Curved drilling operation parameter calculation device, curved drilling operation parameter calculation method and program

Abstract

To calculate appropriate operation parameters even if inappropriate operations are included in past curved drilling operations that are the basis of machine learning.SOLUTION: An information processing apparatus 100 calculates operation parameters using a predetermined physical model, separately from acquiring operation parameters using machine learning, outputs information regarding the operation parameters obtained from machine learning when the difference between these two operation parameters does not exceed the threshold, and outputs information regarding the operation parameters obtained from the physical model when the difference between the two operation parameters exceeds the threshold. By using the operation parameters output in this manner, the main work of curved drilling using a curved drilling machine 1 is performed.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a technique for managing curved hole drilling. [Background technology]

[0002] One known method of countering liquefaction directly beneath existing structures is to drill holes from the ground surface away from the existing structure directly beneath it, and then inject chemicals to improve the ground (curved drilling infiltration solidification treatment method).This method requires drilling holes along a diagonal or curved approach from the ground surface, where drilling begins, to directly beneath the existing structure, and requires advanced drilling trajectory management to avoid the risk of damaging buried objects and ensure quality.

[0003] For example, Patent Document 1 proposes a technology that performs high-quality drilling without relying on the skill level of the operator by using machine learning to analyze data accumulated from previous operations performed by skilled operators when performing curved drilling, and then calculating the optimal operating parameters for the next step. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2019-167728 Summary of the Invention [Problem to be solved by the invention]

[0005] However, past curved drilling operations may have included inappropriate operations, and a learning model that includes such inappropriate operations may not always be able to calculate optimal operation parameters.

[0006] The present invention aims to provide a mechanism that can output appropriate operation parameters even if past curved drilling operations that form the basis of machine learning include inappropriate operations. [Means for solving the problem]

[0007] In order to solve the above problems, the present invention provides a trained algorithm that has been machine-learned using results of accumulating over time of position and attitude parameters indicating the position and attitude of a drilling bit of a curved drilling device and operation parameters for operating the curved drilling device as explanatory variables and objective variables, the trained algorithm comprising: an acquisition unit that inputs information about a planned drilling line at a certain step of curved drilling, the position and attitude parameters, and the operation parameters, and acquires first target operation parameters for the next step; a calculation unit that calculates second target operation parameters for the next step based on physical calculations from the information about the planned drilling line at a certain step of curved drilling, the position and attitude parameters, and the operation parameters; and an output unit that outputs information about the first target operation parameters when the difference between the first target operation parameters acquired by the acquisition unit and the second target operation parameters calculated by the calculation unit does not exceed a threshold, and outputs information about the second target operation parameters when the difference exceeds the threshold. The first target operation parameter and the second target operation parameter are operation parameters related to the curve rate and / or the rotation angle of the drill bit. A curved drilling operation parameter calculation device is provided.

[0009] The calculation unit may also calculate an operating parameter related to the rotation angle of the drilling bit for the next step so that the difference between the planned drilling line and the position of the drilling bit in the next step is smaller than the difference between the planned drilling line and the position of the drilling bit in the certain step.

[0010] The calculation unit may calculate an operating parameter related to the bending rate of the drilling bit for the next step based on the inclination angle of the rod at the certain step, the rotation angle of the drilling bit at the certain step, the bending rate at the certain step, and the correction angle of the rod from the certain step to the next step.

[0011] The present invention also provides a trained algorithm that has been machine-trained using results of accumulating position and attitude parameters indicating the position and attitude of a drilling bit of a curved drilling device and operation parameters for operating the curved drilling device over time as explanatory variables and objective variables, the trained algorithm comprising the steps of: inputting information about a planned drilling line at a certain step of curved drilling, the position and attitude parameters, and the operation parameters, and acquiring first target operation parameters for the next step; calculating second target operation parameters for the next step based on physical calculations from information about the planned drilling line at a certain step of curved drilling, the position and attitude parameters, and the operation parameters; and outputting information about the first target operation parameters when the difference between the acquired first target operation parameters and the calculated second target operation parameters does not exceed a threshold, and outputting information about the second target operation parameters when the difference exceeds the threshold. The first target operation parameter and the second target operation parameter are operation parameters related to the curve rate and / or the rotation angle of the drill bit. A method for calculating curved drilling operation parameters is provided.

[0012] The present invention also provides a computer that executes the following steps to a trained algorithm that has been machine-learned using results of accumulating over time of position and attitude parameters indicating the position and attitude of a drilling bit of a curved drilling device and operation parameters for operating the curved drilling device as explanatory variables and objective variables: inputting information about a planned drilling line at a certain step of curved drilling, the position and attitude parameters, and the operation parameters, and acquiring first target operation parameters for the next step; calculating second target operation parameters for the next step based on physical calculations from information about the planned drilling line at a certain step of curved drilling, the position and attitude parameters, and the operation parameters; and outputting information about the first target operation parameters if the difference between the acquired first target operation parameters and the calculated second target operation parameters does not exceed a threshold, and outputting information about the second target operation parameters if the difference exceeds the threshold. The first target operation parameter and the second target operation parameter are operation parameters related to a curve rate and / or a rotation angle of a drill bit. We provide programs that: [Effects of the Invention]

[0013] According to the present invention, even if the learning data that forms the basis of machine learning contains inappropriate operations in past curved drilling operations, appropriate operation parameters can be output. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a diagram showing an overview of a curved hole drilling method according to one embodiment of the present invention. [Figure 2] FIG. 1 is a side view showing an overview of a curved drilling system. [Figure 3] FIG. 10 is a partially enlarged cross-sectional view showing the progress of the drill bit. [Figure 4] FIG. 2 is a block diagram showing a functional configuration of the information processing device. [Figure 5] FIG. 10 is a diagram illustrating the target correction angle of the drill bit in the YZ plane. [Figure 6] FIG. 10 is a diagram illustrating the rotation angle of a drilling bit. [Figure 7] FIG. 10 is a diagram illustrating a target correction angle in the XY plane of the drill bit. [Figure 8] FIG. 10 is a diagram illustrating an example of classification of curve ratios. [Figure 9] This is a diagram explaining the relationship between the bending rate of the next point, the inclination angle of the current point, the rotation angle of the drilling bit, the bending rate, and the target correction angle of the drilling bit. [Figure 10] FIG. 10 is a diagram illustrating an example of calculation of a curve ratio. [Figure 11] 10 is a flowchart showing a processing procedure of the information processing device. DETAILED DESCRIPTION OF THE INVENTION

[0015] An example of an embodiment of the present invention will be described. In Fig. 1, an information processing device 100 generates a trained algorithm through data collection, which collects data for machine learning from data accumulated in past curved drilling operations, and pre-learning based on the collected data. The information processing device 100 inputs data related to the current point of curved drilling into this trained algorithm, and displays optimal control values, i.e., operation parameters, for controlling the drilling machine 1 at the next point of curved drilling as target operation parameters, or automatically controls the drilling machine 1 based on the target operation parameters.

[0016] However, since past curved hole drilling operations may include inappropriate operations, the learned algorithm may not always be able to obtain optimal target operation parameters.

[0017] Therefore, the information processing device 100 calculates a target operation parameter (second target operation parameter) using a predetermined physical model, separately from calculating a target operation parameter (first target operation parameter) using machine learning, and if the difference between them does not exceed a threshold, outputs information about the target operation parameter (first target operation parameter) obtained from the machine learning, and if the difference between them exceeds the threshold, outputs information about the target operation parameter (second target operation parameter) obtained from the physical model. The actual work of curved drilling using the curved drilling machine 1 is performed using the target operation parameter (first target operation parameter or second target operation parameter) output in this way.

[0018] The information processing device 100 shown in Figure 1 includes hardware for implementing a computer, such as a processor, memory, storage, communication device, input device, output device, and buses connecting these. Each function of the information processing device 100 is realized by loading predetermined software (programs) onto hardware such as the processor and memory, causing the processor to perform calculations, control communications via the communication device, acquire data transmitted from other devices, and control at least one of reading and writing data from and to the memory and storage. The information processing device 100 functions as a curved drilling operation parameter calculation device according to the present invention.

[0019] As shown in Figure 2, the drilling machine 1 includes a flexible drilling rod 2, a drilling bit 3 supported at the tip of the drilling rod 2, a water supply unit 5 that supplies drilling water to a jet nozzle 4 at the tip of the drilling bit 3, a rotation drive unit that rotates the drilling bit 3 via the drilling rod 2, and a propulsion drive unit that propels the drilling bit 3 via the drilling rod 2. The drilling machine main body 6 includes the rotation drive unit and the propulsion drive unit.

[0020] As shown in Figure 3, the drill bit 3 has a tapered surface 3a formed on one side and a drilling jet water injection nozzle 4 at its tip, which extends in a direction parallel to the tapered surface 3a. Drilling water consisting of bentonite mud is supplied from an aboveground water supply unit 5 (a drilling mud pressure pump) through the rod 2 and is injected at high pressure from the injection nozzle 4. When the drill bit 3 is pushed in without rotating, a curved hole is drilled in the direction of the tapered surface 3a, as shown in Figure 3(a). On the other hand, when the drill bit 3 is pushed in while rotating, a straight hole is drilled, as shown in Figure 3(b).

[0021] As shown in Fig. 2, the drilling machine 1 is provided with a position and attitude measurement unit 7 that is disposed at the tip of the drilling rod 2 and that measures position and attitude parameters that indicate the position and attitude of the drilling bit 3 over time. The operator sequentially checks the position of the drilling bit 3 and the position and attitude parameters such as the rotation angle and tilt angle of the drilling bit 3 on the information processing device 100, and operates or automatically controls the rotation drive unit and the propulsion drive unit while visually checking the control values, etc. of the information processing device 100, to drill along a predetermined planned drilling line.

[0022] The position and orientation measurement unit 7 has built-in measuring instruments such as a gyroscope and accelerometer at the tip of the rod 2, for example, at the leading part of the rod 2, and measures the position, rotation angle and inclination angle of the drill bit 3 from the measured values.

[0023] In addition, as shown in Figure 4, the drilling machine 1 is equipped with an operating value measuring unit 8 that measures control values ​​(water supply pressure, rotational pressure (rotational torque) of the rod 2, rotational speed, propulsion pressure, etc.) for controlling the water supply unit 5, rotation drive unit, and propulsion drive unit, respectively, i.e., operating parameters, over time.

[0024] This operation value measurement unit 8 acquires operation parameters, which are control values ​​to be controlled by the control unit that controls the rotation drive unit and the propulsion drive unit, i.e., water supply pressure, rotation pressure of the rod 2, rotation speed, propulsion pressure, etc. The operation parameters may include the operation amounts of the operating tools of the drilling machine 1, such as levers, dials, switches, and numeric keypads for inputting numerical values ​​that operate the water supply unit 5, rotation drive unit, and propulsion drive unit, and the operation amounts may be quantified and acquired using sensors, encoders, etc.

[0025] The information processing device 100 connected to the drilling machine 1 is equipped with a data logger 9 that records position and attitude parameters and operation parameters. The data logger 9 records the position of the drilling bit 3 measured during drilling work and position and attitude parameters such as the rotation speed, rotation angle, and tilt angle of the drilling bit 3 over time, and also records the control values ​​for controlling the rotation drive unit and the propulsion drive unit, i.e., the operation parameters, over time.

[0026] The information processing device 100 also includes a data generation unit 10 that generates machine learning data and input data based on a data group of recorded position and orientation parameters and operation parameters. The data generation unit 10 generates machine learning data and input data in which the measured position and orientation parameters and operation parameters are synchronized so that they are at the same time, and outputs the data to the machine learning unit 11 or the analysis unit 12. The output to the machine learning unit 11 may be transmitted by a data transfer device, or a removable data storage device such as a USB memory may be used.

[0027] The machine learning unit 11 may be provided in the information processing device 100, or in a computer device separate from the information processing device 100, or may be provided on the cloud. The machine learning unit 11 performs machine learning based on a data group of position and attitude parameters and operation parameters recorded and accumulated in past work, thereby generating a trained algorithm that calculates, as target operation parameters, optimal operation parameters for the next step corresponding to the input planned drilling line, position and attitude parameters, and operation parameters. The method of this machine learning is not particularly limited, and for example, a multivariate analysis method, a support vector machine method, a neural network method, etc. may be used.

[0028] The analysis unit 12 provided in the information processing device 100 includes a calculation unit 12a that performs calculation processing and a memory unit 12b that stores a learned algorithm executed by the calculation unit 12a. By executing the learned algorithm using the calculation unit 12a, optimal operation parameters for the next step corresponding to position and attitude parameters and operation parameters input manually, from the position and attitude measurement unit 7, the operation value measurement unit 8, or the data generation unit 10 are calculated, and output as target operation parameters to an operation screen for the operator or an automatic control device. In other words, the analysis unit 12 functions as an acquisition unit that inputs information about the planned drilling line at a certain step of curved drilling, the position and attitude parameters, and the operation parameters to a learned algorithm that has been machine-trained using the results of accumulating position and attitude parameters indicating the position and attitude of the drilling bit of the curved drilling device and the operation parameters for operating the curved drilling device over time as explanatory variables and objective variables, and acquires first target operation parameters for the next step.

[0029] As described above, the information processing device 100 acquires target operation parameters (first target operation parameters) using machine learning, and calculates target operation parameters (second target operation parameters) using a predetermined physical model. The calculation unit 13 calculates target operation parameters (second target operation parameters) for the next step based on physical calculations from information about the planned drilling line, position and attitude parameters, and operation parameters in a certain step of curved drilling.

[0030] If the difference between the target operation parameter (first target operation parameter) acquired by the analysis unit 12 (acquisition unit) and the target operation parameter (second target operation parameter) calculated by the calculation unit 13 does not exceed a threshold, the output unit 14 outputs information about the target operation parameter (first target operation parameter) acquired by the analysis unit 12 (acquisition unit), and if the difference exceeds the threshold, outputs information about the target operation parameter (second target operation parameter) calculated by the calculation unit 13.

[0031] (A. Machine Learning) Next, machine learning according to this embodiment will be described. The specific steps of machine learning are as follows. (A-1. Data Collection) In multiple curved drilling operations that have been performed in the past, position and attitude parameters that indicate the position and attitude of the drilling bit 3 are recorded over time to accumulate position and attitude parameter data, and operation parameter data is collected by recording operation parameters for controlling the water supply unit 5, rotation drive unit, and propulsion drive unit over time to accumulate operation parameter data.

[0032] Normally, actual curved hole drilling work is performed in the following procedure, and the position and posture parameters and operation parameters at each step are acquired and recorded along with the time series.

[0033] First, the drilling machine 1 is set up on the surface of the ground at the drilling start position, and a guide pipe is installed if necessary. The rod 2 is then connected to the rotary drive shaft of the drilling machine body 6, and the drilling bit 3 is inserted into the ground by rotary propulsion or propulsion alone, and drilling begins. The water supply unit 5 also starts to supply drilling water, such as bentonite mud, through the rod 2.

[0034] At that time, the position and attitude measurement unit 7 starts measuring the position and attitude parameters, i.e., the three-dimensional position, rotation angle, and inclination angle of the drilling bit 3, and transmits these to the data logger 9 as needed, and the operation value measurement unit 8 starts measuring the operation parameters, i.e., the water supply pressure of the drilling water by the water supply unit 5, the rotation pressure and rotation speed of the rod 2 of the rotation drive unit, and the thrust pressure of the rod 2 of the thrust drive unit, and transmits these to the data logger 9 as needed. In addition, if necessary, the amount of operation of the operating tools used to operate these items is measured, and transmitted to the data logger 9 as needed.

[0035] Then, while each of the pipes making up the rod 2 is added, the drilling machine main body 6 applies rotational and propulsive forces to the rod 2, and the drilling bit 3 at the tip of the rod 2 drills a hole while advancing it linearly underground to the specified position. If the advancing direction or advancing speed changes due to ground conditions or other reasons, the operator adjusts the rotational pressure and rotational speed of the rod 2 in the rotation drive unit, the propulsion pressure of the propulsion drive unit, and the water supply pressure of the water supply unit 5 so that the planned drilling line and drilling speed are achieved. Then, while the work is being carried out, the position and attitude measurement unit 7 constantly measures the position and attitude parameters, i.e., the position, rotation angle, and tilt angle of the drilling bit 3, and transmits these to the data logger 9.

[0036] Furthermore, the operation value measuring unit 8 measures the changing rotation pressure and rotation speed of the rod 2 of the rotation drive unit, the propulsion pressure of the propulsion unit, and the water supply pressure of the water supply unit 5 over time, and transmits these to the data logger 9. If necessary, the amount of operation of an operating tool such as a lever operated during adjustment may also be quantitatively measured and transmitted to the data logger 9.

[0037] On the other hand, when performing curved drilling, the rotational pressure and rotational speed of the rod 2 of the rotary drive unit are adjusted to align the rotational angle of the drill bit 3 to the desired position, and the rotation of the rod 2 is stopped. In this state, drilling water is sprayed while the thrust pressure of the thrust drive unit is adjusted to provide only thrust force to the rod 2. In this case, as shown in Figure 3(a), the thrust direction of the drill bit 3 gradually changes due to the force acting from the ground on the tapered surface 3a of the drill bit 3. The water supply pressure of the water supply unit 5 and the thrust pressure of the thrust drive unit are adjusted to propel the drill bit 3 curvedly underground along the planned drilling line. The rotational pressure and rotational speed of the rod 2 of the rotary drive unit are also adjusted as necessary. During the drilling operation, the position and orientation measurement unit 7 continuously measures the position, rotational angle, and tilt angle of the drill bit 3 over time and transmits the results to the data logger 9.

[0038] The operation value measuring unit measures the rotation pressure and rotation of the rod 2 of the rotation drive unit, which change depending on the operation. The speed, the propulsion pressure of the propulsion drive unit, and the water supply pressure of the water supply unit 5 are measured as time passes and transmitted to the data logger 9. If necessary, the amount of operation of the operating tool operated during adjustment is also quantitatively measured and transmitted to the data logger 9.

[0039] Next, once the drilling bit 3 is pointed in the specified direction, the rotational pressure and rotational speed of the rod 2 of the rotation drive unit, the thrust pressure of the thrust drive unit, and the water supply pressure of the water supply unit 5 are adjusted, and the drilling bit 3 moves forward in a straight line in the specified direction, and the position, posture, and operation parameters are measured in the same manner.

[0040] Straight and curved thrusting is performed in accordance with the planned drilling line in this way, and the position and posture parameters and operation parameters at each point in time are measured over time and transmitted to the data logger 9 for recording.

[0041] (A-2. Data generation) The data generation unit 10 synchronizes the position and orientation parameters and operation parameters that have been measured separately and transmitted to the data logger 9 so that they are at the same time, and generates and stores data for machine learning. The data group collected in the above-mentioned data collection is then generated as data for machine learning and registered in the machine learning unit 11.

[0042] (A-3. Preliminary Study) In the pre-learning, the machine learning unit 11 performs machine learning on a group of data on position and posture parameters and operation parameters from accumulated past work, and generates a learning algorithm for obtaining the optimal target operation parameters (first target operation parameters) for the next step that correspond to the planned drilling line, position and posture parameters, and operation parameters input during this work.

[0043] This pre-learning assumes a number of situations in actual work, associates these assumptions with data groups of position and attitude parameters and operation parameters, and performs machine learning to derive rules for calculating optimal operation parameters for each situation. For example, in data generation, data from skilled operators is classified and aggregated, and machine learning data related to the skilled operators is subjected to machine learning using a neural network or the like to generate a trained algorithm (analysis formula) that outputs optimal target operation parameters (first target operation parameters) that imitate the operation of the skilled operator for the input planned drilling line, position and attitude parameters, and operation parameters. Each generated trained algorithm is stored in the memory unit 12b of the analysis unit 12.

[0044] (B. Physical Model) Next, a procedure for calculating the target operation parameters (second target operation parameters) using a physical model will be described.

[0045] In curved drilling, the main operations that the operator performs, or the main operations that the operator pays attention to during operation, are the operation parameters related to the "rotation angle of the drilling bit" and the "curve rate." This is because, as explained using FIG. 3, the rotation angle of the drilling bit (i.e., the orientation of the tapered surface 3a) is the main factor that determines the direction of travel of the drilling bit. Furthermore, the curve rate is the rate at which the operator performs a curved operation during the period in which drilling progresses a predetermined distance (e.g., 1 m), and this rate at which the operator performs a curved operation (curve rate) reflects the operator's intention to curve the drilling bit. Therefore, in this embodiment, a target operation parameter (second target operation parameter) is calculated for these two operation parameters using a physical model.

[0046] (B-1. Drilling bit rotation angle) First, we will explain the physical model for calculating the drilling bit rotation angle. Figure 5 is a diagram for explaining the target correction angle in the YZ plane, and illustrates the actual drilling position relative to the planned drilling line as viewed from a vertical cross section. Here, the Z axis is the up-down direction (vertical direction), and the Y direction is the horizontal component of the drilling direction. In Figure 5, the target correction angle θ in the YZ plane of the drilling direction z is set as the angle relative to a target point that will reduce the current deviation ΔZ in the Z direction to a specified ratio at a specified distance ahead in the drilling direction (for example, 3 m ahead) (for example, if the specified ratio is 1 / 2, then the target point will be such that the deviation in the Z direction is 1 / 2 x ΔZ). Note that in this case, since one rod is 3 m long in curved drilling and the operator normally operates with 3 m ahead from the current point in mind, 3 m ahead is used as the specified distance ahead in the drilling direction as described above.

[0047] Next, Figure 6 is a diagram illustrating the drill bit rotation angle. The drill bit rotation angle is the angle (clockwise angle based on the positive direction of the Z axis) formed by the positive direction of the Z axis and the direction in which the tapered surface 3a of the tip of the drill bit described in Figure 3 is facing (normal direction of the tapered surface 3a). Here, the drill bit rotation angle θ that causes the tip of the drill bit to move from the current point toward the target point described above is b The purpose is to find out.

[0048] Here, the target correction angle θ z Assuming that drilling has progressed 1 m, the Z coordinate Z of the current point t is expressed by the following equation: Z t =tanθ z

[0049] Next, Figure 7 is a diagram for explaining the target correction angle in the XY plane, and illustrates the state when the actual drilling position relative to the planned drilling line is viewed from the vertical direction. In Figure 7, the target correction angle θ xis set as the angle relative to the target point where the current deviation amount ΔX in the X direction is reduced to a predetermined ratio at a predetermined distance ahead in the drilling direction (for example, 3 m ahead). (For example, if the predetermined ratio is set to 1 / 2, the deviation amount in the X direction will be 1 / 2 × ΔX.) Then, the target correction angle θ z Using the same concept, the target correction angle θ x Assuming that drilling has progressed 1m, the X coordinate X of the current point t is expressed by the following equation: X t =tanθ x

[0050] From the above, the target correction angle θ x and θ z Assuming that the drilling has progressed 1m, the X and Z coordinates (X t , Z t ) is expressed by the following formula: (X t , Z t )=(tanθ x , tanθ z ) In other words, in Figure 6, the positive direction of the Z axis and the origin and point (tanθ x , tanθ z The angle between the line segment connecting these two points (clockwise angle based on the positive direction of the Z axis) is the drill bit rotation angle θ b is.

[0051] The above is a physical model for calculating the drill bit rotation angle. In other words, this physical model is an algorithm that calculates the operating parameters related to the drill bit rotation angle for the next step (next point) of a certain step (current point) so that the difference between the planned drilling line and the drill bit position at the next step (next point) is smaller than the difference between the planned drilling line and the drill bit position at the previous step.

[0052] (B-2. Curve ratio) Next, a physical model for calculating the bending ratio will be described. As mentioned above, the bending ratio is the ratio of bending operations performed by the operator during the period in which drilling progresses a predetermined distance (for example, 1 m), and the minimum value is "0" and the maximum value is "1." In this embodiment, as shown in Fig. 8, the range of values ​​that the bending ratio can take is classified into four ranges in increments of 0.25: 0 or more and less than 0.25, 0.25 or more and less than 0.5, 0.5 or more and less than 0.75, and 0.75 or more and less than 1.

[0053] Here, by verifying the operations of skilled operators, the inventors found that each classification of the bending rate from the current drilling point to the next drilling point has a relationship with the average values ​​of the inclination angle of the current point, the drilling bit rotation angle, the bending rate at the current point, and the target correction angle, as shown in Figure 9. For example, when the bending rate from the current drilling point to the next drilling point is greater than or equal to 0 and less than 0.25, the average value of the inclination angle of the current point is "-2.99", the average value of the drilling bit rotation angle is "0.25", the average value of the bending rate at the current point is "0.28", and the average value of the target correction angle is "0.65".

[0054] In this embodiment, the curve ratio from the current point of drilling to the next point is calculated by the method illustrated in Fig. 10, taking into consideration the relationship illustrated in Fig. 9. Specifically, four threshold values ​​corresponding to the curve ratios are set for each of the inclination angle of the current point, the drilling bit rotation angle, the curve ratio at the current point, and the target correction angle, and a score is calculated based on the magnitude relationship between the inclination angle of the current point, the drilling bit rotation angle, the curve ratio at the current point, and the target correction angle measured or calculated during drilling and these threshold values, and the curve ratio classification is counted, and the optimal curve ratio is determined based on the sum of these scores. In the example of Figure 10, since the inclination angle of the current point is "-4", a score of "1" is counted for category "B", which corresponds to the range between the thresholds "-3.1" and "-4.5", since the drilling bit rotation angle of the current point is "0.3", a score of "1" is counted for category "B", which corresponds to the range between the thresholds "0.26" and "0.34", since the curvature ratio of the current point is "0.35", a score of "1" is counted for category "C", which corresponds to the range between the thresholds "0.3" and "0.42", and since the target correction angle is "1.7", a score of "1" is counted for category "B", which corresponds to the range between the thresholds "0.8" and "1.9". As a result, the sum of these scores is "3" for category "B" and "1" for category "C," so the optimal turning ratio is determined to be, for example, the median value "0.375" of the range of "0.25 or more and less than 0.5" for category B. Note that "0.3" is used as the threshold value for the turning ratio for both categories "A" and "B," but this is only an example because no distinction is made between categories "A" and "B" in terms of turning ratio.

[0055] The above is the physical model for calculating the bending rate. In other words, this physical model is an algorithm that calculates the operation parameters related to the bending rate of the drill bit for the next step (next point) based on the rod tilt angle at a certain step (current point), the drill bit rotation angle at a certain step, the bending rate at a certain step, and the rod correction angle from one step to the next.

[0056] (operation) Next, the operation of this embodiment will be described with reference to Figure 11. First, the drilling machine main body 6 is installed on the surface layer of the ground at the drilling start position, and the planned drilling line is input into the information processing device 100 (step S11). In addition, a guide pipe is installed as necessary, and the rod 2 is connected to the rotary drive shaft of the drilling machine main body 6. The drilling bit 3 is inserted into the ground by rotary propulsion or propulsion alone, and drilling begins. In addition, the water supply unit 5 starts to supply drilling water, such as bentonite mud water, through the rod 2.

[0057] At this time, the position and attitude parameters, i.e., the three-dimensional position, rotation angle and inclination angle of the drilling bit 3, are measured by the position and attitude measurement unit 7 and transmitted to the data logger 9 at any time, and the operation parameters, i.e., the water supply pressure of the drilling water by the water supply unit 5, the rotation pressure and rotation speed of the rod 2 of the rotation drive unit, the thrust pressure of the rod 2 of the thrust drive unit and the operation amount of the operating tool when operating them are measured by the operation value measurement unit and transmitted to the data logger 9 at any time (steps S12, S13).

[0058] The data of the position and attitude parameters and operation parameters transmitted to the data logger 9 are synchronized by the data generation unit 10 so that they are at the same time, thereby generating data for machine learning and input data, and the data for machine learning is stored, while the input data is sequentially output in real time to the analysis unit 12 as explanatory variables for machine learning (step S14).

[0059] The analysis unit 12 inputs the input data transmitted from the data generation unit 10 into the learned algorithm, and acquires optimal operation parameters as first target operation parameters (steps S15 and S16).

[0060] Meanwhile, the calculation unit 13 calculates optimal operation parameters related to the bit rotation angle and the curve rate as second target operation parameters in accordance with the physical calculations exemplified in FIGS. 5 to 10 (steps S17, S18).

[0061] Next, the output unit 14 determines whether the difference between the first target operation parameter and the second target operation parameter exceeds a threshold value (step S19). Note that, among the first target operation parameters, for the target operation parameters that cannot be calculated using a physical model other than the bit rotation angle and curvature rate, which are the second target operation parameters, the processing of step S19 is skipped and the process proceeds to step S20.

[0062] More specifically, in step S19, if the first target operation parameter related to the drill bit rotation angle is within ±30 degrees of the second target operation parameter related to the drill bit rotation angle, the difference between the two is determined to be within a threshold value; if not, the difference is determined to exceed the threshold value. Note that this threshold value of 30 degrees can be determined arbitrarily. Furthermore, if the first target operation parameter related to the turn rate is included in the category indicated by the second target operation parameter related to the turn rate, the difference between the two is determined to be within the threshold value; if not, the difference is determined to exceed the threshold value.

[0063] If the difference between the first target operation parameter and the second target operation parameter does not exceed the threshold (step S19; YES), the output unit 14 determines that the accuracy of the machine learning is high and outputs the first target operation parameter (step S20). On the other hand, if the difference between the first target operation parameter and the second target operation parameter exceeds the threshold (step S19; NO), the output unit 14 determines that the accuracy of the machine learning is not high and outputs the second target operation parameter (step S21).

[0064] These output target operation parameters are displayed on the information processing device 100, and the operator performs operations in accordance with the target operation parameters. The above processing is repeated throughout the period during which the drilling work is being performed.

[0065] According to this embodiment, by working according to the output target operation parameters, the operator can perform high-quality drilling at a predetermined drilling speed regardless of their level of proficiency. Furthermore, even if the learning data that forms the basis of machine learning contains inappropriate operations for past curved drilling operations, it is possible to calculate and output appropriate operation parameters. The machine learning data recorded in this work can be sent to the machine learning unit 11 via a data transfer device or a removable storage device. This data can be used as data from past work and fed back into the above-mentioned data collection and pre-learning, further improving the accuracy of the constructed trained algorithm.

[0066] The present invention may also be a curved drilling operation parameter calculation method comprising the steps of: inputting information about a planned drilling line at a certain step of curved drilling, the position and attitude parameters, and the operation parameters to a trained algorithm that has been machine-learned using the results of accumulating position and attitude parameters indicating the position and attitude of a drilling bit of a curved drilling device and operation parameters for operating the curved drilling device over time as explanatory variables and target variables; calculating second target operation parameters for the next step based on physical calculations from the information about the planned drilling line at a certain step of curved drilling, the position and attitude parameters, and the operation parameters; outputting information about the first target operation parameters if the difference between the acquired first target operation parameters and the calculated second target operation parameters does not exceed a threshold; and outputting information about the second target operation parameters if the difference exceeds the threshold. The present invention may also be a program for a computer to execute the above method. [Explanation of symbols]

[0067] 1 drilling machine, 2 rod, 3 drilling bit, 4 injection nozzle, 5 water supply unit, 6 drilling machine body, 7 position and attitude measurement unit, 8 operation value measurement unit, 9 data logger, 10 data generation unit, 11 machine learning unit, 12 analysis unit, 12a calculation unit, 12b storage unit, 13 calculation unit, 14 output unit, 100 information processing device

Claims

1. an acquisition unit that inputs information about the planned drilling line at a certain step of curved drilling, the position and attitude parameters, and the operation parameters to a learned algorithm that has been machine-trained using the results of accumulating over time of position and attitude parameters indicating the position and attitude of the drilling bit of the curved drilling device and operation parameters for operating the curved drilling device as explanatory variables and objective variables, and acquires first target operation parameters for the next step; a calculation unit that calculates second target operation parameters for a next step based on physical calculations from information about the planned drilling line, the position and attitude parameters, and the operation parameters in a certain step of the curved drilling; an output unit that outputs information about the first target operation parameter when a difference between the first target operation parameter acquired by the acquisition unit and the second target operation parameter calculated by the calculation unit does not exceed a threshold, and outputs information about the second target operation parameter when the difference exceeds the threshold; Equipped with A curved drilling operation parameter calculation device, wherein the first target operation parameter and the second target operation parameter are operation parameters relating to the curve rate and / or the rotation angle of the drilling bit.

2. The calculation unit calculates an operation parameter related to a rotation angle of the drilling bit for the next step so that a difference between the planned drilling line and the position of the drilling bit in the next step is smaller than a difference between the planned drilling line and the position of the drilling bit in the certain step. The curved drilling operation parameter calculation device according to claim 1.

3. The calculation unit calculates an operation parameter related to the bending rate of the drill bit for the next step based on the tilt angle of the rod at the certain step, the rotation angle of the drill bit at the certain step, the bending rate at the certain step, and the correction angle of the rod from the certain step to the next step. The curved drilling operation parameter calculation device according to claim 1 or 2.

4. a step of inputting information about the planned drilling line at a certain step of curved drilling, the position and attitude parameters and the operation parameters to a trained algorithm that has been machine-learned using the results of accumulating over time of position and attitude parameters indicating the position and attitude of the drilling bit of the curved drilling device and operation parameters for operating the curved drilling device as explanatory variables and objective variables, and obtaining first target operation parameters for the next step; calculating second target operation parameters for a next step based on physical calculations from information about the planned drilling line, the position and attitude parameters, and the operation parameters in a certain step of the curved drilling; outputting information about the first target operation parameter when a difference between the acquired first target operation parameter and the calculated second target operation parameter does not exceed a threshold value, and outputting information about the second target operation parameter when the difference exceeds the threshold value; Equipped with A curved drilling operation parameter calculation method, wherein the first target operation parameter and the second target operation parameter are operation parameters relating to the curve rate and / or the rotation angle of the drilling bit.

5. On the computer, a step of inputting information about the planned drilling line at a certain step of curved drilling, the position and attitude parameters and the operation parameters to a trained algorithm that has been machine-learned using the results of accumulating over time of position and attitude parameters indicating the position and attitude of the drilling bit of the curved drilling device and operation parameters for operating the curved drilling device as explanatory variables and objective variables, and obtaining first target operation parameters for the next step; calculating second target operation parameters for a next step based on physical calculations from information about the planned drilling line, the position and attitude parameters, and the operation parameters in a certain step of the curved drilling; outputting information about the first target operation parameter when a difference between the acquired first target operation parameter and the calculated second target operation parameter does not exceed a threshold value, and outputting information about the second target operation parameter when the difference exceeds the threshold value; Execute A program in which the first target operation parameter and the second target operation parameter are operation parameters related to the curvature rate and / or the rotation angle of the drill bit.

Images