oil change to a terminal device or the like carried by the manager of the shovel 100. Furthermore, the output unit 260 may output information to be displayed on the display device 40 of the shovel 100.

[0082] The storage unit 270 stores analysis result information 271, operation information 272, an estimation model 273, and estimation result information 274.

[0083] In the storage unit 270, the analysis result information 271 may be stored as a data set in which the analysis result information 271 is associated with the corresponding operation information 272.

[0084] The estimation model 273 is mainly configured with, for example, a neural network. In this embodiment, the neural network may be a so-called deep neural network having one or more intermediate layers (hidden layers) between an input layer and an output layer. When the operation information 272 is input, the estimation model 273 of this embodiment outputs a predicted value of the oil properties as a result of estimating the oil state.

[0085] The estimation result information 274 is information in which the operation information 272 input to the estimation model 273 is associated with the predicted value of the oil property output from the estimation model 273.

[0086] Next, a description will be given of the operation of the management device 200. Fig. 4 is a first flowchart illustrating the processing of the management device. Fig. 4 shows the processing of generating or updating the estimation model 273.

[0087] The management device 200 of this embodiment determines whether the analysis result information 271 has been acquired by the analysis result acquisition unit 210 (step S401).

[0088] In step S401, if the analysis result information 271 is acquired, the process proceeds to step S403 described below. In step S401, if the analysis result information 271 is not acquired, the management device 200 determines whether or not the operation information 272 is acquired by the operation information acquisition unit 220 (step S402). In step S402, if the operation information 272 is not acquired, the management device 200 returns to step S401.

[0089] When the analysis result information 271 is acquired in step S401, or when the operation information 272 is acquired in step S402, the management device 200 specifies a range from which learning data is to be extracted by the learning data extraction unit 230 (step S403). The range from which learning data is to be extracted in this embodiment may indicate a period during which information that becomes the learning data is acquired, or may indicate the amount of information that becomes the learning data.

[0090] Next, the management device 200 extracts information within the range specified in step S403 as learning data from the analysis result information 271 and the operation information 272 stored in the storage unit 270 (step S404). The details of the processing by the learning data extraction unit 230 in steps S403 and S404 will be described later.

[0091] Next, the management device 200 causes the learning unit 240 to perform machine learning using the learning data extracted in step S404, and generates or updates the estimation model 273, which is learned data (step S405). In this embodiment, if the estimation model 273 has already been stored in the storage unit 270, the learning unit 240 updates the estimation model 273 stored in the storage unit 270.

[0092] Fig. 5 is a second flowchart for explaining the processing of the management device. Fig. 5 shows a process of estimating the state of oil using the estimation model 273. Note that the management device 200 may execute the process shown in Fig. 4 and the process shown in Fig. 5 as a series of processes.

[0093] The management device 200 determines whether or not the operation information 272 has been acquired from the excavator 100 by the operation information acquisition unit 220 (step S501). If the operation information 272 has not been acquired in step S501, the management device 200 waits.

[0094] In step S501, when the operation information 272 is acquired, the management device 200 inputs the operation information 272 to the estimation model 273 via the state estimation unit 250 (step S502).

[0095] Next, the state estimation unit 250 acquires a predicted value of the oil properties output from the estimation model 273 (step S503), and stores in the memory unit 270 the estimation result information 274 associated with the operation information 272 input to the estimation model 273 in step S502 (step S504).

[0096] Next, the process of the learning data extraction unit 230 of this embodiment will be described with reference to Fig. 6. Fig. 6 is a diagram for explaining extraction of learning data. Fig. 6 describes the details of the process of the learning data extraction unit 230 in steps S403 and S404 in Fig. 4.

[0097] 6(A), (B), and (C) are diagrams for explaining the processing of the learning data extraction unit 230 when the operation information 272 is acquired.

[0098] 6(A), (B), and (C), sensor A and sensor B are examples of values ​​included in operation information 272, and the state of the oil may be the value of the oil property indicated by the analysis result information or the predicted value of the oil property output from estimation model 273. Figures 6(A), (B), and (C) show the changes over time of sensor A, sensor B, and the state of the oil.

[0099] In this embodiment, when updating the estimation model 273 in response to the input of the operation information 272, the range for extracting learning data may be a data set of the analysis result information 271 acquired during a certain period in the past starting from the date and time when the operation information 272 was acquired, and the operation information 272 associated with the analysis result information 271. In this case, the learning data extraction unit 230 extracts, from the storage unit 270, the analysis result information 271 acquired during this certain period and the operation information 272 corresponding to the analysis result information 271, as learning data.

[0100] The operation information 272 corresponding to the analysis result information 271 is the operation information 272 acquired at the timing closest to the date and time indicated by the collection date and time information included in the analysis result information 271.

[0101] FIG. 6A shows a case where the management device 200 generates an estimation model 273 when it acquires operation information 272x, and estimates the state of oil using the generated estimation model 273.

[0102] In this case, the date and time when the operation information 272x was acquired is timing T1. Therefore, the learning data extraction unit 230 specifies a certain period K in the past starting from timing T1. In other words, in FIG. 6A, the period K starting from timing T1 and going back to timing T2 is set as the range for extracting learning data.

[0103] Then, the learning data extraction unit 230 extracts a data set of the analysis result information 271 acquired during a certain period K and the operation information 272 corresponding to the analysis result information 271 as learning data.

[0104] Analysis result information 271 acquired during a period K from timing T1 to timing T2 includes analysis result information 271a and 271b.

[0105] 6A, the operation information 272a is the operation information acquired at the timing closest to that of the analysis result information 271a. Therefore, the operation information 272a is the operation information corresponding to the analysis result information 271a.

[0106] 6A, the operation information 272b is the operation information acquired at the timing closest to that of the analysis result information 271b. Therefore, the operation information 272b is the operation information corresponding to the analysis result information 271b.

[0107] Therefore, the learning data extraction unit 230 extracts learning data including a data set in which the analysis result information 271a and the operation information 272a are associated with each other, and a data set in which the analysis result information 271b and the operation information 272b are associated with each other. Note that the period K may be set in advance by an administrator of the management device 200, etc.

[0108] The management device 200 generates an estimation model 273 using the learning data extracted in this manner, and inputs the operation information 272x acquired at timing T1 to the generated estimation model 273. Then, the management device 200 acquires a predicted value 275x of the oil property output from the estimation model 273.

[0109] Next, Fig. 6(B) will be described. In the example of Fig. 6(B), when the management device 200 acquires the operation information 272y, the estimation model 273 is updated, and the state of the oil is estimated using the updated estimation model 273.

[0110] In this case, the date and time when the operation information 272y was acquired is timing T5. Therefore, the learning data extraction unit 230 specifies a certain period K in the past starting from timing T5. In other words, in FIG. 6B, the period K starting from timing T5 and going back to timing T6 is set as the range for extracting learning data.

[0111] Then, the learning data extraction unit 230 extracts a data set of the analysis result information 271 acquired during a certain period K and the operation information 272 corresponding to the analysis result information 271 as learning data.

[0112] The analysis result information 271 acquired during the period K from timing T5 to timing T6 includes analysis result information 271c, 271c.

[0113] 6(B), the operation information 272c is the operation information acquired at the timing closest to that of the analysis result information 271c. Therefore, the operation information 272c is the operation information corresponding to the analysis result information 271c. Moreover, the operation information 272d is the operation information acquired at the timing closest to that of the analysis result information 271d. Therefore, the operation information 272d is the operation information corresponding to the analysis result information 271d.

[0114] Therefore, the learning data extraction unit 230 extracts learning data including a data set in which the analysis result information 271c and the operation information 272c are associated with each other, and a data set in which the analysis result information 271d and the operation information 272d are associated with each other.

[0115] In this embodiment, the estimation model 273 is updated using the learning data extracted in this manner, and the operation information 272y acquired at timing T5 is input to the updated estimation model 273. Then, the management device 200 acquires a predicted value 275y of the oil property output from the updated estimation model 273.

[0116] Next, Fig. 6(C) will be described. In the example of Fig. 6(C), when the management device 200 acquires the operation information 272z, the estimation model 273 is updated, and the oil state is estimated using the updated estimation model 273.

[0117] In this case, the date and time when the operation information 272z was acquired is timing T8. Therefore, the learning data extraction unit 230 specifies a certain period K in the past starting from timing T8. In other words, in FIG. 6C, the period K starting from timing T8 and going back to timing T9 is set as the range for extracting learning data.

[0118] Then, the learning data extraction unit 230 extracts a data set of the analysis result information 271 acquired during a certain period K and the operation information 272 corresponding to the analysis result information 271 as learning data.

[0119] The analysis result information 271 acquired during the period K from timing T8 to timing T9 includes analysis result information 271e and 271f.

[0120] 6(C), the operation information 272e is the operation information acquired at the timing closest to that of the analysis result information 271e. Therefore, the operation information 272e is the operation information corresponding to the analysis result information 271e. Moreover, the operation information 272f is the operation information acquired at the timing closest to that of the analysis result information 271f. Therefore, the operation information 272f is the operation information corresponding to the analysis result information 271f.

[0121] Therefore, the learning data extraction unit 230 extracts learning data including a data set in which the analysis result information 271e and the operation information 272e are associated with each other, and a data set in which the analysis result information 271f and the operation information 272f are associated with each other.

[0122] In this embodiment, the estimation model 273 is updated using the learning data extracted in this manner, and the operation information 272z acquired at timing T8 is input to the updated estimation model 273. Then, the management device 200 acquires a predicted value 275z of the oil property output from the updated estimation model 273.

[0123] In this manner, in this embodiment, the range for extracting learning data is changed and the estimation model 273 is updated depending on the timing for predicting the oil state.

[0124] Therefore, in this embodiment, when estimating the state of the oil, the estimation model 273 updated to the latest state can be used, and even if the estimation model 273 has explanatory variables that change over time, an appropriate predicted value of the oil properties can be obtained.

[0125] In the example of FIG. 6, the range for extracting learning data is the period during which information was acquired, but the range is not limited to this.

[0126] The range from which learning data is extracted may be, for example, the amount of information to be the learning data. Specifically, for example, the range from which learning data is extracted may be a predetermined number of pieces of information acquired before a predetermined timing.

[0127] For example, if the predetermined timing is the timing T1 when new operation information is acquired and the predetermined number is 100, in Fig. 6(A), the learning data extraction unit 230 may extract 100 sets of analysis result information 271 acquired before timing T1 and operation information 272 corresponding to each analysis result information 271 from the storage unit 270, and use these sets as learning data. The predetermined timing and the predetermined number may be set in advance.

[0128] Also, in FIG. 6(B), the learning data extraction unit 230 extracts from the memory unit 270 a total of 100 data sets of analysis result information 271 acquired before timing T5 and operation information 272 corresponding to each analysis result information 271, and sets these as learning data.

[0129] Also, in FIG. 6(C), the learning data extraction unit 230 extracts from the memory unit 270 a total of 100 data sets of analysis result information 271 acquired before timing T8 and operation information 272 corresponding to each analysis result information 271, and sets these as learning data.

[0130] In this embodiment, the accuracy of machine learning can be adjusted by setting the range of learning data extraction to the amount of information (the number of data sets) in this way. Also, the estimation model 273 can be generated or updated according to the amount of information stored in the storage unit 270.

[0131] Furthermore, in this embodiment, the range for extracting the learning data may be a specific time period. The specific time period may be set in advance. Specifically, for example, the specific time period may be the same season as the season when the operation information 272 is acquired. In this case, if the timing T1 when the operation information 272 is acquired is spring, the learning data extraction unit 230 extracts a data set of the analysis result information 271 and the operation information 272 acquired during the past spring period (for example, from March to May) and sets it as learning data.

[0132] In this embodiment, the range from which learning data is extracted may be a period during which work was performed at the same work site in the past. The work site may be identified from location information included in the operation information, or may be input by a manager of the excavator 100, for example.

[0133] In this manner, in this embodiment, the range from which learning data is extracted can be changed depending on the situation, and the estimation model 273 can be updated with learning data according to the usage status of the shovel 100.

[0134] In addition, in FIG. 6, a case has been described in which the management device 200 updates the estimation model 273 when it acquires the operation information 272. However, the learning data extraction unit 230 of this embodiment may specify and extract a range of learning data even when new analysis result information 271 is acquired, and may update the estimation model 273 with the extracted learning data.

[0135] Specifically, when new analysis result information 271 is acquired, the learning data extraction unit 230 extracts, as learning data, a data set of analysis result information 271 acquired during a certain period in the past starting from the date and time indicated by collection date and time information included in the newly acquired analysis result information 271, and operation information 272 corresponding to the analysis result information 271. The newly acquired analysis result information 271 is included in the learning data extracted here.

[0136] Then, the management device 200 may update the estimation model 273 by the learning unit 240 using the extracted learning data.

[0137] In this embodiment, by doing so, the latest analysis result information 271 can be reflected in the estimation model 273 every time oil is sampled during regular maintenance.

[0138] Furthermore, in this embodiment, by using the estimation model 273, the oil state can be estimated every time the operation information 272 is acquired. That is, according to this embodiment, the oil state can be estimated more frequently than in periodic maintenance.

[0139] Specifically, for example, in a period K from timing T1 to timing T2 in FIG. 6(A), the estimation model 273 is not generated. Therefore, during this period, the oil state is grasped with the same frequency as regular maintenance. In contrast, for example, in a period K from timing T5 to timing T6 in FIG. 6(B) and a period K from timing T8 to timing T9 in FIG. 6(C), the estimation model 273 is generated, and the oil state is estimated every time the operation information 272 is acquired.

[0140] Therefore, according to this embodiment, it is possible to prompt the manager of the shovel 100 to change the oil at an appropriate time.

[0141] Next, the estimation result information 274 of this embodiment will be described with reference to Fig. 7. Fig. 7 is a diagram showing an example of the estimation result information.

[0142] The estimation result information 274 in this embodiment includes, as information items, date and time, model, machine number, pilot pressure, engine RPM, engine load factor, lever operation signal, operating state, predicted oil property value, etc. The values ​​of the items “date and time”, “model”, “machine number”, “pilot pressure”, “engine RPM”, “engine load factor”, “lever operation signal”, and “operating state” included in the estimation result information 274 are operation information, and are information input to the estimation model 273.

[0143] The value of the item "date and time" indicates the date and time when the operation information was acquired. The value of the item "model" indicates the model of the shovel 100, and the value of the item "machine number" indicates the machine identification number for identifying the shovel 100. The items "pilot pressure," "engine speed," "engine load factor," "lever operation signal," "operating state," etc. are values ​​acquired from various sensors possessed by the shovel 100.

[0144] In addition, the value of the item “oil property predicted value” included in the estimation result information 274 is a value resulting from estimating the state of the oil, and is output from the estimation model 273 .

[0145] In the example of FIG. 7, in estimation result information 274x, operation information 272x is associated with a predicted value 275x of the oil property, which is information indicating the estimation result of the oil.

[0146] In other words, in the example of Figure 7, the management device 200 inputs operation information 272x into the estimation model 273 and obtains a predicted value 275x of the oil properties that indicates the state of the oil in the shovel 100 at the time the operation information 272x was obtained.

[0147] Further, the management device 200 of the present embodiment estimates the state of the oil using the estimation model 273 every time the operation information 272 is acquired, but the present invention is not limited to this.

[0148] In this embodiment, for example, the state of the oil may be estimated when the operating state of the shovel 100 indicated by the operation information is a predetermined specified state.

[0149] Specifically, for example, when the operating state of the excavator 100 indicated by the operation information is an idling state, the operation information 272 may be input to the estimation model 273 to estimate the state of the oil.

[0150] When oil is collected from the shovel 100, it is expected that the shovel 100 is in a state close to idling. In other words, it is expected that the estimation model 273 is generated using, as learning data, a data set in which the analysis result information 271 and the operation information 272 acquired in an idling state are associated with each other.

[0151] Therefore, in this embodiment, by executing the process of FIG. 5 when the operating state of the shovel 100 is an idling state, it is possible to use the operating information 272 suitable for calculation by the estimation model 273 as input data, thereby improving the accuracy of the estimation.

[0152] In this embodiment, when the operating state of the shovel 100 is a state other than an idling state, the management device 200 may estimate the state of the oil by using a coefficient or the like according to the operating state. An operating state other than an idling state includes, for example, a state during excavation work.

[0153] In this embodiment, by doing so, a predicted value of the oil property according to the operating state of the shovel 100 can be obtained, and the oil state can be estimated with high accuracy.

[0154] Furthermore, in this embodiment, when the operating state of the shovel 100 indicated by the operation information is other than an idling state, a message or the like instructing the shovel 100 to change the operating state to an idling state may be sent to the shovel 100 and displayed on the display device 40 of the shovel 100.

[0155] Fig. 8 is a diagram showing an example of a display on a display device of a shovel. A message 40a is displayed on the display device 40 shown in Fig. 8, instructing the operation state of the shovel 100 to be put into an idling state in order to estimate the state of the oil.

[0156] When the operating state of the shovel 100 is set to an idling state after this message 40 a is displayed on the display device 40 , the shovel 100 acquires the operation information 272 and transmits it to the management device 200 .

[0157] In this embodiment, the message 40a is displayed on the display device 40 of the shovel 100, but this is not limiting. In this embodiment, for example, when the shovel 100 is operated from a remote control room, the message 40a may be displayed on the display device in the remote control room.

[0158] Furthermore, the management device 200 of this embodiment may transmit, instead of a message, a control signal to instruct the shovel 100 to be in an idling state. When the shovel 100 receives from the management device 200 a control signal instructing the shovel 100 to be in an idling mode, the shovel 100 may set its own operating state to an idling state in response to the instruction of the control signal, and obtain and transmit the operation information 272 to the management device 200.

[0159] Next, a notification when it is determined that the oil needs to be changed based on the estimation result of the oil condition will be described with reference to Fig. 9. Fig. 9 is a diagram showing an example of a display of the estimation result of the oil condition.

[0160] A screen 90 shown in FIG. 9 is an example of a screen displayed on a terminal device or the like carried by the manager of the shovel 100, for example.

[0161] The management device 200 of this embodiment may determine that an oil change is necessary and display a screen 90 on the manager's terminal device or the like when a predicted value of the oil properties obtained by inputting newly acquired operation information 272 into the estimation model 273 is equal to or greater than a predetermined threshold value.

[0162] The screen 90 includes display areas 90a, 90b, and 90c. The display area 90a displays information indicating changes over time in the sensor values ​​included in the operation information input to the estimation model 273, and the predicted values ​​and analysis result information of the oil properties output from the estimation model 273.

[0163] The display area 90b displays the model number of the excavator 100 and the date and time when the oil condition was estimated. The display area 90c displays a message indicating that the oil has deteriorated and needs to be changed.

[0164] In the example of FIG. 9, a marker 90d indicating the range of learning data used to update the estimation model 273, and a marker 90e indicating the newly acquired operating information 272 and the predicted value of the oil properties may be displayed in the display area 90a.

[0165] In this embodiment, the sensor values ​​contained in the operation information 272 and the changes over time in the predicted oil properties are displayed in correspondence with a message, allowing the manager of the shovel 100 to understand the changes in the oil condition.

[0166] Furthermore, in this embodiment, by displaying the extraction range of the learning data, the manager of the shovel 100 or the like can understand how the estimation model 273 is updated.

[0167] In the above-described embodiment, the state of the oil (hydraulic oil) of the shovel 100 is estimated, but the present invention is not limited to this. This embodiment can be applied to an object whose state is monitored by manually collecting and analyzing the oil. For example, this embodiment can be applied to a monitoring target discharged from a plant, factory, etc.

[0168] The preferred embodiments of the present invention have been described above in detail. However, the present invention is not limited to the above-described embodiments. Various modifications, substitutions, etc. may be applied to the above-described embodiments without departing from the scope of the present invention. Furthermore, features described separately may be combined unless technical contradictions arise. [Explanation of symbols]

[0169] 30 Controller 40 Display device 100 Shovel 200 Management device 210 Analysis result acquisition section 220 Operation Information Acquisition Department 230 Learning Data Extraction Unit 240 Learning Department 250 State Estimation Unit 260 Output section 270 Storage section

Claims

1. A management system including a work machine and a management device for the work machine, The management device includes: a model that learns a relationship between operation information of the work machine and a state of hydraulic oil of the work machine; an operation information acquisition unit that acquires the operation information from the work machine; a state estimation unit that inputs the operation information acquired by the operation information acquisition unit into the model and acquires estimation result information that estimates a state of the hydraulic oil.

2. The management device includes:

2. The work machine management system according to claim 1, further comprising a learning unit that generates the model using a data set that associates analysis result information indicating an analysis result of the hydraulic oil sampled from the work machine with operation information corresponding to the analysis result information as learning data.

3. 3. A management system for a work machine according to claim 2, wherein the estimation result information includes the operation information input to the model and a predicted value indicating a state of the hydraulic oil output from the model.

4. The management device includes: A storage unit in which the data set is stored; and a learning data extraction unit that specifies a range from which learning data is extracted; 4. The management system for a work machine according to claim 3, wherein the learning data extraction unit further extracts, from the storage unit, a data set acquired during a period indicated by the range, and sets the data as learning data.

5. The management device includes: A storage unit in which the data set is stored; and a learning data extraction unit that specifies a range from which learning data is extracted; 4. The work machine management system according to claim 3, wherein the learning data extraction unit further extracts a data set from the storage unit so that the amount of information is equal to the amount indicated by the range, and sets the data as learning data.

6. The state estimation unit is 2. A work machine management system according to claim 1, wherein when an operating state of the work machine indicated by the operation information acquired by the operation information acquisition unit is a predetermined state, the operation information is input to the model.

7. The management device includes:

4. A management system for a work machine according to claim 3, further comprising an output section which determines whether or not the hydraulic oil needs to be replaced based on a predicted value indicating the state of the hydraulic oil, and outputs a result of the determination.

Citation Information

Patent Citations

  • Work machinery

    JP7009408B2