Load estimation device, load estimation system, load estimation method, and load estimation program

Abstract

A controller (200) of a load estimation device (2): determines an operation of a work machine (3), using operation data of the work machine (3); selects at least one load estimation model corresponding to the determined operation of the work machine (3) from among a plurality of load estimation models respectively associated with a plurality of operations of the work machine; and estimates a load on the work machine (3), using the selected at least one load estimation model and the operation data.

Description

LOAD ESTIMATION DEVICE, LOAD ESTIMATION SYSTEM, LOAD ESTIMATION METHOD, AND LOAD ESTIMATION PROGRAM

[0001] The present disclosure relates to a technique for estimating a load on a work machine.

[0002] Patent Document 1 discloses a technique for estimating damage parameters relating to damage to a predetermined portion of a work machine.

[0003] Work machines perform a variety of operations. Specifically, a hydraulic excavator, which is one example of a work machine, performs a variety of operations, such as excavation, swinging, and earth removal. Therefore, it is desirable to develop a technology that can estimate the load on a work machine depending on the type of operation of the work machine.

[0004] Japanese Patent Application Laid-Open No. 2020-128656

[0005] An object of the present disclosure is to provide a technique that can estimate the load on a work machine according to the type of operation of the work machine.

[0006] The provided load estimation device includes a controller, which determines the operation of a work machine using operation data of the work machine, selects at least one load estimation model corresponding to the determined operation of the work machine from a plurality of load estimation models respectively associated with a plurality of operations of the work machine, and estimates the load on the work machine using the selected at least one load estimation model and the operation data.

[0007] FIG. 1 is a diagram showing the overall configuration of a load estimation system according to an embodiment. FIG. 2 is a diagram showing a work machine of the load estimation system. FIG. 3 is a diagram showing an example of multiple operations of the work machine. FIG. 4 is a block diagram showing the configuration of the work machine. FIG. 5 is a block diagram showing the configuration of a load estimation device of the load estimation system. FIG. 6 is a diagram showing an example of multiple load estimation models stored in a load estimation model storage unit of the load estimation system. FIG. 7 is a diagram showing another example of multiple load estimation models stored in a load estimation model storage unit of the load estimation system. FIG. 8 is a diagram showing yet another example of multiple load estimation models stored in a load estimation model storage unit of the load estimation system. A flowchart showing an example of calculation processing performed by a machine controller of the work machine. A flowchart showing an example of calculation processing performed by a controller of the load estimation device. A diagram showing an example of processing to determine the operation of the work machine using operation data of the work machine. A diagram showing an example of output from the controller. A diagram for explaining database selection. A diagram for explaining a local load estimation model.

[0008] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. Note that the following embodiments are examples that embody the present disclosure and do not limit the technical scope of the present disclosure.

[0009] FIG. 1 is a diagram showing the overall configuration of a load estimation system 1 according to this embodiment.

[0010] 1, the load estimation system 1 includes a load estimation device 2, a work machine 3, and a display device 4. The load estimation device 2 is connected to the work machine 3 and the display device 4 via a network 5 so that they can communicate with each other. The display device 4 is an example of an output device in the present disclosure.

[0011] The network 5 is an information communication network such as the Internet, a telephone network, a mobile phone network, a satellite communication network, a wide area network (WAN), a local area network (LAN), a dedicated line, etc. The network 5 may be any one of these communication networks, or may be configured by combining a plurality of types of communication networks.

[0012] FIG. 2 is a diagram showing the work machine 3 of the load estimation system 1.

[0013] The work machine 3 includes a lower traveling body 10 capable of traveling on the ground G, an upper rotating body 12 supported by the lower traveling body 10, and a work implement 14 supported by the upper rotating body 12. Note that in the present embodiment, a hydraulic excavator is shown as an example of the work machine 3, but the present disclosure is not limited to this, and the work machine 3 may be another work machine such as a crane or a bulldozer, for example.

[0014] The upper rotating body 12 has a rotating frame 16 and a plurality of elements mounted on the rotating frame 16. The plurality of elements include a machine room 17 that houses a power source (not shown) and a cab 18 that serves as a driver's room. The power source may be an engine or a battery. The upper rotating body 12 is attached to the lower traveling body 10 so as to be able to rotate. The upper rotating body 12 is an example of a movable part in the present disclosure.

[0015] The work device 14 includes a boom 21, an arm 22, and a bucket 24. The boom 21 has a base end supported on the front of the rotating frame 16 so as to be able to rise and fall, i.e., so as to be rotatable about a horizontal axis, and a tip end opposite the base end. The arm 22 has a base end attached to the tip end of the boom 21 so as to be rotatable about a horizontal axis, and a tip end opposite the base end. The bucket 24 is rotatably attached to the tip end of the arm 22. Each of the boom 21, the arm 22, and the bucket 24 is an example of a movable part in the present disclosure.

[0016] The work machine 3 is equipped with a plurality of actuators. The plurality of actuators include a boom cylinder 26, an arm cylinder 27, a bucket cylinder 28, a swing motor 29, a traveling motor 30L, and a traveling motor 30R (see FIG. 4). The boom cylinder 26, the arm cylinder 27, and the bucket cylinder 28 are each hydraulic cylinders that operate upon a supply of hydraulic oil discharged from a hydraulic pump (not shown). The swing motor 29, the traveling motor 30L, and the traveling motor 30R are each hydraulic motors that operate upon a supply of hydraulic oil discharged from the hydraulic pump.

[0017] The boom cylinder 26 is interposed between the upper rotating body 12 and the boom 21 and extends and retracts to cause the boom 21 to perform a raising and lowering operation. Specifically, the boom cylinder 26 has a head-side chamber and a rod-side chamber. When hydraulic oil is supplied to the head-side chamber, the boom cylinder 26 extends, moving the boom 21 in the boom-up direction and discharging hydraulic oil from the rod-side chamber. On the other hand, when hydraulic oil is supplied to the rod-side chamber, the boom cylinder 26 contracts, moving the boom 21 in the boom-down direction and discharging hydraulic oil from the head-side chamber.

[0018] The arm cylinder 27 is interposed between the boom 21 and the arm 22 and extends and retracts to cause the arm 22 to rotate. Specifically, the arm cylinder 27 has a head-side chamber and a rod-side chamber. When hydraulic oil is supplied to the head-side chamber, the arm cylinder 27 extends, moving the arm 22 in the arm-pushing direction and discharging hydraulic oil from the rod-side chamber. The arm-pushing direction is the direction in which the tip of the arm 22 approaches the boom 21. On the other hand, when hydraulic oil is supplied to the rod-side chamber, the arm cylinder 27 contracts, moving the arm 22 in the arm-pushing direction and discharging hydraulic oil from the head-side chamber. The arm-pushing direction is the direction in which the tip of the arm 22 moves away from the boom 21.

[0019] The bucket cylinder 28 is interposed between the arm 22 and the bucket 24 and extends and contracts to cause the bucket 24 to rotate. Specifically, the bucket cylinder 28 has a head-side chamber and a rod-side chamber. When hydraulic oil is supplied to the head-side chamber, the bucket cylinder 28 extends, rotating the bucket 24 in the scooping direction and discharging hydraulic oil from the rod-side chamber. The scooping direction is the direction in which the tip 25 of the bucket 24 approaches the upper rotating body 12. On the other hand, when hydraulic oil is supplied to the rod-side chamber, the bucket cylinder 28 contracts, rotating the bucket 24 in the opening direction and discharging hydraulic oil from the head-side chamber. The opening direction is the direction in which the tip 25 of the bucket 24 moves away from the upper rotating body 12.

[0020] The swing motor 29 has a pair of ports. When hydraulic oil is supplied to one of the pair of ports, the swing motor 29 rotates to rotate the upper swing body 12 to the right and discharges hydraulic oil from the other of the pair of ports. When hydraulic oil is supplied to the other of the pair of ports, the swing motor 29 rotates to rotate the upper swing body 12 to the left and discharges hydraulic oil from one of the pair of ports.

[0021] Each of the left and right travel motors 30L, 30R has a motor output shaft that rotates in both directions by receiving hydraulic oil from a hydraulic pump, and causes the undercarriage 10 connected to the motor output shaft to travel forward or backward. The travel motors 30L and 30R rotate at the same speed, causing the undercarriage 10 to move forward or backward. On the other hand, the travel motors 30L and 30R rotate at different speeds, causing the undercarriage 10 to turn relative to the ground G.

[0022] The work machine 3 performs a plurality of operations. The plurality of operations may include, for example, the excavation operation shown in Fig. 3(A), the lifting and swinging operation shown in Fig. 3(B), the earth-discharging operation shown in Fig. 3(C), and the return swinging operation shown in Fig. 3(D), or may include operations other than these. The excavation operation, lifting and swinging operation, earth-discharging operation, and return swinging operation are a series of operations performed in this order.

[0023] The excavation operation shown in Figure 3(A) is an operation for storing earth and sand into the bucket 24 by moving the bucket 24 in a direction approaching the undercarriage 10 with part of the bucket 24 placed in the soil. This excavation operation includes an arm retraction operation, a boom raising operation, and a bucket scooping operation. The arm retraction operation is an operation in which the arm 22 rotates in the arm retraction direction. The boom raising operation is an operation in which the boom 21 rotates in the boom raising direction. The bucket scooping operation is an operation in which the bucket 24 rotates in the scooping direction.

[0024] The lifting and swinging operation shown in Figure 3(B) is an operation performed after the excavation operation. The lifting and swinging operation includes a boom-raising operation and a swinging operation that are performed with earth and sand loaded in the bucket 24 during the excavation operation. By performing this lifting and swinging operation, the bucket 24 is positioned directly above the bed of the earth and sand transport vehicle. The swinging operation is an operation in which the upper swinging body 12 swings right or left relative to the lower traveling body 10 around a swing axis Z that extends vertically.

[0025] The earth-discharging operation shown in Figure 3(C) is an operation performed after the lifting and swinging operation. The earth-discharging operation is an operation in which the bucket 24 is positioned directly above the bed of the earth-carrying vehicle and the earth and sand in the bucket 24 is discharged onto the bed. This earth-discharging operation includes an arm-pushing operation and a bucket-opening operation. The arm-pushing operation is an operation in which the arm 22 rotates in the arm-pushing direction. The bucket-opening operation is an operation in which the bucket 24 rotates in the opening direction.

[0026] The return swing operation shown in Figure 3 (D) is an operation performed after the soil discharge operation. The return swing operation includes a boom lowering operation and a swing operation. The boom lowering operation is an operation in which the boom 21 rotates in the boom lowering direction relative to the upper swing structure 12. The swing operation in the return swing operation is an operation in which the upper swing structure 12 rotates in the direction opposite to the swing direction in the lifting swing operation. By performing this return swing operation, the bucket 24 is positioned directly above the soil to be excavated.

[0027] 4 is a block diagram showing the configuration of the work machine 3. The work machine 3 is equipped with a machine controller 100, a boom cylinder pressure sensor 111, an arm cylinder pressure sensor 112, a bucket cylinder pressure sensor 113, a swing motor pressure sensor 114, a swing sensor 115, an attitude sensor 116, an operation device 117, a communication unit 118, and a hydraulic circuit 119.

[0028] The hydraulic circuit 119 includes a boom cylinder 26, an arm cylinder 27, a bucket cylinder 28, a swing motor 29, a pair of left and right travel motors 30L, 30R, a pair of boom solenoid valves 31, a pair of arm solenoid valves 32, a pair of bucket solenoid valves 33, a pair of swing solenoid valves 34, a pair of left travel solenoid valves 35L, a pair of right travel solenoid valves 35R, a boom control valve 36, an arm control valve 37, a bucket control valve 38, a swing control valve 39, and a pair of left and right travel control valves 40L, 40R.

[0029] The boom control valve 36 is, for example, a hydraulic pilot changeover valve having a pair of boom pilot ports. When boom pilot pressure is input to one of the pair of boom pilot ports, the boom control valve 36 opens in a direction corresponding to the boom pilot port and with a stroke corresponding to the magnitude of the boom pilot pressure. This changes the direction and flow rate of hydraulic oil supplied to the boom cylinder 26.

[0030] The arm control valve 37 is, for example, a hydraulic pilot switching valve having a pair of arm pilot ports. When arm pilot pressure is input to one of the pair of arm pilot ports, the arm control valve 37 opens in a direction corresponding to the arm pilot port and with a stroke corresponding to the magnitude of the arm pilot pressure. This changes the direction and flow rate of the hydraulic oil supplied to the arm cylinder 27.

[0031] The bucket control valve 38 is, for example, a hydraulic pilot changeover valve having a pair of bucket pilot ports. When bucket pilot pressure is input to one of the pair of bucket pilot ports, the bucket control valve 38 opens in a direction corresponding to the bucket pilot port and with a stroke corresponding to the magnitude of the bucket pilot pressure. This changes the direction and flow rate of hydraulic oil supplied to the bucket cylinder 28.

[0032] The swing control valve 39 is, for example, a hydraulic pilot switching valve having a pair of swing pilot ports. When a swing pilot pressure is input to one of the pair of swing pilot ports, the swing control valve 39 opens in a direction corresponding to the swing pilot port with a stroke corresponding to the magnitude of the swing pilot pressure. This changes the direction and flow rate of the hydraulic oil supplied to the swing motor 29.

[0033] Each of the travel control valves 40L, 40R is a hydraulic pilot changeover valve having a pair of travel pilot ports. When a travel pilot pressure is input to one of the pair of travel pilot ports, each of the travel control valves 40L, 40R opens in a direction corresponding to the travel pilot port with a stroke corresponding to the magnitude of the travel pilot pressure. This changes the direction and flow rate of hydraulic oil supplied to the travel motors 30L, 30R.

[0034] The pair of boom solenoid valves 31 are proportional solenoid valves respectively interposed between the pilot pump and a pair of boom pilot ports of the boom control valve 36, and perform opening and closing operations in response to input of a boom command signal, which is an electrical signal output by the machine controller 100. Upon receiving input of a boom command signal, the pair of boom solenoid valves 31 adjust the boom pilot pressure to a degree corresponding to the boom command signal.

[0035] The pair of arm solenoid valves 32 are proportional solenoid valves respectively interposed between the pilot pump and a pair of arm pilot ports of the arm control valve 37, and perform opening and closing operations in response to input of an arm command signal, which is an electric signal output by the machine controller 100. When input of an arm command signal, the pair of arm solenoid valves 32 adjust the arm pilot pressure to a degree according to the arm command signal.

[0036] The pair of bucket solenoid valves 33 are proportional solenoid valves respectively interposed between the pilot pump and a pair of arm pilot ports of the bucket control valve 38, and perform opening and closing operations in response to input of a bucket command signal, which is an electrical signal output by the machine controller 100. When input of a bucket command signal, the pair of bucket solenoid valves 33 adjust the bucket pilot pressure to a degree according to the bucket command signal.

[0037] The pair of swing solenoid valves 34 are proportional solenoid valves respectively interposed between the pilot pump and a pair of swing pilot ports of the swing control valve 39, and perform opening and closing operations in response to an input of a swing command signal, which is an electrical signal output by the machine controller 100. Upon receiving an input of a swing command signal, the swing solenoid valve 34 adjusts the swing pilot pressure to a degree corresponding to the swing command signal.

[0038] The pair of travel solenoid valves 35L are proportional solenoid valves respectively interposed between the pilot pump and a pair of travel pilot ports of the travel control valve 40L, and perform opening and closing operations in response to input of a travel command signal, which is an electrical signal output by the machine controller 100. When input of a travel command signal, the pair of travel solenoid valves 35L adjust the travel pilot pressure to a degree corresponding to the travel command signal.

[0039] The pair of travel solenoid valves 35R are proportional solenoid valves respectively interposed between the pilot pump and a pair of travel pilot ports of the travel control valve 40R, and perform opening and closing operations in response to input of a travel command signal, which is an electrical signal output by the machine controller 100. When input of a travel command signal, the travel solenoid valve 35R adjusts the travel pilot pressure to a degree corresponding to the travel command signal.

[0040] The boom cylinder pressure sensor 111 detects the pressure value of the boom cylinder 26. Specifically, the boom cylinder pressure sensor 111 includes a boom cylinder head pressure sensor and a boom cylinder rod pressure sensor. The boom cylinder head pressure sensor detects the boom cylinder head pressure, which is the pressure of hydraulic oil in the head-side chamber of the boom cylinder 26. The boom cylinder rod pressure sensor detects the boom cylinder rod pressure, which is the pressure of hydraulic oil in the rod-side chamber of the boom cylinder 26. The boom cylinder pressure sensor 111 converts the detected boom cylinder head pressure and boom cylinder rod pressure into detection signals, which are corresponding electrical signals, and inputs them to the machine controller 100.

[0041] The arm cylinder pressure sensor 112 detects the pressure value of the arm cylinder 27. Specifically, the arm cylinder pressure sensor 112 includes an arm cylinder head pressure sensor and an arm cylinder rod pressure sensor. The arm cylinder head pressure sensor detects the arm cylinder head pressure, which is the pressure of hydraulic oil in the head side chamber of the arm cylinder 27. The arm cylinder rod pressure sensor detects the arm cylinder rod pressure, which is the pressure of hydraulic oil in the rod side chamber of the arm cylinder 27. The arm cylinder pressure sensor 112 converts the detected arm cylinder head pressure and arm cylinder rod pressure into detection signals, which are corresponding electric signals, and inputs them to the machine controller 100.

[0042] The bucket cylinder pressure sensor 113 detects the pressure value of the bucket cylinder 28. Specifically, the bucket cylinder pressure sensor 113 includes a bucket cylinder head pressure sensor and a bucket cylinder rod pressure sensor. The bucket cylinder head pressure sensor detects the bucket cylinder head pressure, which is the pressure of hydraulic oil in the head-side chamber of the bucket cylinder 28. The bucket cylinder rod pressure sensor detects the bucket cylinder rod pressure, which is the pressure of hydraulic oil in the rod-side chamber of the bucket cylinder 28. The bucket cylinder pressure sensor 113 converts the detected bucket cylinder head pressure and bucket cylinder rod pressure into detection signals, which are corresponding electrical signals, and inputs them to the machine controller 100.

[0043] The swing motor pressure sensor 114 detects the operating pressure value of the swing motor 29, i.e., the motor differential pressure. Specifically, the swing motor pressure sensor 114 includes a right swing port pressure sensor and a left swing port pressure sensor. The right swing port pressure sensor detects the right swing port pressure, which is the pressure of hydraulic oil in the right swing port of the swing motor 29. The left swing port pressure sensor detects the left swing port pressure, which is the pressure of hydraulic oil in the left swing port of the swing motor 29. The swing motor pressure sensor 114 converts the detected differential pressure between the right swing port pressure and the left swing port pressure into a corresponding detection signal, which is an electrical signal, and inputs it to the machine controller 100.

[0044] In addition, the swing motor pressure sensor 114 may convert the detected right swing port pressure into a corresponding detection signal, which is an electrical signal, and input it to the machine controller 100, or may convert the detected left swing port pressure into a corresponding detection signal, which is an electrical signal, and input it to the machine controller 100.

[0045] The rotation sensor 115 is configured by, for example, a resolver or a rotary encoder, and detects the rotation angle of the upper rotating body 12 relative to the lower traveling body 10. The rotation sensor 115 converts the detected rotation angle into a corresponding detection signal, which is an electrical signal, and inputs it to the machine controller 100.

[0046] The attitude sensor 116 detects the attitude of the work implement 14. The attitude sensor 116 may include the boom attitude sensor 61, the arm attitude sensor 62, and the bucket attitude sensor 64 shown in FIG.

[0047] The boom attitude sensor 61 detects the boom attitude, which is the attitude of the boom 21. The boom attitude may be the rotation angle of the boom 21 relative to the upper rotating body 12, the attitude of the boom 21 relative to the horizontal plane, or the cylinder length of the boom cylinder 26. The arm attitude sensor 62 detects the arm attitude, which is the attitude of the arm 22. The arm attitude may be the rotation angle of the arm 22 relative to the boom 21, the attitude of the arm 22 relative to the horizontal plane, or the cylinder length of the arm cylinder 27. The bucket attitude sensor 64 detects the bucket attitude, which is the attitude of the bucket 24. The bucket attitude may be the rotation angle of the bucket 24 relative to the arm 22, the attitude of the bucket 24 relative to the horizontal plane, or the cylinder length of the bucket cylinder 28.

[0048] The boom attitude sensor 61, the arm attitude sensor 62, and the bucket attitude sensor 64 may each be, for example, a resolver, a rotary encoder, a potentiometer, or an IMU (Inertial Measurement Unit). The attitude sensor 116 converts the detected boom attitude, arm attitude, and bucket attitude into detection signals, which are corresponding electrical signals, and inputs them to the machine controller 100.

[0049] The operation device 117 receives operations by an operator for operating the work device 14, rotating the upper rotating body 12, and traveling the undercarriage 10. The operation device 117 includes a boom operation device, an arm operation device, a bucket operation device, a swing operation device, and a traveling operation device.

[0050] The boom operation device has a boom operation lever that receives operation from an operator for boom-raising or boom-lowering operation. The boom operation device may be configured as an electric lever device that includes an operation signal generation unit that inputs the operation amount of the boom operation lever to the machine controller 100.

[0051] The arm operating device has an arm operating lever that receives an operation from an operator for an arm pulling operation or an arm pushing operation. The arm operating device may be configured as an electric lever device that includes an operation signal generating unit that inputs the operation amount of the arm operating lever to the machine controller 100.

[0052] The bucket operating device has a bucket operating lever that receives operation from an operator for the bucket scooping operation or the bucket opening operation. The bucket operating device may be configured as an electric lever device that includes an operation signal generating unit that inputs the operation amount of the bucket operating lever to machine controller 100.

[0053] The swing operation device has a swing operation lever that receives an operation from an operator to swing the upper swing body 12 right or left. The swing operation device may be configured as an electric lever device that includes an operation signal generation unit that inputs the operation amount of the swing operation lever to the machine controller 100.

[0054] The travel operation device has a travel operation lever that receives operation from an operator to move the undercarriage 10 forward or backward. The travel operation device may be configured as an electric lever device that includes an operation signal generation unit that inputs the operation amount of the travel operation lever to the machine controller 100.

[0055] The boom operation lever, arm operation lever, bucket operation lever, and swing operation lever may be configured as two operation levers, in which case one of the two operation levers plays the role of two of the boom operation lever, arm operation lever, bucket operation lever, and swing operation lever, and the other of the two operation levers plays the role of the remaining two.

[0056] The machine controller 100 comprises a computer including a processor and a memory. The machine controller 100 comprises an operation data acquisition unit 102 and a command unit 103. The functions of the operation data acquisition unit 102 and the command unit 103 are realized by the processor executing a control program stored in the memory.

[0057] The operation data acquisition unit 102 acquires operation data of the work machine 3. The operation data may include data related to the operation of a plurality of movable parts. The plurality of movable parts include the boom 21, the arm 22, the bucket 24, and the upper rotating body 12. In this case, the operation data may include sensor values ​​that detect physical quantities that change in accordance with the operation of the work machine 3. That is, the operation data may include sensor values ​​detected by a group of sensors such as sensors 111 to 116. Specifically, the operation data may include, for example, the pressure value of the boom cylinder 26, the pressure value of the arm cylinder 27, the pressure value of the bucket cylinder 28, a posture value related to the boom posture, a posture value related to the arm posture, a posture value related to the bucket posture, an operating pressure value of the swing motor 29, and the swing angle. However, the operation data does not necessarily have to include all of these sensor values. Furthermore, the operation data may include sensor values ​​other than these sensor values.

[0058] The operation data acquiring unit 102 may store operation data acquired from the sensors 111 to 116. The operation data acquiring unit 102 may store operation data acquired multiple times within a predetermined time period. The predetermined time period may be, for example, one minute, ten minutes, one hour, or another time period.

[0059] The command unit 103 controls the operation of each element included in the hydraulic circuit 119. The command unit 103 includes a boom command unit, an arm command unit, a bucket command unit, a swing command unit, and a travel command unit.

[0060] The boom command unit inputs a boom command signal having a value corresponding to the amount of operation of the boom operation device to the boom solenoid valve 31. As a result, the flow rate of hydraulic oil supplied to the boom cylinder 26 increases as the amount of operation of the boom operation device increases.

[0061] The arm command unit inputs an arm command signal having a value corresponding to the amount of operation of the arm operating device to the arm electromagnetic valve 32. As a result, the flow rate of hydraulic oil supplied to the arm cylinder 27 increases as the amount of operation of the arm operating device increases.

[0062] The bucket command unit inputs a bucket command signal having a value corresponding to the operation amount of the bucket operating device to the bucket solenoid valve 33. As a result, the flow rate of hydraulic oil supplied to the bucket cylinder 28 increases as the operation amount of the bucket operating device increases.

[0063] The swing command unit inputs a swing command signal having a value corresponding to the amount of operation of the swing operation device to the swing solenoid valve 34. As a result, the flow rate of hydraulic oil supplied to the swing motor 29 increases as the amount of operation of the swing operation device increases.

[0064] The travel command unit inputs a travel command signal having a value corresponding to the amount of operation of the travel operation device to the travel solenoid valve 35L and the travel solenoid valve 35R. As a result, the flow rate of hydraulic oil supplied to the travel motors 30L, 30R increases as the amount of operation of the travel operation device increases.

[0065] The communication unit 118 includes an operational data transmission unit 106. The operational data transmission unit 106 transmits the operational data acquired by the operational data acquisition unit 102 to the load estimation device 2.

[0066] The load estimation device 2 may be, for example, an information terminal such as a tablet computer (a so-called tablet), a smartphone, a laptop personal computer, or a desktop personal computer, or may be, for example, a management device such as a server. The management device may be, for example, a computer (cloud server) in a cloud service provided as a service over a network such as the Internet.

[0067] 5 is a block diagram showing the configuration of the load estimation device 2. The load estimation device 2 includes a controller 200 and a communication unit 210.

[0068] The controller 200 includes a computer including a processor 220 and a memory 230. The processor 220 of the controller 200 includes an operation determination unit 221, a model selection unit 222, and a load estimation unit 223. The processor 220 may further include one or both of a lifespan calculation unit 224 and a display information generation unit 225. However, the lifespan calculation unit 224 and the display information generation unit 225 are not essential components of the load estimation device of the present disclosure. The functions of the operation determination unit 221, the model selection unit 222, the load estimation unit 223, the lifespan calculation unit 224, and the display information generation unit 225 are realized by the processor 220 executing a control program (load estimation program) stored in the memory 230.

[0069] The movement determination unit 221 determines the movement of the work machine 3 using the operation data of the work machine 3. Specifically, the movement determination unit 221 acquires the operation data of the work machine 3 that is the estimation target, and determines the movement of the work machine 3 by inputting the operation data received by the operation data receiving unit 211 into the movement determination model stored in the movement determination model storage unit 231.

[0070] The model selection unit 222 selects at least one load estimation model that corresponds to the operation of the work machine 3 determined by the operation determination unit 221 from a plurality of load estimation models that are respectively associated with a plurality of operations of the work machine 3 .

[0071] The load estimation unit 223 estimates the load on the work machine 3 using the at least one load estimation model selected by the model selection unit 222 and the operation data.

[0072] The lifespan calculation unit 224 calculates the lifespan of the work machine 3 based on the load on the work machine 3 estimated by the load estimation unit 223. The lifespan calculation unit 224 may perform a frequency analysis of stress using the rainflow method from the time change in the load on the work machine 3 estimated by the load estimation unit 223 (for example, stress occurring at a predetermined part of the work machine 3). The lifespan calculation unit 224 may calculate the degree of damage increased per unit time using Miner's rule from the analysis results. The lifespan calculation unit 224 may calculate the degree of damage up to the present time by adding the calculated degree of damage to the degree of damage calculated up to the previous time. The lifespan calculation unit 224 may then calculate the remaining lifespan by subtracting the degree of damage up to the present time from the design lifespan of the work machine. The lifespan calculation unit 224 is also capable of calculating the lifespan using various conventional techniques other than the above-mentioned methods.

[0073] The display information generation unit 225 generates display information for presenting to the workers the load on the work machine 3 calculated by the load estimator 223. The display information generation unit 225 may also generate display information for presenting to the workers the lifespan of the work machine 3 calculated by the lifespan calculation unit 224.

[0074] The memory 230 of the controller 200 includes an operation determination model storage unit 231 and a load estimation model storage unit 232 .

[0075] The movement determination model storage unit 231 stores a movement determination model that uses operation data of the work machine 3 as input values ​​and the movement of the work machine 3 as output values. Specifically, for example, the movement determination model storage unit 231 may store a movement determination model that has been constructed by machine learning, with operation data of the work machine 3 as input values ​​and the movement of the work machine 3 as output values. The machine learning may be machine learning using teacher data.

[0076] The load estimation model storage unit 232 stores a database including a plurality of load estimation models respectively associated with a plurality of operations of the work machine 3. For example, the load estimation model storage unit 232 may store a database including a plurality of load estimation models that differ for each operation of the work machine 3. The load estimation model storage unit 232 may store each of a plurality of operations of the work machine 3 that have been set in advance, in association with one of a plurality of load estimation models or at least one of a plurality of load estimation models.

[0077] The database may be constructed by creating an initial database using past operational data and then further improving the estimation accuracy of the initial database through learning, as shown in the left diagram of FIG. 14, for example.

[0078] The load on the work machine 3 varies depending on the operation performed by the work machine 3. Furthermore, even when the work machine 3 performs the same operation, the load on the work machine 3 may vary depending on factors such as the operator's level of skill, weather conditions, and the surrounding environment. Therefore, it is preferable that the database include a plurality of load estimation models calculated under a variety of conditions.

[0079] Specifically, when the operation of the work machine 3 differs, the external forces acting on the work machine 3 also change. For example, when the operation of the work machine 3 is an excavation operation, the load on the work implement 14, including the bucket 24, arm 22, and boom 21, increases, and the stress corresponding to this increases. On the other hand, when the operation of the work machine 3 is a lifting and swinging operation, an earth-discharging operation, or a return swinging operation, the load and stress on the work implement 14 become relatively small. Furthermore, when the operation of the work machine 3 is a lifting and swinging operation or a return swinging operation, the load on the parts of the upper rotating body 12 involved in the swinging operation becomes large, and the stress corresponding to this becomes large, whereas when the operation of the work machine 3 is an excavation operation or an earth-discharging operation, the load and stress on the parts of the upper rotating body 12 involved in the swinging operation become relatively small.

[0080] In this way, differences in the operation of the work machine 3 affect the load on the work machine 3. Furthermore, a certain correlation exists between the operation of the work machine 3 and the operation data of the work machine 3. Therefore, a certain correlation exists between the load on the work machine 3 and the operation data. Therefore, it is possible to construct multiple load estimation models that use the operation data of the work machine 3 as input values ​​and the load on the work machine 3 as output values.

[0081] The database stores load estimation models that are suitable for a plurality of operations of the work machine 3. In other words, the database stores a plurality of load estimation models that correspond to a plurality of operations of the work machine 3. The correlation between the operation of the work machine 3 and the load on the work machine 3 may be acquired using a method such as experimentation, simulation, or machine learning. A plurality of load estimation models that correspond to a plurality of operations of the work machine 3 are designed based on the correlation between the operation of the work machine 3 and the load on the work machine 3 acquired by this method, and a database containing the designed plurality of load estimation models is stored in the load estimation model storage unit 232. By constructing a database that stores load estimation models set for each operation of the work machine 3 in this way, the controller 200 can estimate the load for a plurality of operations of the work machine 3.

[0082] In this embodiment, each of the plurality of load estimation models stored in the database includes an operating status parameter and a model parameter θ associated with the operating status parameter.

[0083] The operating status parameters may be set, for example, as follows. For example, when the work machine 3 performs an excavation operation, the sensor values ​​detected by the group of sensors (e.g., sensors 111 to 116) change in a somewhat characteristic manner, and the sensor values ​​also change in a somewhat characteristic manner for each of the lifting and swinging operation, the earth-discharging operation, and the return swinging operation. Therefore, the operating status parameters corresponding to the excavation operation can be designed in advance based on the characteristic changes in the sensor values ​​acquired using techniques such as experiments, simulations, and machine learning. Similarly, the operating status parameters corresponding to the lifting and swinging operation, the earth-discharging operation, and the return swinging operation can be designed in advance based on the characteristic changes in the sensor values ​​acquired using the techniques.

[0084] The model parameter θ is a parameter used by the controller 200 to estimate the load on the work machine 3. As described above, the load on the work machine 3 varies depending on the operation of the work machine 3, and there is a certain degree of correlation between the load and the operation of the work machine 3. Therefore, the model parameter θ corresponding to the excavation operation can be designed in advance using techniques such as experiments, simulations, and machine learning. Similarly, the model parameters θ corresponding to each of the lifting swing operation, the soil discharge operation, and the return swing operation can be designed in advance using the above techniques.

[0085] The model selection unit 222 extracts a load estimation model that includes operating status parameters similar to the actual operation data from among a plurality of load estimation models included in the database, and the load estimation unit 223 estimates the load on the work machine 3 using the model parameter θ included in the extracted load estimation model and the operation data. Specific examples of these processes are as follows:

[0086] The model selection unit 222 compares operation data acquired at a certain time t with multiple load estimation models in the database, and extracts from the multiple load estimation models a load estimation model that includes an operation status parameter that has a high degree of similarity to the operation data. More specifically, the model selection unit 222 may extract from the multiple load estimation models a load estimation model that includes an operation status parameter whose similarity to the operation data is equal to or greater than a predetermined threshold. The predetermined threshold is a similarity threshold (similarity threshold) that serves as a criterion for extracting a load estimation model. The similarity threshold is not particularly limited, and may be set to, for example, 90%. The load estimation unit 223 calculates the load on the work machine 3 by substituting the model parameter θ included in the extracted load estimation model and the operation data into equation (1) or equation (11) described below.

[0087] When each of the operation status parameters of two or more load estimation models has a similarity to the operation data that is equal to or greater than a similarity threshold, the model selection unit 222 may extract these two or more load estimation models and generate a local load estimation model that is a local load estimation model using the extracted two or more load estimation models. In this case, the controller 200 may generate the local load estimation model by weighting the two or more load estimation models in order of magnitude of similarity and adding them together.

[0088] The method for evaluating the similarity is not particularly limited, and various known methods can be used. An example of the similarity evaluation method will be described later.

[0089] Figure 6 is a diagram showing an example of a plurality of load estimation models stored in the load estimation model storage unit 232. The load estimation model storage unit 232 stores a plurality of databases. Each of the plurality of databases includes at least one load estimation model. In the specific example shown in Figure 6, the plurality of databases are databases corresponding to a series of operations performed by the work machine 3 at a work site. The plurality of databases include a first database DB1 for excavation operations, a second database DB2 for lifting and swinging operations, a third database DB3 for earth removal operations, and a fourth database DB4 for return swing operations.

[0090] [Regarding the first database DB1] The first database DB1 includes at least one load estimation model associated with an excavation operation. The first database DB1 may include a plurality of load estimation models associated with an excavation operation (a first load estimation model, a second load estimation model, ..., an n-th load estimation model).

[0091] Specifically, for example, one excavation operation may be divided into n stages performed in this order from a first stage (e.g., the beginning of the excavation operation) to an nth stage (e.g., the end of the excavation operation). The first database DB1 may include a plurality of load estimation models (a first load estimation model, a second load estimation model, ..., an nth load estimation model) respectively associated with the n stages of one excavation operation. In this case, the first load estimation model may include, for example, an operating status parameter P1 corresponding to a sensor value in the first stage of the excavation operation and a model parameter θ1 designed to estimate the load in the first stage of the excavation operation. The second load estimation model may include, for example, an operating status parameter P2 corresponding to a sensor value in the second stage of the excavation operation and a model parameter θ2 designed to estimate the load in the second stage of the excavation operation. The nth load estimation model may include, for example, an operating status parameter Pn corresponding to a sensor value in the nth stage of the excavation operation and a model parameter θn designed to estimate the load in the nth stage of the excavation operation.

[0092] Furthermore, the first load estimation model may include, for example, an operating status parameter P1 corresponding to a sensor value during excavation operation performed under a first condition, and a model parameter θ1 designed to estimate the load during excavation operation performed under the first condition. The second load estimation model may include, for example, an operating status parameter P2 corresponding to a sensor value during excavation operation performed under a second condition, and a model parameter θ2 designed to estimate the load during excavation operation performed under the second condition. The nth load estimation model may include, for example, an operating status parameter Pn corresponding to a sensor value during excavation operation performed under the nth condition, and a model parameter θn designed to estimate the load during excavation operation performed under the nth condition.

[0093] [Regarding the Second Database DB2] The second database DB2 includes at least one load estimation model associated with the lifting and turning motion. The second database DB2 may include multiple load estimation models associated with the lifting and turning motion (first load estimation model, second load estimation model, ..., n-th load estimation model).

[0094] Specifically, for example, one lifting and turning operation may be divided into n stages, which are performed in this order, from a first stage (e.g., the beginning of the lifting and turning operation) to an nth stage (e.g., the end of the lifting and turning operation). The second database DB2 may include multiple load estimation models (a first load estimation model, a second load estimation model, ..., an nth load estimation model) each associated with the n stages of one lifting and turning operation. In this case, the first load estimation model may include an operating status parameter P1 corresponding to a sensor value in the first stage of the lifting and turning operation and a model parameter θ1 designed to estimate the load in the first stage of the lifting and turning operation. The second load estimation model may include, for example, an operating status parameter P2 corresponding to a sensor value in the second stage of the lifting and turning operation and a model parameter θ2 designed to estimate the load in the second stage of the lifting and turning operation. The nth load estimation model may include, for example, an operating status parameter Pn corresponding to a sensor value at the nth stage of a lifting and turning operation, and a model parameter θn designed to estimate the load at the nth stage of a lifting and turning operation.

[0095] The first load estimation model may include, for example, an operating status parameter P1 corresponding to a sensor value in a lifting and turning operation performed under a first condition, and a model parameter θ1 designed to estimate the load in a lifting and turning operation performed under the first condition. The second load estimation model may include, for example, an operating status parameter P2 corresponding to a sensor value in a lifting and turning operation performed under a second condition, and a model parameter θ2 designed to estimate the load in a lifting and turning operation performed under the second condition. The nth load estimation model may include, for example, an operating status parameter Pn corresponding to a sensor value in a lifting and turning operation performed under the nth condition, and a model parameter θn designed to estimate the load in a lifting and turning operation performed under the nth condition.

[0096] [Regarding the third database DB3] The third database DB3 includes at least one load estimation model associated with the earth unloading operation. The third database DB3 may include a plurality of load estimation models associated with the earth unloading operation (a first load estimation model, a second load estimation model, ..., an n-th load estimation model).

[0097] Specifically, for example, one earth unloading operation may be divided into n stages, which are performed in this order from a first stage (e.g., the beginning of the earth unloading operation) to an nth stage (e.g., the end of the earth unloading operation). The third database DB3 may include a plurality of load estimation models (a first load estimation model, a second load estimation model, ..., an nth load estimation model) respectively associated with the n stages of one earth unloading operation. In this case, the first load estimation model may include, for example, an operating status parameter P1 corresponding to a sensor value in the first stage of the earth unloading operation and a model parameter θ1 designed to estimate the load in the first stage of the earth unloading operation. The second load estimation model may include, for example, an operating status parameter P2 corresponding to a sensor value in the second stage of the earth unloading operation and a model parameter θ2 designed to estimate the load in the second stage of the earth unloading operation. The nth load estimation model may include, for example, an operating status parameter Pn corresponding to a sensor value in the nth stage of the earth unloading operation and a model parameter θn designed to estimate the load in the nth stage of the earth unloading operation.

[0098] Furthermore, the first load estimation model may include, for example, an operating status parameter P1 corresponding to a sensor value in an earth unloading operation performed under a first condition, and a model parameter θ1 designed to estimate the load in an earth unloading operation performed under the first condition. The second load estimation model may include, for example, an operating status parameter P2 corresponding to a sensor value in an earth unloading operation performed under a second condition, and a model parameter θ2 designed to estimate the load in an earth unloading operation performed under the second condition. The nth load estimation model may include, for example, an operating status parameter Pn corresponding to a sensor value in an earth unloading operation performed under an nth condition, and a model parameter θn designed to estimate the load in an earth unloading operation performed under the nth condition.

[0099] [Regarding the Fourth Database DB4] The fourth database DB4 includes at least one load estimation model associated with the return turning motion. The fourth database DB4 may include a plurality of load estimation models associated with the return turning motion (first load estimation model, second load estimation model, ..., n-th load estimation model).

[0100] Specifically, for example, one return swing operation may be divided into n stages, which are performed in this order from a first stage (e.g., the beginning of the return swing operation) to an nth stage (e.g., the end of the return swing operation). The fourth database DB4 may include a plurality of load estimation models (a first load estimation model, a second load estimation model, ..., an nth load estimation model) respectively associated with the n stages of one return swing operation. In this case, the first load estimation model may include, for example, an operating status parameter P1 corresponding to a sensor value in the first stage of the return swing operation and a model parameter θ1 designed to estimate the load in the first stage of the return swing operation. The second load estimation model may include, for example, an operating status parameter P2 corresponding to a sensor value in the second stage of the return swing operation and a model parameter θ2 designed to estimate the load in the second stage of the return swing operation. The nth load estimation model may include, for example, an operating status parameter Pn corresponding to a sensor value at the nth stage of the return turning operation, and a model parameter θn designed to estimate the load at the nth stage of the return turning operation.

[0101] Furthermore, the first load estimation model may include, for example, an operating status parameter P1 corresponding to a sensor value in a return swing operation performed under a first condition, and a model parameter θ1 designed to estimate the load in the return swing operation performed under the first condition. The second load estimation model may include, for example, an operating status parameter P2 corresponding to a sensor value in a return swing operation performed under a second condition, and a model parameter θ2 designed to estimate the load in the return swing operation performed under the second condition. The nth load estimation model may include, for example, an operating status parameter Pn corresponding to a sensor value in a return swing operation performed under an nth condition, and a model parameter θn designed to estimate the load in the return swing operation performed under the nth condition.

[0102] Fig. 7 is a diagram showing another example of a plurality of load estimation models stored in the load estimation model storage unit 232. Fig. 7 shows an example of the inside of the first database DB1.

[0103] In the specific example shown in Figure 7, the multiple load estimation models included in the first database DB1 include a first load estimation model, a second load estimation model, and a third load estimation model. The first load estimation model is associated with operations in the early stages of an excavation operation. The second load estimation model is associated with operations in the middle stages of the excavation operation. The third load estimation model is associated with operations in the final stages of the excavation operation. The load on the work machine 3 in the early stages of an excavation operation, the load on the work machine 3 in the middle stages of the excavation operation, and the load on the work machine 3 in the final stages of the excavation operation are not the same but are slightly different. Therefore, by subdividing the classification of the multiple load estimation models as shown in Figure 7, it is possible to estimate the load on the work machine 3 with greater accuracy.

[0104] The first load estimation model includes an operating status parameter P1 corresponding to the sensor value at the beginning of the excavation operation and a model parameter θ1 designed to estimate the load at the beginning of the excavation operation. The second load estimation model includes an operating status parameter P2 corresponding to the sensor value at the middle of the excavation operation and a model parameter θ2 designed to estimate the load at the middle of the excavation operation. The third load estimation model includes an operating status parameter P3 corresponding to the sensor value at the end of the excavation operation and a model parameter θ3 designed to estimate the load at the end of the excavation operation.

[0105] Fig. 8 is a diagram showing yet another example of a plurality of load estimation models stored in the load estimation model storage unit 232. Fig. 8 shows another example of the inside of the first database DB1.

[0106] In the specific example shown in Figure 8, the multiple load estimation models included in the first database DB1 include a fourth load estimation model, a fifth load estimation model, and a sixth load estimation model. The fourth load estimation model is associated with excavation operation performed under fourth conditions. The fifth load estimation model is associated with excavation operation performed under fifth conditions. The sixth load estimation model is associated with excavation operation performed under sixth conditions. The load on the work machine 3 during excavation operation may vary depending on factors such as the operator's proficiency, the environment of the work site, and weather conditions. Therefore, by subdividing the multiple load estimation models as shown in Figure 8, it is possible to estimate the load on the work machine 3 with greater accuracy.

[0107] The fourth load estimation model includes an operating status parameter P4 corresponding to the sensor value during excavation operation performed under the fourth condition, and a model parameter θ4 designed to estimate the load during excavation operation performed under the fourth condition. The fifth load estimation model includes an operating status parameter P5 corresponding to the sensor value during excavation operation performed under the fifth condition, and a model parameter θ5 designed to estimate the load during excavation operation performed under the fifth condition. The sixth load estimation model includes an operating status parameter P6 corresponding to the sensor value during excavation operation performed under the sixth condition, and a model parameter θ6 designed to estimate the load during excavation operation performed under the sixth condition.

[0108] The fourth condition, the fifth condition, and the sixth condition may be conditions related to the proficiency of the operator who operates the work machine 3, for example. In this case, the fourth condition may be that the operator's proficiency is equal to or greater than a predetermined first threshold, the fifth condition may be that the operator's proficiency is smaller than the first threshold and greater than a second threshold, and the sixth condition may be that the operator's proficiency is equal to or less than a predetermined second threshold. The fourth condition, the fifth condition, and the sixth condition may be conditions related to the environment of the work site where work is performed by the work machine 3. The environment of the work site may be, for example, the firmness of the ground. In this case, the fourth condition may be that the firmness of the ground is equal to or greater than a predetermined first threshold, the fifth condition may be that the firmness of the ground is smaller than the first threshold and greater than a second threshold, and the sixth condition may be that the firmness of the ground is equal to or less than a predetermined second threshold. The fourth condition, the fifth condition, and the sixth condition may also be weather conditions at the time when work is performed by the work machine 3.

[0109] The specific examples shown in Figures 6 to 8 have been explained using excavation operations as an example of the operation of the work machine 3, but the specific examples shown in Figures 6 to 8 can also be applied to various operations of the work machine 3 other than excavation operations.

[0110] The load estimation model storage unit 232 may store a plurality of load estimation models constructed by machine learning.

[0111] The communication unit 210 includes an operational data receiving unit 211 and a display information transmitting unit 212 .

[0112] The operation data receiving unit 211 receives operation data of the work machine 3. Specifically, the operation data receiving unit 211 receives operation data transmitted by the work machine 3.

[0113] The display information transmitting unit 212 transmits the display information generated by the display information generating unit 225 to the display device 4 .

[0114] FIG. 9 is a flowchart showing an example of calculation processing performed by the machine controller 100 of the work machine 3.

[0115] First, in step S11, the operation data acquisition unit 102 of the machine controller 100 acquires, as operation data, the sensor values ​​output by the sensors 111 to 116. In addition, the machine controller 100 starts measuring time when starting the calculation process shown in FIG.

[0116] Next, in step S12, the operation data acquisition unit 102 stores the acquired operation data in a storage medium. The storage medium may be provided in the machine controller 100 or in a device separate from the machine controller 100.

[0117] Next, in step S13, the operational data acquiring unit 102 determines whether a predetermined time has elapsed. The start point of measurement of the predetermined time may be, for example, the point at which the calculation process shown in FIG. 9 is started as described above, or may be the point at which the process of step S14 described below is completed (the point at which the elapsed time is reset). The predetermined time may be, for example, a value within a range of 10 minutes to 1 hour.

[0118] If the predetermined time has not elapsed (NO in step S13), the operational data acquiring unit 102 performs the processes from step S11 onward. That is, until the predetermined time has elapsed, the operational data acquiring unit 102 repeatedly acquires operational data (step S11) and stores the repeatedly acquired operational data in the storage medium (step S12). This allows the storage medium to sequentially store operational data that is sequentially acquired at an approximately constant cycle.

[0119] If the predetermined time has elapsed (YES in step S13), the operational data transmission unit 106 of the communication unit 118 transmits the operational data stored in the storage medium to the load estimation device 2 (step S14). Then, the operational data acquisition unit 102 resets the elapsed time for determining whether the predetermined time has elapsed, and performs the processes from step S11 onwards.

[0120] FIG. 10 is a flowchart showing an example of a calculation process performed by the controller 200 of the load estimation device 2.

[0121] The calculation processing shown in Figure 10 includes processing for determining the operation of the work machine 3, processing for selecting a load estimation model, and processing for estimating the load on the work machine 3. It is preferable that the controller 200 performs these processes sequentially while the work machine 3 is operating. This enables the controller 200 to more appropriately estimate the load in accordance with changes in the operation of the work machine 3, even if the operation of the work machine 3 changes frequently. Note that performing the above processes sequentially means continuously repeating the processing of steps S21 to S23 shown in Figure 10, as will be explained below.

[0122] [Determining the Operation of the Work Machine] First, in step S21, the operation determination section 221 of the controller 200 determines the operation of the work machine 3 using the operation data transmitted from the work machine 3.

[0123] Specifically, for example, when the work machine 3 performs a series of operations including an excavation operation, a lifting and swinging operation, an earth-discharging operation, and a return swinging operation, the operation data includes data corresponding to the excavation operation (excavation operation data), data corresponding to the lifting and swinging operation (lifting and swinging operation data), data corresponding to the earth-discharging operation (earth-discharging operation data), and data corresponding to the return swinging operation (return swing operation data). The operation data may be time-series data including sensor values ​​obtained at approximately constant intervals over time.

[0124] The excavation operation, lifting and swinging operation, earth-discharging operation, and return swinging operation are each made up of specific operations of the boom 21, arm 22, bucket 24, and upper swing body 12. Therefore, a correlation exists between each of the excavation operation, lifting and swinging operation, earth-discharging operation, and return swinging operation and the sensor values ​​(operation data) of sensors 111 to 116. Therefore, operation determination unit 221 can determine the operation of work machine 3 using a model and / or conditions set based on the correlation and the operation data.

[0125] Specifically, for example, the operation determination model storage unit 231 may store in advance conditions for determining each of the excavation operation, lifting and swinging operation, soil discharge operation, and return swing operation. The conditions may be conditions that compare operation data including the sensor values ​​with a preset threshold value. In this case, the operation determination unit 221 can determine the operation of the work machine 3 using the operation data and the conditions. Furthermore, the operation determination model storage unit 231 may determine the operation of the work machine 3 using a machine learning model, which will be described later.

[0126] [Load Estimation Model Selection] Next, in step S22, the model selection unit 222 of the controller 200 selects a load estimation model that corresponds to the operation of the work machine 3 determined by the operation determination unit 221 from among the multiple load estimation models stored in the load estimation model storage unit 232. Specifically, the model selection unit 222 may select a database that corresponds to the operation of the work machine 3 determined by the operation determination unit 221 from among the multiple databases, and select a load estimation model that is similar to the operation determined by the operation determination unit 221 from among the multiple load estimation models included in the selected database.

[0127] As described above, each of the multiple load estimation models is stored in the database of the load estimation model storage unit 232 as a set of an operating status parameter representing the operating status and a model parameter θ for estimating the load.

[0128] The model selection unit 222 extracts a load estimation model that includes operating status parameters that have a high degree of similarity to operation data acquired at a certain time t from a plurality of load estimation models included in the database. The model selection unit 222 can then calculate the load y(t) on the work machine 3 by substituting the model parameter θ and operation data included in the extracted load estimation model into equation (1) or equation (11), which will be described later. In other words, the combination of each of the plurality of load estimation models with equation (1) or equation (11) is a model that uses the operation data x i (t) or operation data φ(t) of the work machine 3 as an input value and the load y(t) on the work machine 3 as an output value. Note that equation (1) and equation (11) are each an example of an equation designed in advance based on the correlation between operation data and the load on the work machine 3, and therefore the load estimation device in the present disclosure does not necessarily need to use equation (1) or equation (11), and may calculate the load on the work machine 3 using another relational expression.

[0129] As described above, the operating status parameters are parameters that can be compared with the operating data for similarity. For example, if the operating data includes multiple sensor values ​​detected by the sensors 111 to 116, the operating status parameters may include multiple parameters for comparison with the multiple sensor values.

[0130] Specifically, when the operation of the work machine 3 determined by the operation determination unit 221 is an excavation operation, the model selection unit 222 selects a first database DB1 corresponding to an excavation operation from among a plurality of databases. Then, the model selection unit 222 selects a load estimation model similar to the sensor value included in the operation data from among the plurality of load estimation models included in the first database DB1.

[0131] If the operation of the work machine 3 determined by the operation determination unit 221 is a lifting and swinging operation, the model selection unit 222 selects a second database DB2 corresponding to the lifting and swinging operation from among the multiple databases. Then, the model selection unit 222 selects a load estimation model similar to the sensor value included in the operation data from among the multiple load estimation models included in the second database DB2.

[0132] If the operation of the work machine 3 determined by the operation determination unit 221 is an earth unloading operation, the model selection unit 222 selects a third database DB3 corresponding to an earth unloading operation from among the multiple databases. Then, the model selection unit 222 selects a load estimation model similar to the sensor value included in the operation data from the multiple load estimation models included in the third database DB3.

[0133] If the operation of the work machine 3 determined by the operation determination unit 221 is a return swing operation, the model selection unit 222 selects a fourth database DB4 corresponding to the return swing operation from among the multiple databases. Then, the model selection unit 222 selects a load estimation model similar to the sensor value included in the operation data from among the multiple load estimation models included in the fourth database DB4.

[0134] The model selection unit 222 may select two or more load estimation models from one database and generate a local load estimation model using the two or more load estimation models. A method for generating this local load estimation model will be described later.

[0135] [Estimation of Load on Work Machine] Next, in step S23, the load estimation unit 223 of the controller 200 estimates the load y(t) on the work machine 3 using the model parameter θ and operation data x i (t) included in the load estimation model selected by the model selection unit 222, and the following equation (1). The following equation (1) is a relational expression that is designed in advance using techniques such as experiments, simulations, and machine learning, based on the correlation between the load on the work machine and the operation data.

[0136]

[0137] As described above, the sets of operating status parameters and model parameters θ included in each of the multiple load estimation models are designed in advance to be suitable for the operation of the work machine 3 corresponding to that set, and are stored in the load estimation model storage unit 232. In other words, the sets of operating status parameters and model parameters θ are designed in advance in association with the load acting on the work machine 3 in accordance with the operation of the work machine 3 corresponding to that set, and are stored in the load estimation model storage unit 232.

[0138] The above are the main features of the determination of work machine operation, selection of a load estimation model, and estimation of the load on the work machine 3 performed by the controller 200 of the load estimation device 2 according to this embodiment. Below, more preferred specific examples of the determination of work machine operation, selection of a load estimation model, and estimation of the load on the work machine 3 will be described.

[0139] [Determining movement using a machine learning model] Figure 11 is a diagram showing an example of processing for determining the movement of the work machine 3 using operation data and a machine learning model of the work machine 3. As shown in Figure 11, the movement determination unit 221 of the controller 200 may use operation data and a machine learning model to determine the movement of the work machine 3. In this case, the movement determination model storage unit 231 stores the movement determination model N (machine learned model) of the neural network shown in Figure 11.

[0140] The input xi(t) of the operation determination model N is operation data. The output value sk(t) of the operation determination model N is an output value related to the operation of the work machine 3.

[0141] The structure of the motion determination model N is not particularly limited as long as it satisfies the above-mentioned relationship between the input and output of data, and various neural network model structures can be adopted.

[0142] 11, the motion determination model N includes an LSTM layer, a fully connected layer, and a softmax function (LSTM: Long Short Term Memory). The LSTM layer is a type of recurrent neural network (RNN: Recurrent Neural Network) suitable for time-series data that changes over time. The softmax function is expressed by, for example, the following equation:

[0143]

[0144] The output values ​​sk(t) of the motion determination model N are output values ​​corresponding to a plurality of predetermined motions, and these output values ​​correspond to normalized probability values, and the sum of these output values ​​is 1. The plurality of motions includes, for example, an excavation motion, a lifting and swinging motion, an earth-discharging motion, and a return and swinging motion.

[0145] The operation determination unit 221 uses the operation determination model N to perform calculations in the LSTM layer and in the fully connected layer on the operation data (e.g., time series data) of the work machine 3, and normalizes and outputs output values ​​corresponding to the multiple operations using a softmax function.

[0146] Fig. 12 is a diagram showing an example of the output value sk(t) of the movement determination unit 221. As shown in Fig. 12, the output value sk(t) of the movement determination unit 221 is a probability value (predicted probability) for each of the plurality of movements.

[0147] 12, the operation data includes data (e.g., sensor values ​​of sensors 111 to 116) acquired at predetermined time intervals in a time range from time t1 to time tn. In this case, the operation determination unit 221 may perform an operation in the LSTM layer and an operation in the fully connected layer for each of the data at time t1, the data at time t2, ..., and the data at time tn, and may normalize and output output values ​​corresponding to the plurality of operations using a softmax function.

[0148] Specifically, the movement determination unit 221 performs a calculation using the movement determination model N on data at time t1 included in the operation data, and outputs the calculation results "digging operation: 75%," "lifting and swinging operation: 10%," "earth discharge operation: 5%," and "return swing operation: 10%" as output values ​​sk(t). The movement determination unit 221 also performs a calculation using the movement determination model N on data at time t2 included in the operation data, and outputs the calculation results "digging operation: 70%," "lifting and swinging operation: 25%," "earth discharge operation: 3%," and "return swing operation: 2%" as output values ​​sk(t). Similarly, the movement determination unit 221 performs a calculation using the movement determination model N on data at time tn included in the operation data, and outputs the calculation results "digging operation: 5%," "lifting and swinging operation: 15%," "earth discharge operation: 10%," and "return swing operation: 70%" as output values ​​sk(t).

[0149] For example, the movement determination unit 221 may determine, for each of times t1 to tn, the movement corresponding to the highest probability value among the probability values ​​for each of the plurality of movements as the movement of the work machine 3. In this case, the movement determination unit 221 determines that the movement of the work machine 3 at time t1 is an "digging movement", determines that the movement of the work machine 3 at time t2 is an "digging movement", and determines that the movement of the work machine 3 at time tn is a "return swing movement".

[0150] For example, for each of times t1 to tn, the movement determination unit 221 may determine, as the movement of the work machine 3, a movement corresponding to a probability value equal to or greater than a predetermined threshold value among the probability values ​​for each of the plurality of movements. For example, if the predetermined threshold is set to 20%, the movement determination unit 221 may determine that the movement of the work machine 3 at time t1 is an "digging movement", and that the movement of the work machine 3 at time tn is a "return swing movement". Furthermore, at time t2, the probability value corresponding to the excavation movement and the probability value corresponding to the lifting swing movement are equal to or greater than the predetermined threshold value, so the movement determination unit 221 may determine that the movement of the work machine 3 at time t2 is an "digging movement" and a "lifting swing movement". In this case, the load estimator 223 may estimate the load on the work machine 3 by weighting the probability values ​​according to the magnitude of the probability values. This load estimation method will be described later.

[0151] [Database Selection] Next, database selection will be described with reference to Fig. 13.

[0152] As described above, the load estimation model storage unit 232 may store a plurality of databases corresponding to a plurality of operations of the work machine 3.

[0153] 13 , the load estimation model storage unit 232 stores a database DB1 related to excavation operations, a database DB2 related to lifting and swinging operations, a database DB3 related to earth-discharging operations, and a database DB4 related to return swing operations. Database DB1 includes one load estimation model or multiple load estimation models related to excavation operations. Database DB2 includes one load estimation model or multiple load estimation models related to lifting and swinging operations. Database DB3 includes one load estimation model or multiple load estimation models related to earth-discharging operations. Database DB4 includes one load estimation model or multiple load estimation models related to return swing operations.

[0154] In the following, a case will be described in which each of the databases DB1, DB2, DB3, and DB4 includes a plurality of load estimation models.

[0155] First, the model selection unit 222 selects at least one database from the plurality of databases DB1, DB2, DB3, and DB4 that is similar to the determination result of the motion determination unit 221. Specifically, the model selection unit 222 may select a database using the output value sk(t) of the motion determination model N of the motion determination unit 221 and the condition of the following equation (3).

[0156]

[0157] In equation (3), "n" is the number of databases. In the case of FIG. 13 , the number of databases is four, so the value of "n" is four. For example, in the specific example shown in FIG. 12 , of the output values ​​sk(t) of the operation determination model N at time t1, the one that satisfies the condition of equation (3) above is "digging operation." Therefore, the model selection unit 222 selects database DB1 corresponding to the excavation operation from among the multiple databases DB1, DB2, DB3, and DB4, as the database that is similar to the determination result of the operation determination unit 221. Then, the model selection unit 222 extracts, from the multiple load estimation models included in database DB1, a load estimation model that includes operating status parameters that have a high degree of similarity to the operation data at time t1. Then, the load estimation unit 223 estimates the load y(t) on the work machine 3 using the model parameter θ and operation data x(t) included in the load estimation model selected by the model selection unit 222, and equation (1) above.

[0158] 12, the output values ​​sk(t) of the motion determination model N at time t2 are "digging motion" and "lifting and turning motion," which satisfy the condition of the above formula (3). Therefore, the model selection unit 222 selects, from the multiple databases DB1, DB2, DB3, and DB4, the database DB1 corresponding to the excavating motion and the database DB2 corresponding to the lifting and turning motion as databases similar to the determination result of the motion determination unit 221.

[0159] The model selection unit 222 then extracts a load estimation model including an operational status parameter that has a high degree of similarity to the operational data at time t2 from among the multiple load estimation models included in the database DB1. The load estimation unit 223 then estimates the load yDB1 on the work machine 3 using the model parameter θ and operational data x(t) included in the load estimation model selected by the model selection unit 222, and the above equation (1).

[0160] Similarly, the model selection unit 222 extracts a load estimation model including an operational status parameter that has a high degree of similarity to the operation data at time t2 from the plurality of load estimation models included in the database DB2. Then, the load estimation unit 223 estimates the load yDB2 on the work machine 3 using the model parameter θ and operation data x(t) included in the load estimation model selected by the model selection unit 222, and the above formula (1).

[0161] The load estimator 223 then calculates the load y(t) on the work machine 3 by substituting the load yDB1 estimated using database DB1, the load yDB2 estimated using database DB2, and the output value sk(t) of the operation determination model N at time t2 into the following equation (4). This equation (4) sums up values ​​weighted in descending order of probability value (prediction probability).

[0162]

[0163] In equation (4), s1(t) is the output corresponding to the excavation operation (70% in FIG. 12) among the output values ​​sk(t) of the motion determination model N at time t2, and s2(t) is the output corresponding to the lifting and turning operation (25% in FIG. 12) among the output values ​​sk(t) of the motion determination model N at time t2.

[0164] [Local Load Estimation Model] FIG. 14 is a diagram for explaining the local load estimation model.

[0165] In the load estimation device 2, when two or more load estimation models are selected from one database by the model selection unit 222, the load estimation unit 223 may generate a local load estimation model using the two or more load estimation models, and estimate the load on the work machine 3 using the local load estimation model and operation data. Even when there are two or more candidate load estimation models that are similar to the operation of the work machine 3, the controller 200 can estimate the load on the work machine 3. This improves the accuracy of load estimation.

[0166] The multiple databases correspond to multiple operations, and each of the multiple load estimation models included in each database includes an operation status parameter that represents an operation status and a model parameter θ for estimating a load, as described above. The model selection unit 222 selects a database that corresponds to the operation determined by the operation determination unit 221 from the multiple databases, and selects a load estimation model that includes an operation status parameter that has a high similarity to the operation data from the multiple load estimation models included in the selected database.

[0167] Hereinafter, a specific description will be given of how to generate a local load estimation model when the load estimation model storage unit 232 stores a plurality of databases.

[0168] First, the model selection unit 222 selects a database from among the multiple databases DB1, DB2, DB3, and DB4 using the output value sk(t) of the motion determination model N of the motion determination unit 221 and the condition of the above equation (3). The specific example in Fig. 14 shows the case where the model selection unit 222 selects the database DB1 corresponding to the excavation motion.

[0169] This database DB1 stores in advance a plurality of load estimation models related to excavation operations, as shown in, for example, Figures 7 and 8. The small black circles drawn inside the database DB1 in Figure 14 represent the plurality of load estimation models related to excavation operations. Furthermore, the star symbols drawn inside the database DB1 in Figure 14 indicate operation data at a certain time. The distance between the star symbols and each of the plurality of black circles represents the degree of similarity of the operation status parameters of the load estimation model with respect to the operation data. The circle drawn inside the database DB1 in Figure 14 indicates a similarity threshold (similarity threshold, for example, 90%) that serves as a criterion for extracting a load estimation model. Load estimation models in the area inside this circle have a similarity equal to or greater than the similarity threshold.

[0170] 14 , each of the operational status parameters of the two load estimation models has a similarity to the operational data that is equal to or greater than a similarity threshold. In this case, the model selection unit 222 extracts these two load estimation models and generates a local load estimation model, which is a local load estimation model, using the two extracted load estimation models.

[0171] Specifically, the model selection unit 222 may generate the model parameter θ in the local load estimation model by weighting and adding together the model parameter θ included in one of the two extracted load estimation models and the model parameter θ included in the other using the following equation (5).

[0172]

[0173] In equation (5), "wi" is expressed by the following equation (6).

[0174]

[0175] In equation (6), "S(φ(t), φ(i))" is the similarity expressed by the following equation (7).

[0176]

[0177] In equation (7), "φi(t)" is an information vector at time t, and is expressed by the following equation (8).

[0178] In addition, in equation (7), "φi(j)" is an information vector stored in the database.

[0179] Also, in equation (7), "hi" is the bandwidth of the Gaussian kernel for the information vector.

[0180]

[0181] In equation (8), “xi(t)” is the operational data at time t.

[0182] In addition, in equation (8), "y(t-1)" is the load estimated in the previous step (time (t-1)).

[0183] When the difference between the operational data and the operational status parameters in the database is zero, the similarity indicated by equation (7) is expressed as in equation (9) below.

[0184]

[0185] The conditional expression based on the similarity is expressed by, for example, the following expression (10).

[0186]

[0187] The load estimation unit 223 can then calculate the load y(t) of the work machine 3 by substituting the model parameter θ of the local load estimation model generated as described above and the operation data φ(t) into the following equation (11):

[0188]

[0189] [Modifications] Although the embodiments of the present disclosure have been described above, the present disclosure is not limited to the above-described embodiments and includes, for example, the following modifications.

[0190] (A) Regarding the Load Estimation Device In the above embodiment, the load estimation device 2 is a device such as a server that is located at a location remote from the work machine 3, but the load estimation device in the present disclosure may also be mounted on the work machine.

[0191] (B) Regarding the Load Estimation Model Storage Unit In the above embodiment, the load estimation model storage unit is mounted on the load estimation device 2, but the load estimation model storage unit in the present disclosure may be mounted on a device separate from the load estimation device. The separate device may be a work machine or a storage device for storing data.

[0192] (C) Regarding the multiple operations of the work machine In the above embodiment, the multiple operations of the work machine include an excavation operation, a lifting and swinging operation, an earth-discharging operation, and a return swinging operation, but the multiple operations in the present disclosure do not have to include some or all of the excavation operation, lifting and swinging operation, earth-discharging operation, and return swinging operation, and may include operations other than the excavation operation, lifting and swinging operation, earth-discharging operation, and return swinging operation.

[0193] (D) Determining the Operation of a Work Machine The operation determination model may be a machine-learned model that uses operation data as an input value and one operation as an output value.

[0194] The movement determination unit 221 may determine the movement of the work machine 3 using judgment conditions that are set in advance based on the correlation between the movement of the work machine 3 and the operation data of the work machine 3 and actual operation data. The movement determination model storage unit 231 may store these judgment conditions. Specifically, the judgment conditions may include an excavation movement judgment condition for determining an excavation movement, a lifting and swing movement judgment condition for determining a lifting and swing movement, an earth dumping movement judgment condition for determining an earth dumping movement, and a return swing movement judgment condition for determining a return swing movement. In this case, the movement determination unit 221 may determine that the movement of the work machine 3 is an excavation movement if the excavation movement judgment condition is satisfied, determine that the movement of the work machine 3 is a lifting and swing movement if the lifting and swing movement judgment condition is satisfied, determine that the movement of the work machine 3 is an earth dumping movement if the earth dumping movement judgment condition is satisfied, and determine that the movement of the work machine 3 is a return swing movement if the return swing movement judgment condition is satisfied.

[0195] (E) Regarding the Output Device In the above embodiment, the display device 4 is exemplified as the output device, but the output device in the present disclosure is not limited to the display device 4 and may be, for example, a device capable of outputting information about the load on the work machine using sound or light.

[0196] (F) Regarding the Operation Device In the above embodiment, the operation device 117 operates the solenoid valves 31 to 35 of the hydraulic circuit 119 via the machine controller 100. However, the present disclosure is not limited to this. The operation device 117 may include a remote control valve, which is a hydraulic device that outputs pressure according to the amount of lever operation. In this case, the command unit 103 and the solenoid valves 31 to 35 are not required, and the boom pilot pressure, arm pilot pressure, bucket pilot pressure, swing pilot pressure, and travel pilot pressure output from the operation device 117 are input to the control valves 36, 37, 38, 39, 40L, and 40R, respectively. Pressure oil is supplied to the operation device 117 from a pilot pump. The operation device 117 reduces the pressure of the supplied pressure oil to a pressure according to the amount of lever operation and outputs it as pilot pressure to the control valves 36, 37, 38, 39, 40L, and 40R. Furthermore, a pressure sensor is installed in the hydraulic piping connecting the operation device 117 and the control valves 36, 37, 38, 39, 40L, and 40R. The pressure sensor detects the pressure value of the pilot pressure output from the operation device 117 to the control valves 36, 37, 38, 39, 40L, and 40R, and inputs a signal of the detected pressure value to the machine controller 100. The machine controller 100 handles the pressure value signals input from the pressure sensor as operation command signals (boom command signal, arm command signal, bucket command signal, swing command signal, and travel command signal).

[0197] (G) Regarding Movable Parts In the above embodiment, the boom 21, arm 22, bucket 24, and upper rotating body 12 are given as examples of multiple movable parts, but the movable parts in the present disclosure are not limited to these and may include, for example, the undercarriage 10. In this case, the operation data of the work machine 3 may include sensor values ​​that detect physical quantities that change in accordance with the operation of the undercarriage 10. In this case, for example, the operation data acquisition unit 102 acquires operation data that includes sensor values ​​from a sensor that detects the operating pressure value of the travel motor of the work machine 3. Furthermore, the operation determination model storage unit 231 may store an operation determination model constructed by machine learning, with operation data including the operating pressure value of the travel motor as input values ​​and the operation of the work machine 3 as output values. This makes it possible to calculate the load on the work machine 3, including the undercarriage 10.

[0198] As described above, the present disclosure provides a technique that can estimate the load on a work machine according to the type of operation.

[0199] The load estimation device according to the first aspect includes a controller, and the controller determines the operation of the work machine using operation data of the work machine, selects at least one load estimation model corresponding to the determined operation of the work machine from among a plurality of load estimation models respectively associated with a plurality of operations of the work machine, and estimates the load on the work machine using the selected at least one load estimation model and the operation data.

[0200] In the load estimation device according to this first aspect, the controller determines the operation of the work machine, selects a load estimation model in accordance with the determined operation, and estimates the load using the selected load estimation model and operation data, so that the load on the work machine can be estimated in accordance with the type of operation of the work machine.

[0201] A load estimation device according to a second aspect preferably comprises the following configuration in addition to the load estimation device according to the first aspect. That is, in the load estimation device according to the second aspect, it is preferable that the at least one load estimation model includes two or more load estimation models, and the controller generates a local load estimation model using the two or more load estimation models, and estimates the load on the work machine using the local load estimation model and the operation data. In this second aspect, even if there are two or more candidate load estimation models that are similar to the operation of the work machine, the controller can generate a local load estimation model using the two or more load estimation models, and estimate the load on the work machine using the generated local load estimation model and the operation data. This improves the accuracy of load estimation.

[0202] This second aspect may have the following configuration. That is, each of the plurality of load estimation models may include an operation status parameter and a model parameter for estimating a load. The controller may select a load estimation model including an operation status parameter having a high similarity to the operation data from the plurality of load estimation models. When two or more load estimation models are selected, the controller may generate the local load estimation model by weighting the two or more load estimation models in order of magnitude of similarity and adding them together.

[0203] A load estimation device according to a third aspect preferably includes the following configuration in addition to the load estimation device according to the first or second aspect. That is, in the load estimation device according to the third aspect, it is preferable that the work machine includes a plurality of actuators and a plurality of movable parts operated by the plurality of actuators, and the operation data includes data related to the operation of the plurality of movable parts. In this third aspect, the controller can estimate the load caused by the operation of the movable parts of the work machine.

[0204] A load estimation device according to a fourth aspect is preferably the load estimation device according to any one of the first to third aspects, further comprising the following configuration. That is, in the load estimation device according to the fourth aspect, it is preferable that the controller sequentially determines the operation of the work machine, selects the at least one load estimation model, and estimates the load on the work machine while the work machine is operating. In this fourth aspect, even if the operation of the work machine changes frequently, the controller can more appropriately estimate the load in accordance with the changes in the operation.

[0205] A load estimation system according to a fifth aspect includes an operation data acquisition unit that acquires operation data of a work machine, and the load estimation device according to any one of the first to fourth aspects.

[0206] It is preferable that the load estimation system according to the sixth aspect further comprises the following configuration in addition to the load estimation system according to the fifth aspect: That is, in the load estimation system according to the sixth aspect, a load estimation model storage unit that stores a plurality of load estimation models respectively associated with a plurality of operations of the work machine may be mounted on the load estimation device or on equipment separate from the load estimation device.

[0207] It is preferable that the load estimation system according to the seventh aspect further comprises the following configuration in addition to the load estimation system according to the fifth or sixth aspect. That is, the load estimation system according to the seventh aspect preferably comprises an output device that outputs information about the estimated load on the work machine. In this seventh aspect, operators, managers, and other related parties in the work can understand the load on the work machine based on the information output by the output device.

[0208] A load estimation method according to an eighth aspect includes determining the operation of a work machine using operation data of the work machine, selecting at least one load estimation model corresponding to the determined operation of the work machine from a plurality of load estimation models respectively associated with a plurality of operations of the work machine, and estimating the load on the work machine using the at least one selected load estimation model and the operation data. The load estimation method according to the eighth aspect determines the operation of the work machine, selects a load estimation model corresponding to the determined operation, and estimates the load using the selected load estimation model and the operation data, so that it is possible to estimate the load on the work machine according to the type of operation of the work machine. The load estimation method according to the eighth aspect may further include acquiring operation data of the work machine.

[0209] A load estimation program according to a ninth aspect causes a computer to execute the following processes: determining the operation of the work machine using operation data of the work machine; selecting at least one load estimation model corresponding to the determined operation of the work machine from a plurality of load estimation models respectively associated with a plurality of operations of the work machine; and estimating the load on the work machine using the at least one selected load estimation model and the operation data. The load estimation program according to the ninth aspect causes a computer to execute the processes of determining the operation of the work machine, selecting a load estimation model corresponding to the determined operation, and estimating the load using the selected load estimation model and the operation data, so that the load on the work machine can be estimated according to the type of operation of the work machine. The load estimation program according to the ninth aspect may further cause the computer to execute the process of acquiring the operation data of the work machine. The load estimation program may be stored in a non-transitory storage medium.

Claims

1. A load estimation device comprising a controller, wherein the controller determines the operation of a work machine using operation data of the work machine, selects at least one load estimation model corresponding to the determined operation of the work machine from a plurality of load estimation models respectively associated with a plurality of operations of the work machine, and estimates the load on the work machine using the at least one selected load estimation model and the operation data.

2. A load estimation device according to claim 1, wherein the at least one load estimation model includes two or more load estimation models, and the controller generates a local load estimation model using the two or more load estimation models, and estimates the load on the work machine using the local load estimation model and the operation data.

3. A load estimation device according to claim 1 or 2, wherein the work machine comprises a plurality of actuators and a plurality of moving parts operated by the plurality of actuators, and the operation data includes data relating to the operation of the plurality of moving parts.

4. A load estimation device according to any one of claims 1 to 3, wherein the controller sequentially determines the operation of the work machine, selects the at least one load estimation model, and estimates the load on the work machine while the work machine is operating.

5. A load estimation system comprising: an operation data acquisition unit that acquires operation data of a work machine; and a load estimation device according to any one of claims 1 to 4.

6. A load estimation system according to claim 5, wherein a load estimation model storage unit that stores a plurality of load estimation models respectively associated with a plurality of operations of the work machine is mounted on the load estimation device or on equipment separate from the load estimation device.

7. A load estimation system according to claim 5 or 6, further comprising an output device that outputs information about the estimated load on the work machine.

8. A load estimation method comprising: determining the operation of a work machine using operation data of the work machine; selecting at least one load estimation model corresponding to the determined operation of the work machine from a plurality of load estimation models respectively associated with a plurality of operations of the work machine; and estimating the load on the work machine using the selected at least one load estimation model and the operation data.

9. A load estimation program that causes a computer to execute the following processes: a process of determining the operation of a work machine using operation data of the work machine; a process of selecting at least one load estimation model that corresponds to the determined operation of the work machine from a plurality of load estimation models that are respectively associated with a plurality of operations of the work machine; and a process of estimating the load on the work machine using the at least one selected load estimation model and the operation data.

Images