Information processing apparatus, information processing method, and program

Abstract

To obtain a sufficient amount of learning data regardless of conditions such as the size and excavation capacity of a shield excavator.SOLUTION: A machine condition acquisition unit configured to acquire a machine condition of a shield machine in which an operation corresponding to the operation data has been performed, a standardized data generation unit configured to calculate standardized data which is operation data independent of the machine condition by standardizing the operation data with the machine condition, and a learning data generation unit configured to generate learning data corresponding to an operation of the shield machine to be estimated by performing an operation reverse to an operation used for standardization by the standardized data generation unit on the standardized data.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an information processing device, an information processing method, and a program. [Background technology]

[0002] In excavation using a shield machine, there is a technology that uses AI (machine learning) to predict the operation of the propulsion jack that determines the excavation direction. For example, Patent Document 1 discloses a technology that estimates the position coordinate of the force point by using an estimation model generated by machine learning learning data that associates directional data such as indicated values ​​and current values ​​with the position coordinate of the force point to be excavated. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2019-143385 Summary of the Invention [Problem to be solved by the invention]

[0004] Shield tunneling machines vary in size and excavation capacity (machine conditions, described below). The size and excavation capacity of a shield tunneling machine must be selected based on the geological and groundwater conditions of the excavated ground, so that the face can be stabilized. In addition to these conditions, the size and size of the excavation surface, economic efficiency, etc. are also considered, and the size and excavation capacity of the shield tunneling machine are determined from a comprehensive perspective.

[0005] In generating the above-mentioned estimation model, in order to ensure estimation accuracy, past operation data of shield tunneling machines of the same size and excavation capacity, etc., was used for training. Therefore, at a site where excavation is newly started, if the machine conditions are different from those of shield tunneling machines used at other sites, it may be difficult to obtain a sufficient amount of training data. Generally, when the amount of training data is insufficient, a phenomenon called overfitting becomes more pronounced, resulting in overfitting to previously trained data while being unable to fit to unknown data, resulting in reduced estimation accuracy. Therefore, depending on the conditions of the shield tunneling machine, such as its size and excavation capacity, there is a problem that the amount of training data is insufficient, resulting in reduced estimation accuracy of the estimation model.

[0006] In view of the above-mentioned problems, the present invention aims to provide an information processing device, an information processing method, and a program that can obtain a sufficient amount of learning data regardless of conditions such as the size and excavation capacity of the shield excavator. [Means for solving the problem]

[0007] An information processing device according to one embodiment of the present invention comprises an operation data acquisition unit that acquires operation data indicating the operating history of a shield excavation machine; a machine condition acquisition unit that acquires the machine conditions of the shield excavation machine in which the operation corresponding to the operation data was performed; a standardized data generation unit that calculates standardized data, which is operation data that does not depend on the machine conditions, by standardizing the operation data using the machine conditions; and a learning data generation unit that generates learning data corresponding to the operation of the shield excavation machine to be estimated by performing an operation that is the inverse of the operation used for standardization by the standardized data generation unit using the machine conditions of the shield excavation machine to be estimated.

[0008] An information processing method according to one aspect of the present invention is an information processing method performed by a computer, in which an operation data acquisition unit acquires operation data indicating the operating history of a shield excavation machine, a machine condition acquisition unit acquires the machine conditions of the shield excavation machine on which the operation corresponding to the operation data was performed, a standardization data generation unit standardizes the operation data using the machine conditions to calculate standardization data, which is operation data that does not depend on the machine conditions, and a learning data generation unit generates learning data corresponding to the operation of the shield excavation machine to be estimated by performing an operation that is the inverse of the operation used for standardization by the standardization data generation unit using the machine conditions of the shield excavation machine to be estimated.

[0009] A program according to one embodiment of the present invention causes a computer to acquire operation data indicating the operating history of a shield excavation machine, acquire the machine conditions of the shield excavation machine in which the operation corresponding to the operation data was performed, calculate standardized data, which is operation data that is independent of the machine conditions, by standardizing the operation data using the machine conditions of the shield excavation machine to be estimated, and generate learning data corresponding to the operation of the shield excavation machine to be estimated by performing an operation on the standardized data that is the inverse of the operation used for the standardization using the machine conditions of the shield excavation machine to be estimated. [Effects of the Invention]

[0010] According to the present invention, a sufficient amount of learning data can be obtained regardless of conditions such as the size and excavation capacity of the shield excavator. [Brief explanation of the drawings]

[0011] [Figure 1A] 1 is a diagram for explaining a shield boring machine 10 according to an embodiment. [Figure 1B] 1 is a diagram for explaining a shield boring machine 10 according to an embodiment. [Figure 2] 1 is a block diagram showing a configuration of an information processing device 30 according to an embodiment. [Figure 3]2 is a diagram for explaining a process performed by an information processing device 30 according to an embodiment. FIG. [Figure 4] 5 is a diagram showing an example of information stored in an operation data storage unit 34 according to the embodiment. FIG. [Figure 5] FIG. 4 is a diagram illustrating an example of information stored in a machine condition storage unit 35 according to the embodiment. [Figure 6] FIG. 4 is a diagram showing an example of information stored in a standardized data storage unit 36 ​​according to the embodiment. [Figure 7] 10 is a flowchart showing a flow of processing performed by an information processing device 30 according to an embodiment. [Figure 8] FIG. 10 is a diagram for explaining the effects of the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0013] <About the Shield Drilling Machine 10> First, a shield tunneling machine 10 according to an embodiment will be described with reference to Fig. 1 (Figs. 1A and 1B). Fig. 1 is a diagram for explaining the shield tunneling machine 10 according to an embodiment.

[0014] FIG. 1A shows a conceptual side view of a shield excavator 10. As shown in FIG. 1A, the shield excavator 10 excavates the ground while assembling segment rings using an erector (not shown) behind a cylindrical skin plate 11 and constructing a primary lining SG. The shield excavator 10 has a chamber 12 at the rear of a circular, face-plate-shaped cutter head 15 equipped with a cutter bit 16. Multiple earth pressure gauges D are installed on the side walls of the chamber 12. The earth pressure gauges D measure the pressure (controlled earth pressure) of the mud in the chamber 12. An additive 14 is injected into the chamber 12 through a mud-making material injection pipe 13. The excavated soil accumulated in the chamber 12 is mixed with the additive 14 by a mixing blade (not shown) and converted into mud. A screw conveyor 17 discharges the mud from the chamber 12 onto a conveyor 18 through a soil discharge gate G. The conveyor 18 carries the mud discharged from the screw conveyor 17 out of the tunnel via a conveyor 19. A platform M supports the screw conveyor 17 and the conveyors 18 and 19.

[0015] FIG. 1B shows a conceptual diagram illustrating the propulsion jacks that propel the shield excavator 10. As shown in FIG. 1B, the shield excavator 10 is provided with a plurality of propulsion jacks 20 (propulsion jacks 20-1 to 20-12). The propulsion jacks 20 are provided so as to surround the inner periphery of the skin plate 11. The propulsion jacks 20 are disposed between the skin plate 11 and the segment ring. The propulsion direction and excavation speed of the shield excavator 10 are controlled by extending the propulsion jacks 20 using hydraulic control or the like. In the example shown in this figure, 12 propulsion jacks 20 are provided in the shield tunneling machine 10, but this number is not limited to this. The number, arrangement, and thrust of the shield jacks are determined taking into consideration, for example, the outer diameter of the shield, the total propulsion force, the segment structure, the linearity of the tunnel, and propulsion by a jack on one side when constructing a curve or correcting meandering.

[0016] The position of the point of force that the shield excavator 10 applies to the excavation surface is set depending on which of the propulsion jacks 20 is extended. The propulsion direction of the shield excavator 10 is controlled by applying propulsion pressure to a position corresponding to the force point set in response to the operation of extending the propulsion jack 20. For example, the propulsion direction is controlled by a jack pattern that indicates which of the propulsion jacks 20 is extended.

[0017] The operator controls the direction of advancement of the shield tunneling machine 10 by operating the propulsion jacks 20 while monitoring data from various measuring instruments in accordance with instructions from staff members or the like (or instructions prepared in advance). The operator controls the direction of advancement of the shield tunneling machine 10 by selecting whether or not to use each of the propulsion jacks 20 so that the current values ​​determined based on data from various measuring instruments approach the values ​​specified in the instructions (instructed values).

[0018] The operator also controls the balance between the face pressure and the mud pressure (controlled earth pressure) in the chamber 12 so that the excavation face (face) is kept stable. For example, the operator controls the controlled earth pressure by controlling the amount of earth discharged and excavated, as well as the cutter torque and jack thrust, and propels the shield excavator 10 forward while keeping the face stable.

[0019] In this way, the operator controls the thrust direction to follow the tunnel planning line and controls the thrust force to maintain the stability of the face, and proceeds with excavation. In this embodiment, operation data showing such past operations by the operator is collected and trained into an AI (machine learning) learning model, thereby generating an estimation model for estimating the operation of the shield boring machine 10.

[0020] In the past, estimation models were generated using the operation data of a shield tunneling machine with the same machine conditions as training data, which resulted in a problem of insufficient training data depending on the machine conditions, resulting in a decrease in the estimation accuracy of the estimation model.

[0021] The machine conditions here refer to conditions related to the size and excavation capacity of the shield machine, such as machine diameter, machine length, equipped cutter torque, and equipped total thrust. The machine diameter is the outer diameter of the circle corresponding to the bottom of a cylindrical shield excavation machine. The machine length is the height of the cylinder in a cylindrical shield excavation machine. The equipped cutter torque is the upper limit of the torque that can be applied to the cutter head 15 by the electric motor (motor) equipped to rotate the cutter head 15 of the shield excavation machine. The equipped total thrust is the sum of the thrust in the shield excavation machine, which is the sum of the thrust from all the propulsion jacks 20 equipped on the shield excavation machine.

[0022] To address this issue, in this embodiment, the operation data of shield excavators with different machine conditions can be used for learning. Specifically, the operation data is converted into standardized data that is independent of the machine conditions. In other words, the standardized data is operation data that is independent of the machine conditions. The standardized data is then used to generate learning data that corresponds to the machine conditions of the shield excavator that is the target of the estimation model. This makes it possible to use the operation data of shield excavators with different machine conditions as learning data, and a sufficient amount of learning data can be obtained regardless of the machine conditions.

[0023] In the following, it is assumed that the information processing device 30 converts the operation data into standardized data, and the configuration of the information processing device 30 will be described in detail.

[0024] <Regarding the information processing device 30> Fig. 2 is a block diagram showing the configuration of an information processing device 30 according to an embodiment. As shown in Fig. 2, the information processing device 30 includes, for example, an operation data acquisition unit 31, a machine condition acquisition unit 32, a standardized data generation unit 33, an operation data storage unit 34, a machine condition storage unit 35, a standardized data storage unit 36, a learning data generation unit 37, and an estimation model generation unit 38.

[0025] The operation data acquisition unit 31 acquires operation data. The operation data is data indicating the operation results performed by the shield tunneling machine, and includes, for example, current values ​​and command values. The current values ​​are the values ​​of the time-series changes of the operations performed by the operator. The command values ​​are the command values ​​corresponding to the operation of the current values.

[0026] The operation data acquisition unit 31 acquires operation data, such as operation data relating to the operation of the propulsion direction and operation data relating to the operation of the propulsion force.

[0027] The operation data acquisition unit 31 acquires information indicating time-series changes in the stroke difference of the jack selected by the operation of the propulsion jack 20, for example, the operator, as the operation in the propulsion direction.

[0028] Based on the acquired operation data, the operation data acquisition unit 31 calculates the current value of the left-right jack stroke difference and the current value of the up-down jack stroke difference as operation indicators indicating the propulsion direction. The current value of the left-right jack stroke difference is the component corresponding to the left-right direction with respect to the direction of travel of the current value of the extension difference of the propulsion jack 20 that occurs as a result of extending the propulsion jack 20. The up-down jack stroke difference is the component corresponding to the up-down direction with respect to the direction of travel of the current value of the extension difference of the propulsion jack 20 that occurs as a result of extending the propulsion jack 20.

[0029] For example, suppose that propulsion jacks 20-1, 20-4, and 20-9 are selected from the propulsion jacks 20 shown in Figure 1B and these selected propulsion jacks 20 are extended. In this case, the operation data acquisition unit 31 calculates the current value of the left-right jack stroke difference LR_JS using the following equation (1). Also, the operation data acquisition unit 31 calculates the current value of the upper-lower jack stroke difference UD_JS using the following equation (2).

[0030] LR_JS=J9-J4...Formula (1) UD_JS=2×{J1-(J4+J9) / 2}…Formula (2) however, LR_JS is a value that indicates the current value of the difference between the left and right jack strokes. UD_JS is a value that indicates the current value of the upper and lower jack stroke difference. J1 is a value indicating the current value of the stroke of the propulsion jack 20-1. J4 is a value indicating the current value of the stroke of the propulsion jack 20-4. J9 is a value indicating the current value of the stroke of the propulsion jack 20-9.

[0031] Furthermore, the operation data acquisition unit 31 acquires the command values. The operation data acquisition unit 31 acquires the command value LR_JS_S of the left and right jack stroke difference and the command value UD_JS_S of the upper and lower jack stroke difference. Then, the operation data acquisition unit 31 calculates the difference between the current value and the command value as an error. The operation data acquisition unit 31 calculates the left and right jack stroke error LR_JS_E using the following equation (3). Furthermore, the operation data acquisition unit 31 calculates the upper and lower jack stroke error UD_JS_E using the following equation (4).

[0032] LR_JS_E=(LR_JS)-(LR_JS_S) …Formula (3) however, LR_JS_E is a value indicating the left and right jack stroke error. LR_JS is a value that indicates the current value of the difference between the left and right jack strokes. LR_JS_S is a value indicating the indicated value of the difference between the left and right jack strokes.

[0033] UD_JS_E=(UD_JS)-(UD_JS_S) …Formula (4) however, UD_JS_E is a value indicating the upper and lower jack stroke error. UD_JS is a value that indicates the current value of the upper and lower jack stroke difference. UD_JS_S is a value indicating the indicated value of the upper and lower jack stroke difference.

[0034] The operation data acquisition unit 31 may acquire operation data indicating the azimuth angle and pitch angle together with or instead of the jack stroke difference as operation data indicating the propulsion direction. For example, the operation data acquisition unit 31 acquires the current values ​​and command values ​​of the azimuth angle and pitch angle. The operation data acquisition unit 31 calculates the differences between the acquired current values ​​and command values ​​of the azimuth angle and pitch angle as the errors of the azimuth angle and pitch angle. The operation data acquisition unit 31 calculates the azimuth angle error YD_E using the following equation (5). Furthermore, the operation data acquisition unit 31 calculates the pitch angle error PD_E using the following equation (6).

[0035] YD_E=YD-YD_S ...Equation (5) however, YD_E is a value indicating the azimuth angle error. YD is a value indicating the current value of the azimuth angle. YD_S is a value indicating the indicated value of the azimuth angle.

[0036] PD_E=PD-PD_S ...Equation (6) however, PD_E is a value indicating the pitch angle error. PD is a value indicating the current value of the pitch angle. PD_S is a value indicating the command value of the pitch angle.

[0037] Furthermore, the operation data acquisition unit 31 acquires information indicating time-series changes in the total thrust and the cutter torque as operation data related to the operation of the thrust. The total thrust is the sum of the forces that propel the shield excavator 10. The operation data acquisition unit 31 acquires, for example, the total thrust by multiplying the thrust of each propulsion jack by the number of propulsion jacks 20 selected by the operator. The cutter torque is a rotational force that rotates the cutter head 15. The operation data acquisition unit 31 acquires, as the cutter torque, a torque measured by a measuring device, such as a torque meter, provided on the cutter head 15.

[0038] The operation data acquisition unit 31 stores the acquired operation data in the operation data storage unit 34. More specifically, the operation data acquisition unit 31 stores the current values, command values, and errors of the left-right jack stroke difference and the up-down jack stroke difference in the operation data storage unit 34. The operation data acquisition unit 31 stores the current values, command values, and errors of the azimuth angle and pitch angle in the operation data storage unit 34. The operation data acquisition unit 31 stores the total thrust and cutter torque in the operation data storage unit 34.

[0039] The machine condition acquisition unit 32 acquires the machine conditions of the shield excavator 10 for which an operation corresponding to the operation data acquired by the operation data acquisition unit 31 has been performed. The machine condition acquisition unit 32 acquires, as machine conditions, the machine diameter, machine length, equipped cutter torque, and equipped total thrust of the shield excavator 10 for which an operation corresponding to the operation data has been performed. The machine condition acquisition unit 32 stores the acquired machine conditions in the machine condition memory unit 35.

[0040] Furthermore, the machine condition acquisition unit 32 acquires the machine conditions of the shield excavator 10 that use estimated values ​​from the estimation model. The machine condition acquisition unit 32 acquires the machine diameter, machine length, equipped cutter torque, and equipped total thrust of the shield excavator 10 that use estimated values ​​from the estimation model as machine conditions. The machine condition acquisition unit 32 stores the acquired machine conditions in the machine condition memory unit 35.

[0041] The standardized data generation unit 33 generates standardized data. The standardized data generation unit 33 generates standardized data by performing standardization using the size (machine diameter and machine length) as each geometric property of the shield excavator 10.

[0042] Here, a specific method for generating standardized data will be described using Figure 3. Figure 3 is a diagram for explaining the processing performed by an information processing device 30 according to an embodiment. Figure 3 schematically shows two shield excavators 10 (shield excavators 10-1, 10-2) with different machine conditions. The machine conditions of shield excavator 10-1 are machine diameter D1 and machine length L1. The machine conditions of shield excavator 10-2 are machine diameter D2 and machine length L2. In this figure, (D1>D2, L1>L2).

[0043] Assume that the shield excavators 10-1 and 10-2 are operated to change their machine orientation by an azimuth angle θ relative to their direction of travel (z-axis direction). In this case, the current value LR_JS1 of the difference between the left and right jack strokes of the shield excavator 10-1 and the current value LR_JS2 of the difference between the left and right jack strokes of the shield excavator 10-2 have the relationship shown in equation (7) below. The standardized data generator 33 uses the relationship shown in equation (7) to calculate standardized data LR_JS_ST of the difference between the left and right jack strokes using equation (8) below.

[0044] LR_JS1=(D1 / D2)×LR_JS2…Formula (7) LR_JS_ST=(LR_JS1) / D1 =(LR_JS2) / D2 …Equation (8) however, LR_JS1 is the current value of the difference between the left and right jack strokes of the shield excavator 10-1. D1 is the machine diameter of the shield tunneling machine 10-1. LR_JS2 is the current value of the difference between the left and right jack strokes of the shield excavator 10-2. D2 is the machine diameter of the shield excavator 10-2. LR_JS_ST is the standardized data of the current value of the difference between the left and right jack strokes.

[0045] As shown in equation (8), the standardized data LR_JS_ST makes the two left and right jack stroke differences LR_JS1 and LR_JS2, which have different machine conditions, equal. In other words, the standardized data LR_JS_ST is a standardized left and right jack stroke difference that is independent of machine conditions. The standardized data generating unit 33 generates each standardized data by, for example, applying equation (8) to the current value, the indicated value, and the error of the left and right jack stroke difference.

[0046] Similarly, for the upper and lower jack stroke difference, assume that the shield excavators 10-1 and 10-2 are operated to change the orientation of the machines by a pitch angle φ relative to the direction of travel (z-axis direction). In this case, the current value UD_JS1 of the upper and lower jack stroke difference of the shield excavator 10-1 and the current value UD_JS2 of the upper and lower jack stroke difference of the shield excavator 10-2 have the relationship shown in equation (9) below. The standardized data generator 33 utilizes the relationship shown in equation (9) to calculate standardized data UD_JS_ST of the upper and lower jack stroke difference using equation (10) below.

[0047] UD_JS1=(D1 / D2)×UD_JS2…Formula (9) UD_JS_ST=(UD_JS1) / D1=(UD_JS2) / D2 …Formula (10) however, UD_JS1 is the current value of the difference between the upper and lower jack strokes of the shield excavator 10-1. D1 is the machine diameter of the shield tunneling machine 10-1. UD_JS2 is the current value of the difference between the upper and lower jack strokes of the shield excavator 10-2. D2 is the machine diameter of the shield excavator 10-2. UD_JS_ST is the standardized data of the current value of the upper and lower jack stroke difference.

[0048] The standardized data generating unit 33 generates standardized data for each of the current value, the indicated value, and the error of the upper and lower jack stroke difference by applying the formula (10) to each of them. Specifically, the standardized data generating unit 33 can apply a similar concept to data indicating the azimuth angle and pitch angle as operation data indicating the propulsion direction. For example, the standardized data generating unit 33 calculates standardized data YD_ST of the azimuth angle using the following equation (11). The standardized data generating unit 33 calculates standardized data PD_ST of the pitch angle using the following equation (12).

[0049] YD_ST=YD1 / D1 =YD2 / D2 …Equation (11) however, YD_ST is the standardized data for the current value of the azimuth angle. YD1 is the current value of the azimuth angle of the shield excavator 10-1. D1 is the machine diameter of the shield tunneling machine 10-1. YD2 is the current value of the azimuth angle of the shield excavator 10-2. D2 is the machine diameter of the shield excavator 10-2.

[0050] PD_ST=PD1 / L1 =PD2 / L2 …Equation (12) however, PD_ST is the standardized data of the current value of the pitch angle. PD1 is the current value of the pitch angle of the shield excavator 10-1. L1 is the machine length of the shield excavator 10-1. PD2 is the current value of the pitch angle of the shield excavator 10-2. L2 is the machine length of the shield excavator 10-2.

[0051] The standardized data generating unit 33 generates standardized data for data indicating total thrust as operation data indicating propulsive force. The standardized data generating unit 33 uses the ratio of the total thrust used in actual operation to the upper limit of the thrust equipped in the shield excavator 10 (equipped total thrust) as standard data.

[0052] The standardized data generation unit 33 generates the standardized data of the total thrust by dividing the total thrust by the total equipped thrust. The total thrust is the sum of the thrusts by the propulsion jacks 20 selected by the operator. The total equipped thrust is the sum of the thrusts by all the propulsion jacks 20 equipped on the shield excavator. The standardized data generation unit 33 calculates the standardized data N_ST of the total thrust using the following equation (13).

[0053] N_ST=N1 / N_SB1 =N2 / N_SB2 …Equation (13) however, N_ST is the standardized data for the current value of total thrust. N1 is the current value of the total thrust of the shield excavator 10-1. N_SB1 is the total thrust of the shield excavator 10-1. N2 is the current value of the total thrust of the shield excavator 10-2. N_SB2 is the total thrust of the shield excavator 10-2.

[0054] The standardized data generating unit 33 generates standardized data for data indicating cutter torque as operation data indicating propulsive force. The standardized data generating unit 33 uses the ratio of the torque used in actual operation to the upper limit of the cutter torque (equipped cutter torque) equipped in the shield excavator 10 as standard data. Specifically, the standardized data generating unit 33 generates the standardized data of the cutter torque by dividing the cutter torque by the equipped cutter torque. The cutter torque is the rotational force that rotates the cutter head 15, and is a measurement value measured using a measuring device, such as a torque meter, provided on the cutter head 15. The equipped cutter torque is the upper limit of the rotational force that rotates the cutter head 15. The standardized data generating unit 33 calculates the standardized data CT_ST of the cutter torque using the following equation (14).

[0055] CT_ST=CT1 / CT_SB1 =CT2 / CT_SB2 …Equation (14) however, CT_ST is standardized data for the current value of the cutter torque. CT1 is the current value of the cutter torque of the shield excavator 10-1. CT_SB1 is the equipped cutter torque of the shield excavator 10-1. CT2 is the current value of the cutter torque of the shield excavator 10-2. CT_SB2 is the equipped cutter torque of the shield excavator 10-2.

[0056] Here, it is known that the equipped cutter torque is a value proportional to the cube of the machine diameter D. Specifically, the equipped cutter torque is generally designed using the following equation (15).

[0057] CT_SB=α×D 3 …Formula (15) however, CT_SB is the installed cutter torque. D is the machine diameter. α is the torque coefficient.

[0058] In equation (15), the torque coefficient α varies depending on the machine diameter (shield outer diameter), soil quality, etc. Generally, for earth pressure shields, the torque coefficient α is set in the range of approximately 10 to 25. For slurry shields, the torque coefficient α is set in the range of approximately 8 to 20. When excavating hard ground such as gravel, rubble, or bedrock, the torque coefficient α may be set to a value outside the above range. Taking advantage of the property that the installed cutter torque is proportional to the cube of the machine diameter D, for the same type of shield (earth pressure type or mud water pressure type), the standardized data generation unit 33 may calculate the standardized cutter torque data CT_ST using the following equation (16).

[0059] CT_ST=CT1 / (D1) 3 =CT2 / (D2) 3 …Formula (16) however, CT_ST is standardized data for the current value of the cutter torque. CT1 is the current value of the cutter torque of the shield excavator 10-1. D1 is the machine diameter of the shield tunneling machine 10-1. CT2 is the current value of the cutter torque of the shield excavator 10-2. D2 is the machine diameter of the shield excavator 10-2.

[0060] The standardized data generating unit 33 stores the generated standardized data in the standardized data storage unit .

[0061] The operation data storage unit 34 stores operation data. Fig. 4 is a diagram showing an example of information stored in the operation data storage unit 34 according to the embodiment. This diagram shows an example of operation data for the left-right jack stroke difference. The same applies to operation data corresponding to the upper-lower jack stroke difference, azimuth angle, pitch angle, cutter torque, and thrust. In this diagram, the operation data includes information corresponding to time t, the command value s(t) of the difference between the left and right jack strokes, the current value a(t) of the difference between the left and right jack strokes, and the error d(t) of the difference between the left and right jack strokes. Information indicating the time when the operation was performed is stored in time t. The command value s(t) stores the command value corresponding to the operation performed at time t. The current value a(t) stores the operation value performed at time t. The error d(t) stores the error in the operation performed at time t (the difference between the command value and the operation value). In this way, the operation data stores time-series changes in the data corresponding to the instruction value, operation value, and error.

[0062] The machine condition storage unit 35 stores machine conditions. Figure 5 is a diagram showing an example of information stored in the machine condition storage unit 35 according to an embodiment. In this diagram, the machine conditions include information corresponding to the machine diameter, machine length, equipped cutter torque, and equipped total thrust. The machine diameter stores the outer diameter of the circle corresponding to the bottom surface of a cylindrical shield excavator. The machine length stores the height of the cylinder of a cylindrical shield excavator. The equipped cutter torque stores the upper limit of the torque that can be applied to the cutter head 15 by the electric motor (motor that rotates the cutter head 15) equipped on the shield excavator. The equipped total thrust stores the value obtained by multiplying the upper limit of the thrust per propulsion jack 20 by the number of propulsion jacks 20 equipped on the shield excavator. This figure shows the machine conditions for the shield tunneling machine used at site A (e.g., shield tunneling machine 10-2) and the shield tunneling machine used at site B (e.g., shield tunneling machine 10-1). The shield tunneling machine used at site A (e.g., shield tunneling machine 10-2) has a machine diameter of 2.68 m, a machine length of 5.47 m, an equipped cutter torque of 557 kN m, and an equipped total thrust of 7000 kN. The shield tunneling machine used at site B (e.g., shield tunneling machine 10-1) has a machine diameter of 11.93 m, a machine length of 11.23 m, an equipped cutter torque of 26,624 kN m, and an equipped total thrust of 128,000 kN.

[0063] The standardized data storage unit 36 ​​stores standardized data. This figure shows an example of standardized data for the difference between the left and right jack strokes. The same applies to standardized data corresponding to the difference between the top and bottom jack strokes, azimuth angle, pitch angle, cutter torque, and thrust. In this figure, the value obtained by dividing the error in the difference between the left and right jack strokes by the machine diameter D1 is shown as the standardized data for the difference between the left and right jack strokes.

[0064] The learning data generation unit 37 generates learning data. The learning data generation unit 37 converts the standardized data into operation data corresponding to the operation of the shield excavator 10 to be estimated, using the machine conditions of the shield excavator 10 to be estimated. The learning data generation unit 37 uses the data converted into operation data corresponding to the operation of the shield excavator 10 to be estimated as learning data to be used for learning the operation of the shield excavator 10 to be estimated. In this way, the learning data generation unit 37 generates learning data.

[0065] The learning data generation unit 37 performs an operation that is the inverse of the operation used to generate the standardized data, thereby converting the standardized data into operation data corresponding to the operation of the shield excavator 10 to be estimated. For example, in the case of the left and right jack stroke difference, the learning data generation unit 37 multiplies the standardized data for the left and right jack stroke difference by the machine diameter of the shield excavator 10 to be estimated. In this way, the learning data generation unit 37 converts the standardized data into operation data corresponding to the operation of the shield excavator 10 to be estimated. More specifically, the learning data generation unit 37 uses the following equation (17) to convert the standardized data for the left and right jack stroke difference into operation data for the left and right jack stroke difference corresponding to the operation of the shield excavator 10 to be estimated. In equation (17), it is assumed that the shield excavator 10-3 is used as the shield excavator to be estimated.

[0066] LR_JS3=(LR_JS_ST)×D3…Formula (17) however, LR_JS3 is the current value of the difference between the left and right jack strokes corresponding to the operation of the shield excavator 10-3. LR_JS_ST is the standardized data of the current value of the difference between the left and right jack strokes. D3 is the machine diameter of the shield excavator 10-3.

[0067] While the left-right jack stroke difference has been described as an example above, the same applies to the up-down jack stroke difference, azimuth angle, pitch angle, cutter torque, and thrust. The learning data generation unit 37 can generate learning data that corresponds to the operation of the shield excavator 10 being estimated by converting the standardized data based on the machine conditions of the shield excavator 10 being estimated. This makes it possible to generate a sufficient amount of learning data using the standardized data even for a shield excavator 10 with little operating experience. In other words, a sufficient amount of learning data can be obtained regardless of conditions such as the size and excavation capacity of the shield excavator.

[0068] The estimation model generation unit 38 generates an estimation model using the learning data generated by the learning data generation unit 37. The estimation model generation unit 38 causes the learning model to learn the correspondence between the excavation situation and the operation by learning the learning data generated by the learning data generation unit 37. The excavation situation here refers to the excavation situation, such as the magnitude and direction of the difference (error) between the indicated value and the current value. The operation refers to an operation by the operator, such as a jack pattern or cutter torque selected according to the excavation situation. In addition, any algorithm can be used as the machine learning algorithm here, and examples that can be used include linear regression (e.g., multiple regression analysis), support vector machines, random forests, gradient boosting, neural networks, and recurrent neural networks.

[0069] The storage unit (including the operation data storage unit 34, the machine condition storage unit 35, and the standardized data storage unit 36) of the information processing device 30 is configured by a storage medium such as a hard disk drive (HDD), a flash memory, an electrically erasable programmable read-only memory (EEPROM), a random access read / write memory (RAM), a read-only memory (ROM), or a combination of these. The storage unit of the information processing device 30 stores programs for executing various processes to realize the functions of the information processing device 30, and temporary data used when performing the various processes.

[0070] In addition, the functional units (operation data acquisition unit 31, machine condition acquisition unit 32, standardized data generation unit 33, learning data generation unit 37, and estimation model generation unit 38) of the information processing device 30 are realized by having a CPU (Central Processing Unit) and / or GPU (Graphics Processing Unit) that the information processing device 30 has as hardware execute a program.

[0071] <Processing flow performed by information processing device 30> Here, the flow of processing performed by the information processing device 30 will be described with reference to Fig. 7. Fig. 7 is a flowchart showing the flow of processing performed by the information processing device 30 according to the embodiment. Fig. 7 shows the flow of processing in which the information processing device 30 generates and outputs a simple physical model ML.

[0072] As shown in FIG. 7, first, the operation data acquisition unit 31 of the information processing device 30 acquires operation data (step S10). Next, the machine condition acquisition unit 32 of the information processing device 30 acquires the machine conditions of the shield excavator 10 corresponding to the operation data acquired in step S10 (step S11). The standardized data generation unit 33 of the information processing device 30 generates and stores standardized data (step S12). The standardized data generation unit 33 generates standardized data by standardizing the operation data acquired in step S10 using the machine conditions acquired in step S11. For example, the standardized data generation unit 33 generates standardized data by dividing the operation data of the left and right jack stroke difference acquired in step S10 by the machine diameter acquired in step S11. The standardized data generation unit 33 stores the generated standardized data in the standardized data storage unit 36. The learning data generation unit 37 of the information processing device 30 uses the standardized data to generate learning data corresponding to the operation of the shield excavator 10 to be estimated (step S13). The learning data generation unit 37 generates learning data corresponding to the operation of the shield excavator 10 to be estimated by performing an inverse operation of the standardized data generated in step S12 under the machine conditions of the shield excavator 10 to be estimated. For example, the learning data generation unit 37 generates learning data corresponding to the operation of the shield excavator 10 to be estimated by multiplying the standardized data of the left and right jack stroke difference by the machine diameter of the shield excavator 10 to be estimated. The estimation model generation unit 38 of the information processing device 30 generates an estimation model using the learning data (step S14). The estimation model generation unit 38 uses the learning data generated in step S13 to generate an estimation model that estimates the operation of the shield excavator 10 to be estimated.

[0073] As described above, the information processing device 30 of the embodiment includes an operation data acquisition unit 31, a machine condition acquisition unit 32, a standardized data generation unit 33, and a learning data generation unit 37. The operation data acquisition unit 31 acquires operation data indicating the operation record of the shield excavator. The machine condition acquisition unit 32 acquires the machine conditions of the shield excavator 10 on which the operation corresponding to the operation data was performed. The standardized data generation unit 33 standardizes the operation data using the machine conditions to calculate standardized data, which is operation data independent of the machine conditions. The learning data generation unit 37 generates learning data corresponding to the operation of the shield excavator 10 to be estimated by performing an inverse operation of the standardized data using the machine conditions of the shield excavator 10 to be estimated. This allows the information processing device 30 of the embodiment to generate operation data independent of the machine conditions, i.e., standardized data, and convert the standardized data into operation data corresponding to the operation of the shield excavator 10 to be estimated. As a result, even when the estimation target is a shield excavator 10 with little operational experience, it is possible to convert the data into operation data that corresponds to the operation of that shield excavator 10 by using standardized data calculated from operation data obtained from shield excavators 10 under other machine conditions. In other words, a sufficient amount of learning data can be obtained regardless of conditions such as the size and excavation capacity of the shield excavator.

[0074] Here, the effects will be described with reference to Fig. 8. Fig. 8 is a diagram for explaining the effects according to the embodiment. Fig. 8 shows four graphs showing time-series changes in the amount of change in the point of force for the results of estimation using the estimation model and the results of operation by an experienced operator. In the four graphs in Fig. 8, the horizontal axis represents time, and the vertical axis represents the amount of change in the point of force.

[0075] The left side of Figure 8 shows the time series change in the amount of change in the point of force when an estimation model generated by learning operation data obtained at site A is used to predict the operation of the shield excavator 10 at site B. The upper left side of Figure 8 shows the time series change in the amount of change in the point of force in the x-axis direction (left and right relative to the thrust direction). The lower left side of Figure 8 shows the time series change in the amount of change in the point of force in the y-axis direction (up and down relative to the thrust direction). This figure shows that there is a discrepancy between the operation record of a skilled operator and the predicted value.

[0076] The right side of Figure 8 shows the time series changes in the amount of change in the point of force when an estimation model generated by learning learning data generated by generating standardized data using operation data obtained at site A and converting the standardized data according to the machine conditions of the shield excavator 10 at site B is used to predict the operation of the shield excavator 10 at site B. The upper right side of Figure 8 shows the time series changes in the amount of change in the point of force in the x-axis direction (left and right relative to the thrust direction). The lower right side of Figure 8 shows the time series changes in the amount of change in the point of force in the y-axis direction (up and down relative to the thrust direction). In this figure, the deviation between the operation record of the skilled operator and the predicted value is smaller compared to the graph on the left side of Figure 8. Also, in this figure, the estimation model predicts more cases in which the point of force position changes in the same direction as the operation of the skilled operator compared to the graph on the left side of Figure 8. In other words, it is shown that using standardized data to generate learning data that matches the machine conditions of the shield excavator 10 used for excavation can further improve prediction accuracy.

[0077] Furthermore, in the information processing device 30 of the embodiment, the standardized data generation unit 33 calculates standardized data by dividing the operation data by the machine conditions of the shield excavator 10 on which the operation corresponding to the operation data was performed. The learning data generation unit 37 generates learning data corresponding to the operation of the shield excavator 10 to be estimated by multiplying the standardized data by the machine conditions of the shield excavator 10 to be estimated. As a result, the information processing device 30 of the embodiment can convert the operation data into a value independent of the machine conditions by dividing the operation data by a value corresponding to the machine conditions of the shield excavator 10 on which the operation was performed, and can generate learning data corresponding to the operation of the shield excavator 10 to be estimated by multiplying the standardized data by a value corresponding to the machine conditions of the shield excavator 10 to be estimated.

[0078] Furthermore, in the information processing device 30 of the embodiment, the operation data acquisition unit 31 acquires data related to jack operation to change the propulsion direction, such as the difference between left and right jack strokes or the difference between up and down jack strokes, as operation data. The machine condition acquisition unit 32 acquires an index indicating the size of the surface pushed by the propulsion jack in the shield excavator 10, such as the outer diameter of a circle corresponding to the bottom surface of a cylindrical shield excavator 10, as a machine condition. The standardization data generation unit 33 calculates standardization data by dividing the data related to jack operation to change the propulsion direction by the index. As a result, in the information processing device 30 of the embodiment, even if the operation amount for the operation to change the propulsion direction differs depending on the machine conditions, it is possible to standardize the operation amount as an operation amount that is independent of the machine conditions.

[0079] Furthermore, in the information processing device 30 of the embodiment, the operation data acquisition unit 31 acquires the torque of the cutter head 15 provided on the shield excavator 10 as operation data. The machine condition acquisition unit 32 acquires the equipped cutter torque as the machine condition. The equipped cutter torque is the upper limit of the torque that can be applied to the cutter head 15 by an electric motor (motor) equipped to rotate the cutter head 15 of the shield excavator 10. The standardized data generation unit 33 calculates standardized data by dividing the torque acquired by the operation data acquisition unit 31 by the equipped cutter torque. As a result, the information processing device 30 of the embodiment can use the ratio of the actual torque to the equipped cutter torque as reference data, and can calculate a torque that is independent of the machine condition (equipped cutter torque) even if the torque applied to the cutter head 15 during actual propulsion differs depending on the equipped cutter torque equipped on the shield excavator 10.

[0080] Furthermore, in the information processing device 30 of the embodiment, the operation data acquisition unit 31 acquires the total thrust acting on the shield excavator 10 as operation data. The machine condition acquisition unit 32 acquires the equipped total thrust as the machine condition. The equipped total thrust is the upper limit value of the thrust using the propulsion jacks 20 equipped on the shield excavator 10. The standardized data generation unit 33 calculates standardized data by dividing the total thrust acquired by the operation data acquisition unit 31 by the equipped total thrust. As a result, in the information processing device 30 of the embodiment, the ratio of the actual total thrust to the equipped total thrust can be used as standard data, and even if the total thrust used for actual propulsion differs depending on the equipped total thrust equipped on the shield excavator 10, a total thrust that is not dependent on the machine condition (equipped total thrust) can be calculated.

[0081] Furthermore, in the above-described embodiment, a method of dividing operation data such as the jack stroke difference by the machine diameter has been described as an example of a standardization method, but the present invention is not limited to this. The standardized data may be any data that converts operation data into an operation amount that is independent of at least machine conditions. For example, standardized data that sets the upper and lower jack stroke difference independent of machine length may be generated by dividing the upper and lower jack stroke difference by the machine length. Standardized data that sets the upper and lower jack stroke difference independent of (machine diameter x machine length) may be generated by dividing the left and right jack stroke difference by (machine diameter x machine length). Standardized data that sets the upper and lower jack stroke difference independent of (machine diameter x machine length) may be generated by dividing the upper and lower jack stroke difference by (machine diameter x machine length).

[0082] In the above-described embodiment, the operation data is described as the difference between the left and right jack strokes, the difference between the up and down jack strokes, the azimuth angle, the pitch angle, the total thrust, and the cutter torque, but is not limited to this. Any data indicating at least the direction or thrust of the shield excavator 10 can be used as the operation data. For example, the left and right bending angle can be used as the operation data.

[0083] All or part of the information processing device 30 in the above-described embodiment may be implemented by a computer. In this case, a program for implementing this function may be recorded on a computer-readable recording medium, and the program may be loaded into a computer system and executed. Note that the term "computer system" herein includes hardware such as an OS and peripheral devices. Furthermore, the term "computer-readable recording medium" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into a computer system. Furthermore, the term "computer-readable recording medium" may also include devices that dynamically store programs for a short period of time, such as communication lines used when transmitting programs over a network such as the Internet or a telephone line, or devices that store programs for a fixed period of time, such as volatile memory within a computer system serving as a server or client. The program may also be designed to implement part of the above-described functions, or may be capable of implementing the above-described functions in combination with a program already stored in the computer system, or may be implemented using a programmable logic device such as an FPGA.

[0084] Although an embodiment of the present invention has been described above in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and includes designs within the scope of the gist of the present invention. [Explanation of symbols]

[0085] 10... Shield tunneling machine, 20... Propulsion jack, 30... Information processing device, 31... Operation data acquisition unit, 32... Machine condition acquisition unit, 33... Standardization data generation unit, 37... Learning data generation unit, 38... Estimation model generation unit

Claims

1. an operation data acquisition unit that acquires operation data indicating the operation results of the shield excavator; a machine condition acquisition unit that acquires the machine conditions of a shield excavator that has been operated in accordance with the operation data; a standardized data generating unit that calculates standardized data, which is the operation data that does not depend on the machine conditions, by standardizing the operation data with the machine conditions; a learning data generation unit that generates learning data corresponding to the operation of the shield excavator being estimated by performing an inverse operation of the standardized data used for standardization by the standardized data generation unit using the machine conditions of the shield excavator being estimated; An information processing device comprising:

2. The standardized data generation unit calculates the standardized data by dividing the operation data by the machine conditions of the shield excavator that performed the operation corresponding to the operation data, the learning data generation unit calculates learning data corresponding to the operation of the shield excavator to be estimated by multiplying the standardized data by the machine conditions of the shield excavator to be estimated. The information processing device according to claim 1 .

3. the operation data acquisition unit acquires data related to a jack operation for changing a propulsion direction as the operation data, the machine condition acquisition unit acquires, as the machine condition, an index indicating the size of a surface pushed by a propulsion jack in a shield excavator; the standardization data generation unit calculates the standardization data by dividing the operation data related to the jack operation that changes the propulsion direction by the index. The information processing device according to claim 2 .

4. The data regarding the propulsion direction is a difference between left and right jack strokes or a difference between upper and lower jack strokes, The index is the outer diameter of a circle corresponding to the bottom surface of a cylindrical shield excavator, The information processing device according to claim 3 .

5. The operation data acquisition unit acquires, as the operation data, a torque of a cutter head provided in a shield excavator; The machine condition acquisition unit acquires, as the machine condition, an equipment cutter torque, which is an upper limit of torque that an electric motor equipped to rotate a cutter head of a shield excavator can apply to the cutter head; the standardized data generation unit calculates the standardized data by dividing the torque acquired by the operation data acquisition unit by the installed cutter torque; The information processing device according to claim 1 .

6. The operation data acquisition unit acquires a total thrust acting on the shield excavator as the operation data, The machine condition acquisition unit acquires, as the machine condition, an equipment total thrust, which is an upper limit value of the thrust using a propulsion jack equipped on the shield excavator; the standardized data generation unit calculates the standardized data by dividing the total thrust acquired by the operation data acquisition unit by the equipment total thrust. The information processing device according to claim 1 .

7. An information processing method performed by a computer, comprising: The operation data acquisition unit acquires operation data indicating the operation results of the shield excavator, a machine condition acquisition unit acquires the machine conditions of the shield excavator that has been operated in accordance with the operation data; a standardized data generating unit standardizing the operation data with the machine conditions to calculate standardized data, which is the operation data independent of the machine conditions; the learning data generation unit generates learning data corresponding to the operation of the shield excavator to be estimated by performing an inverse operation of the operation used for standardization by the standardization data generation unit using the machine conditions of the shield excavator to be estimated. Information processing methods.

8. On the computer, Acquire operation data showing the operation results of the shield excavator; acquiring machine conditions of a shield excavator that has been operated in accordance with the operation data; Calculating standardized data, which is the operation data that does not depend on the machine conditions, by standardizing the operation data with the machine conditions; and generating learning data corresponding to the operation of the shield excavator to be estimated by performing an inverse operation of the operation used for the standardization of the standardized data using the machine conditions of the shield excavator to be estimated. program.

Images