JP2022052963A - Ground condition learning device, ground condition determination device, ground condition learning method, and ground condition determination method - Google Patents

Abstract

To provide a ground condition learning device, a ground condition determination device, a ground condition learning method, and a ground condition determination method that can realize accurate ground determination in a short time.SOLUTION: The present invention is a ground condition learning device 3 that includes a learning data acquisition unit 31 and a learning processing unit 32 to learn the conditions of a tunnel face. The learning data acquisition unit 31 acquires a set of the survey data before construction at the position of the tunnel face, the construction data obtained at the time of construction, and the ground grade as learning data. The learning processing unit 32 has a learning device, and makes the learning device machine-learn so as to output the ground grade by inputting the survey data and the construction data.SELECTED DRAWING: Figure 6

Description

The present invention relates to a ground condition learning device, a ground condition determination device, a ground condition learning method, and a ground condition determination method.

In the mountain tunnel construction project, geophysical surveys such as geophysical surveys and boring surveys are conducted as preliminary surveys, and the ground grade is determined by ground evaluation using, for example, a ground classification table. At the design stage, it is common to apply a standard support pattern corresponding to the ground grade determined in the preliminary survey. On the other hand, at the construction stage, the geological condition of the face is recorded as a face observation record, evaluation points are given based on the face observation record, and the ground grade is selected by the ground evaluation using the evaluation points (new face evaluation). Point method).

As described above, the ground evaluation at the construction stage is mainly performed based on the photographs observed or taken by the engineer in the vicinity of the face during and immediately after excavation. There are many dangers in the vicinity of the face, and the evaluation by a geological expert is not always constant, and the geological evaluation may be revised later. As a technique for supporting the evaluation of the ground at the construction stage, for example, the technique described in Patent Document 1 has been proposed.

The evaluation support device described in Patent Document 1 uses a learning divided image generated by dividing a captured image of a face in underground observation by a reference number of pixels as input data, and outputs an evaluation classification for an observation item of the face. It is provided with a learning result storage unit that stores a model generated by machine learning, and a control unit connected to an input unit and an output unit. Then, the control unit divides the evaluation area of the face image of the evaluation target acquired from the input unit by the reference pixel number, applies the learning result to the divided evaluation target divided image, acquires the individual evaluation category, and obtains the individual evaluation. Predict the evaluation category of the evaluation area based on the category.

Further, as a technique for supporting the ground evaluation in the preliminary search stage, the ground evaluation method described in Non-Patent Document 1 is being studied. The ground evaluation method described in Non-Patent Document 1 predicts the ground grade using a neural network (ANN: Artificial Neural Network). Specifically, "elastic wave velocity", "rock type", and "overburden" are used as input data, and "ground grade" is output as output data.

Japanese Unexamined Patent Publication No. 2019-0233392

Chikazu Masuda and 3 others, "Results of Ground Evaluation in Mountain Tunnels Classified as Sedimentary Rocks and Igneous Rocks Using Neural Networks", 47th Symposium on Rock Mechanics, Japan Society of Civil Engineers, January 2020, Lecture number 7, p.34-39

The technique according to Patent Document 1 requires a large amount of high-quality photographed images (high-resolution images with appropriate illuminance at the time of photographing) to be used for learning, so that a large amount of time must be spent on preparation. There's a problem. In addition, since the captured image of the face indirectly expresses information on the deformability and stability of the ground that affects the ground determination, the accuracy of the ground determination based on the captured image of the face is not necessarily high. However, there were still cases where the ground evaluation was revised later. Further, since the technique according to Non-Patent Document 1 performs the ground determination only from the information obtained in the preliminary survey of the ground, it does not realize the accuracy of the ground determination required at the construction stage. In other words, there are restrictions on the budget and time required for the preliminary survey, and there are many sites where the amount of data obtained from the preliminary survey is insufficient due to the technical limitations of the geological survey.

From such a viewpoint, the present invention provides a ground condition learning device, a ground condition determination device, a ground condition learning method, and a ground condition determination method that can realize accurate ground determination in a short time. ..

The ground condition learning device according to the present invention is a ground condition learning device including a learning data acquisition unit and a learning processing unit for learning the state of a tunnel face.
The learning data acquisition unit acquires a set of the survey data before construction at the position of the tunnel face, the construction data obtained at the time of construction, and the ground grade as learning data.
The learning processing unit has a learning device, and the learning device is machine-learned so as to output the ground grade by inputting the survey data and the construction data.

In the ground condition learning device according to the present invention, the learner is machine-learned using learning data including both the information obtained in the investigation stage before construction and the information obtained in the construction stage. Therefore, it is more effective as learning data than in the past, and it is possible to learn in a short time, and when a learning device learned by the method is used, it is possible to realize accurate ground determination.

The ground grade is, for example, a graded classification of the condition of the ground, and it is desirable that the ground grade is the one finally adopted in the actual construction (corresponding to the implemented support pattern).
Further, the survey data and the construction data may be, for example, information generally used in determining the ground grade, and are information having some correlation with the ground grade.
The survey data should include any one of the elastic wave velocities, overburdens and rock species of the ground at the location of the tunnel face. It is desirable that the construction data includes any one of the extrusion displacement amount of the face and the rockfall frequency. The elastic wave velocity and rock type may be measured during construction (that is, the elastic wave velocity and rock type may be included in the data at the time of construction).

The elastic wave velocity and the overburden are reference values or statistical values of a predetermined section divided in the tunnel axial direction, and the rock type may be a representative rock type of a predetermined section divided in the tunnel axial direction.
The rock fall frequency is the number of rock falls that occur in the tunnel face within a predetermined time after the completion of the excavation slip removal work, and the extrusion displacement amount of the face is predetermined after the completion of the excavation slip removal work. It may be a measured value obtained by measuring the tunnel face within the specified time.

The ground condition determination device according to the present invention is a ground condition determination device for determining the condition of a tunnel face, which includes a determination data acquisition unit and an estimation processing unit.
The determination data acquisition unit acquires the survey data before construction at the position of the tunnel face and the construction data obtained at the time of construction.
The estimation processing unit has a learned learner that has been machine-learned to output a ground grade by inputting the survey data and the construction data.

In the ground condition determination device according to the present invention, a learning device that has been machine-learned using learning data including both information obtained in the investigation stage before construction and information obtained in the construction stage is used, so that the accuracy is high. It is possible to realize a good ground judgment.

The ground condition learning method according to the present invention is a ground condition learning method for learning the condition of the tunnel face. This ground-based situation learning method has a learning data acquisition step and a learning processing step.
In the learning data acquisition step, a set of the survey data before construction at the position of the tunnel face, the construction data obtained at the time of construction, and the ground grade is acquired as training data.
In the learning processing step, the learning device is machine-learned so as to output the ground grade by inputting the survey data and the construction data into the learning device.

In the ground-based situation learning method according to the present invention, the learner is machine-learned using learning data including both the information obtained in the investigation stage before construction and the information obtained in the construction stage. Therefore, it is more effective as learning data than in the past, and it is possible to learn in a short time, and when a learning device learned by the method is used, it is possible to realize accurate ground determination.

The ground condition determination method according to the present invention is a ground condition determination method for determining the condition of a tunnel face. This ground condition determination method includes a determination data acquisition step and an estimation processing step.
In the determination data acquisition step, the survey data before construction at the position of the tunnel face and the construction data obtained at the time of construction are acquired.
In the estimation processing step, the ground grade is estimated from the data acquired by using the trained learner that has been machine-learned so as to output the ground grade by inputting the survey data and the construction data.

In the ground condition determination method according to the present invention, a learning device that has been machine-learned using learning data including both information obtained in the investigation stage before construction and information obtained in the construction stage is used, so that the accuracy is high. It is possible to realize a good ground judgment.

According to the present invention, it is possible to realize an accurate ground determination in a short time.

It is a figure for demonstrating the outline of this invention, (a) is an image which shows the standard determination method so far, (b) is an image which shows the determination method by this invention. It is a block diagram of the ground condition learning system which concerns on embodiment of this invention. It is a block diagram of the ground condition determination system which concerns on embodiment of this invention. It is an external view of the face extrusion displacement meter. It is an external view of a rockfall detection device. It is a functional block diagram of the ground condition learning apparatus and the ground condition determination apparatus which concerns on embodiment of this invention. It is an image diagram of a neural network.

Hereinafter, embodiments for carrying out the present invention will be described in detail with reference to the drawings as appropriate. Each figure is merely schematically shown to the extent that the present invention can be fully understood. Therefore, the present invention is not limited to the illustrated examples. In each figure, common components and similar components are designated by the same reference numerals, and duplicate description thereof will be omitted.

[Outline of the present invention]
The outline of the present invention will be described with reference to FIG. FIG. 1 is a diagram for explaining an outline of the present invention. As shown in FIG. 1 (a), until now, the geological condition of the face was recorded as a face observation record, an evaluation point was given based on the face observation record, and a technician was evaluated by the ground using the evaluation point. Was selecting the ground grade (new face evaluation point method). In addition, the measurement results at the time of construction were sometimes used as a reference for selecting the ground grade. On the other hand, as shown in FIG. 1 (b), the inventor of the present invention can obtain the survey data (for example, elastic wave velocity, rock type, overburden) obtained in the preliminary geological survey at the construction stage. Therefore, we devised machine learning using construction data (for example, the amount of extrusion displacement of the face and the frequency of rock fall) that are directly related to the deformability and stability of the ground, which affects the judgment of the ground. Hereinafter, it will be described in detail.

<About the configuration of the ground-based situation learning system according to the embodiment>
The ground-based situation learning system 1A according to the embodiment will be described with reference to FIG. FIG. 2 is a block diagram of the ground condition learning system 1A according to the embodiment. The ground situation learning system 1A is a system for learning the situation of the tunnel face. The ground situation learning system 1A can be used for various types of tunnels without limiting the types of tunnels. In this embodiment, a mountain tunnel will be illustrated as an example of the tunnel.

As shown in FIG. 2, the ground condition learning system 1A mainly includes management terminals 2, 2 ... Installed at the construction site of each tunnel, and the ground condition learning device 3. The management terminal 2 and the ground condition learning device 3 can communicate with each other via the network 4. The ground condition learning system 1A can also be configured as a cloud system, in which case the ground condition learning device 3 corresponds to a cloud server. The configuration of the ground situation learning system 1A shown in FIG. 2 is merely an example.

The management terminal 2 is installed in an office outside the tunnel, for example, and is operated by a tunnel construction manager or the like. Various information about the tunnel construction is stored in the management terminal 2. The management terminal 2 stores, for example, the results of geophysical surveys such as geophysical surveys and boring surveys conducted as preliminary surveys. The information stored in the management terminal 2 is, for example, information such as the elastic wave velocity of the ground, the overburden, and the rock type. Further, the management terminal 2 stores information about the tunnel face acquired during the tunnel construction. The information stored in the management terminal 2 is, for example, the amount of extrusion displacement of the tunnel face, the frequency of falling rocks, the ground grade, and the like. The ground grade is a graded classification of the condition of the ground, for example, the one finally adopted in the actual construction (corresponding to the implemented support pattern).

The ground condition learning device 3 is a device that plays a central role in the ground condition learning system 1A, and learns the condition of the tunnel face using the information collected from each construction site. The ground condition learning device 3 is, for example, from the management terminal 2 used at the construction site of the tunnel under construction, the elastic wave velocity at the position of the tunnel face, the overburden and rock type, the amount of extrusion displacement of the tunnel face, and the falling rock. Collect frequency and ground grade (see white arrow in Figure 2). Then, the ground-based situation learning device 3 makes the learning device (model) machine-learn using the collected information. The details of the learning method will be described later. The trained learner (learned model) learned by the ground condition learning device 3 is delivered to, for example, a construction site of a new tunnel and used for determining the ground of the tunnel (see the dot pattern arrow in FIG. 2). ..

A device to which the learned learner learned by the ground condition learning device 3 is distributed and can determine the state of the tunnel face using the learned learner is referred to as a "ground condition determination device". In this embodiment, since it is assumed that the learned learner is reflected in the management terminal 2, it is referred to as "ground condition determination device 2B". The ground condition learning device 3 can also make the learner machine-learn using the information collected from the ground condition determination device 2B (see the striped arrow in FIG. 2). Further, the ground condition learning device 3 corresponding to the cloud server may be configured to determine the state of the tunnel face using the learned learner. In that case, the ground condition learning device 3 is the “ground condition learning device 3”. (That is, the ground condition determination device may be on the cloud).

<About the configuration of the ground condition judgment system according to the embodiment>
The ground condition determination system 1B according to the embodiment will be described with reference to FIG. FIG. 3 is a block diagram of the ground condition determination system 1B according to the embodiment. The ground condition determination system 1B is a system for determining the condition of the tunnel face. The ground condition determination system 1B can be used for various types of tunnels without limiting the types of tunnels. In this embodiment, a mountain tunnel will be illustrated as an example of the tunnel. It is desirable that the tunnel determined by the ground condition determination system 1B is similar to the tunnel trained by the ground condition learning system 1A. The similarity of tunnels is intended to include various similarities such as similar cross-sectional shapes and cross-sectional areas and similar ground conditions.

As shown in FIG. 3, the ground condition determination system 1B is used by a ground condition determination device 2B having a learned learner, a three-point face extrusion displacement meter 5, rock fall detection devices 6 and 6, and a worker. It is provided with a worker terminal 7. The ground condition determination device 2B and the worker terminal 7 can communicate with each other via, for example, a network provided in the mine. The face extrusion displacement meter 5 and the rock fall detection devices 6 and 6 may be configured to be communicable with the ground condition determination device 2B and the worker terminal 7. The configuration of the ground condition determination system 1B shown in FIG. 3 is merely an example.

The face extrusion displacement meter 5 is a device that measures the extrusion displacement amount of the tunnel face (that is, the displacement amount in the tunnel axial direction). The face extrusion displacement meter 5 is arranged at a position separated from the tunnel face by a predetermined distance, and measures, for example, the amount of extrusion displacement of the tunnel face generated within a predetermined time (for example, 10 minutes) after the completion of excavation slipping work. do. FIG. 4 is an external view of the face extrusion displacement meter 5. As shown in FIG. 4, the face extrusion displacement meter 5 of the present embodiment includes three laser type displacement meters 51, and measures the amount of extrusion displacement at three points in the central portion of the tunnel face. For example, the worker registers the average value of the measured displacements of the three points in the ground condition determination device 2B as the extrusion displacement of the tunnel face. The amount of extrusion displacement of the tunnel face is an example of an index showing the deformability of the tunnel face. The index indicating the deformability of the tunnel face may be, for example, the amount of internal air displacement or the amount of subsidence of the crown (and their initial displacement) of the measuring engineer (measurement in the mine). For example, the amount of displacement for one day from the installation of the target, or the amount of displacement at a fixed excavation length (for example, 3 m or 4 m, which is the length that can be excavated in one day).

In this embodiment, it is assumed that the extrusion displacement amount is registered using the worker terminal 7, but for example, the face extrusion displacement meter 5 automatically performs the ground condition without the operation of the worker. It may be registered in the determination device 2B. Further, the face extrusion displacement meter 5 may measure the displacement amount of a plurality of points other than one point or three points (that is, the number of measurement points is not limited), and information to be registered when a plurality of points are measured. May be a representative value (that is, any one measured value) or a statistical value. Further, although the face extrusion displacement meter 5 of the present embodiment is a self-standing type provided with legs 52, it may be attached to the ceiling or wall of a tunnel.

The rockfall detection device 6 is a device for detecting rockfall on the tunnel face. The rockfall detection device 6 is arranged near the tunnel face and monitors the rockfall of the tunnel face. In this embodiment, two rockfall detection devices 6 are used to monitor the tunnel face. That is, the first rockfall detection device 6a monitors the right half of the tunnel face, and the second rockfall detection device 6b monitors the left half of the tunnel face. The rockfall detection device 6 is installed, for example, within a predetermined time after the completion of the excavation slip removal work (for example, 10 minutes from the completion of the excavation slip removal work) as a monitoring period, and detects rockfalls that occur within the monitoring period. do.

FIG. 5 is an external view of the rockfall detection device 6. As shown in FIG. 5, the rockfall detection device 6 of the present embodiment includes a camera 61 and lighting devices 62 and 62 provided above and below the camera 61, and detects rockfalls using images taken by the camera 61. do. The illuminating tool 62 is, for example, LED (Light Emitting Diode) lighting or organic EL lighting (OLED), and includes at least a wavelength component corresponding to the sensitivity (camera sensitivity) of the camera 61. The camera 61 is, for example, a digital video camera, and captures dozens of images (also referred to as “frames”) per second. A control unit (not shown) of the rockfall detection device 6 detects rockfall by using, for example, the inter-frame difference method. In the inter-frame difference method, an inter-frame difference image is generated between two consecutive frames, and a rockfall is detected using the inter-frame difference image. The technology for detecting rockfall from video is not limited to this.

For example, the worker determines the ground condition by using the total number of times the first rockfall detection device 6a and the second rockfall detection device 6b have detected rockfalls within the monitoring period as the number of times rockfalls have occurred at the tunnel face (number of rockfalls). Register with device 2B. The number of rockfalls on the tunnel face is an example of an index showing the stability of the tunnel face.

In this embodiment, it is assumed that the number of rockfalls is registered using the worker terminal 7, but for example, the rockfall detection device 6 automatically determines the ground condition without the operation of the worker. It may be registered in 2B. Further, one or three or more rockfall detection devices 6 may be used to detect rockfalls on the tunnel face, or a worker may visually detect rockfalls on the tunnel face without using the rockfall detection device 6. .. In addition, even if information indicating the frequency of rockfalls other than the number of rockfalls per predetermined time (including the number of rockfalls per unit time) (for example, information in which the number of rockfalls is divided into levels) is registered in the ground condition determination device 2B. good. Further, although the rock fall detection device 6 of the present embodiment is a self-standing device provided with a leg portion 63, it may be mounted on the ceiling or wall of a tunnel.

The ground condition determination device 2B is a device that plays a central role in the ground condition determination system 1B, and was detected during construction by the results of a geological survey conducted in advance and the face extrusion displacement meter 5 and the rockfall detection device 6. Use the information to determine the status of the tunnel face. The ground condition determination device 2B stores, for example, information on the elastic wave velocity, overburden, and rock type of the ground at the position of the tunnel face. Further, the ground condition determination device 2B acquires, for example, the amount of extrusion displacement of the tunnel face from the face extrusion displacement meter 5, and acquires the frequency of rock fall of the tunnel face from the rock fall detection device 6. Then, the ground condition determination device 2B obtains the ground grade of the tunnel face by inputting such information into the learned learner. The details of the determination method will be described later. The ground condition determination device 2B transmits the determined ground grade information to, for example, the worker terminal 7.

The worker terminal 7 is, for example, a tablet terminal, and information on tunnel construction stored in the management terminal 2 can be browsed. Further, the worker inputs the amount of extrusion displacement measured by the face extrusion displacement meter 5 and the frequency of rock fall detected by the rock fall detection device 6 into the worker terminal 7, and the worker instructs the execution of the tunnel face determination. , These input information are transmitted to the ground condition determination device 2B. In addition, the determination result (information on the ground grade) of the tunnel face condition is displayed on the worker terminal 7, and the worker implements the support pattern (implementation support pattern) based on the determination result (information on the ground grade). ) Is determined.

<About the method of learning the ground condition according to the embodiment>
A method of learning the ground condition according to the embodiment will be described with reference to FIG. FIG. 6 is a functional configuration diagram of the ground condition learning device 3 and the ground condition determination device 2B according to the embodiment. When the ground condition learning device 3 itself determines the condition of the tunnel face, the ground condition learning device 3 and the ground condition determination device 2B can be configured as one device.

The ground situation learning device 3 includes a learning data acquisition unit 31 and a learning processing unit 32, and the learning processing unit 32 has a learning device. The learning device is a learning system in machine learning, and it derives good results by comparing the results of classification, prediction, and judgment based on given data with the actual results that are the correct answers, and adjusting various parameters. Will be possible. The learner is, for example, a neural network, and the present embodiment will also be described assuming a neural network. The learning device is also called a "learning model" or the like.

Neural networks are, as is well known, an information processing method created for the purpose of imitating the movement of the human brain by a computer. Neural networks are good at complex nonlinear processing, and have the feature that it is not necessary to formulate the relationship between explanatory variables and objective variables by using a learning function. As shown in FIG. 7, a neural network is generally composed of an input layer, an intermediate layer, and an output layer. FIG. 7 is an image diagram of a neural network. Explanatory variables are input to the input layer, and objective variables are output from the output layer. The middle class plays a role in linking the two. There are no restrictions on the number of layers in the middle layer and the number of neurons in each middle layer, and they can be set arbitrarily.

The learning data acquisition unit 31 acquires learning data for learning the learning device. The training data is composed as a set of elastic wave velocity, rock type, overburden, extrusion displacement, rockfall frequency and geological grade of the ground at the position of the tunnel face. It is desirable that the correspondence between the training data is strict, but it is not always strict, and there may be a slight deviation in the position of the tunnel face between the data. The learning data should be created based on, for example, the construction results when a similar tunnel is constructed.

The elastic wave velocity is a reference value or a statistical value of a predetermined section divided in the tunnel axial direction. The elastic wave velocity may be a value obtained from a preliminary geological survey, or may be, for example, a measured value measured at intervals of 10 m in the direction of the tunnel axis, a value calculated by statistical processing of the measured value, or the like.
The rock type is a representative rock type of a predetermined section divided in the tunnel axial direction. The rock type may be determined from a preliminary geological survey, or may be determined by observing the exposed ground and excavation scraps in the tunnel. Numerical values are associated with rock types, and the associated numerical values are used for learning data. The correspondence between rock species and numerical values is, for example, "1" for "igneous rock" and "2" for "sedimentary rock".
The overburden is a reference value or a statistical value of a predetermined section divided in the tunnel axis direction. For the overburden, the value obtained from the result of the preliminary geological survey (longitudinal view, etc.) is used.

The extrusion displacement amount is the displacement amount per predetermined time generated in the tunnel face during tunnel construction. The amount of extrusion displacement is measured, for example, within a predetermined time (for example, 10 minutes later) after the completion of the excavation slip-out operation.
The fall frequency is the number of rockfalls per predetermined time that occurred around the tunnel face during tunnel construction, for example, within a predetermined time after the completion of the excavation scraping work (for example, 10 minutes from the completion of the excavation scraping work). It is the number of rockfalls that occurred at the tunnel face.
The ground grade is a ground grade actually adopted at the tunnel construction site, and is acquired for each predetermined section divided in the tunnel axial direction, for example. Numerical values are associated with the ground grade, and the associated numerical values are used for learning data. The correspondence between the ground grade and the numerical value is, for example, "D ground" is "1", "CII ground" is "2", "CI ground" is "3", and "B ground" is "4". "And so on.

The learning processing unit 32 uses the learning data configured as a set of the elastic wave velocity, rock type, overburden, extrusion displacement amount, rockfall frequency, and ground grade of the ground at the position of the tunnel face to machine the learning device. Let them learn. The method of training the learner is not particularly limited, and the learner is trained by, for example, the error back propagation method. Specifically, five pieces of information such as elastic wave velocity, rock type, overburden, face extrusion displacement, and rockfall frequency of training data are input to the input layer, and the results output from the output layer and the ground grade of the training data. Adjust the middle layer based on the error with. That is, the learning processing unit 32 machine-learns the learning device so that an appropriate ground grade is output by inputting the elastic wave velocity, the rock type, the overburden, the displacement amount of the face extrusion, and the rockfall frequency. The number of learning data to be learned is not particularly limited, and for example, learning ends when the expected accuracy (correct answer rate) is reached. The learned learner is delivered to the ground condition determination device 2B.

<Regarding the method for determining the ground condition according to the embodiment>
The method for determining the ground condition according to the embodiment will be described with reference to FIG.
The ground condition determination device 2B includes a determination data acquisition unit 21 and an estimation processing unit 22, and the estimation processing unit 22 has a learned device. The learned learner is machine-learned by the ground-based situation learning device 3. In other words, the trained learner has been trained to output an appropriate ground grade by inputting elastic wave velocity, rock type, overburden, face extrusion displacement amount, and rockfall frequency.

The determination data acquisition unit 21 acquires determination data that is the basis for determining the ground grade. The determination data is composed of a set of elastic wave velocity, rock type, overburden, extrusion displacement amount and rockfall frequency of the ground at the position of the tunnel face. It is desirable that the correspondence between the determination data is strict, but it is not always strict, and there may be a slight deviation in the position of the tunnel face between the data. The information on the elastic wave velocity, rock type, overburden, extrusion displacement amount, and rockfall frequency constituting the determination data may be measured by the same method as the learning data.

The estimation processing unit 22 estimates the condition of the tunnel face using the determination data composed of the elastic wave velocity of the ground at the position of the tunnel face, the rock type, the overburden, the amount of extrusion displacement, and the frequency of rock fall. A trained learner is used to estimate the tunnel face information. As shown in FIG. 7, five pieces of information such as elastic wave velocity, rock type, overburden, face extrusion displacement amount, and rock fall frequency are input to the input layer. In addition, information on the ground grade is output from the output layer. The information of the ground grade output from the output layer is output as a probability that it is, for example, each ground grade (for example, D ground, CII ground, CI ground, B ground).

As described above, in the ground condition learning device 3 according to the present embodiment, learning data including both information (survey data) obtained in the survey stage before construction and information (data at the time of construction) obtained in the construction stage. Make the learner machine-learn using. Therefore, the effectiveness as learning data is higher than in the past, learning is possible in a short time, and the ground condition determination device 2B using the learning device learned by the method performs accurate ground determination. It is feasible.

Although the embodiments of the present invention have been described above, the present invention is not limited to this, and can be carried out without changing the gist of the claims.

In the embodiment, learning is performed using learning data including five items of elastic wave velocity, rock type, overburden, extrusion displacement amount, and rockfall frequency, and judgment data including these five pieces of information is used for learning. I was judging the situation. However, the structure of the training data and the determination data is not limited to this.
For example, a learning device is prepared for each rock type, and learning is performed by inputting learning data including elastic wave velocity, overburden, extrusion displacement amount, and rockfall frequency into the learning device prepared for each rock type. It is also possible to do it. In that case, the condition of the ground is determined by inputting the determination data including the elastic wave velocity, the overburden, the extrusion displacement amount, and the rockfall frequency into the learning device of the corresponding rock type.
In addition, the learner is machine-learned using training data that includes either information on the deformation of the ground (extruded displacement of the tunnel face) or information on stability (frequency of falling rocks) that affects the judgment of the ground. You may. It is desirable to make the learner machine-learn using learning data that includes information on stability (rockfall frequency).

1A ground condition learning system 1B ground condition judgment system 2 management terminal 2B ground condition learning device 3 ground condition learning device 5 face extrusion displacement meter 6 rockfall detection device 7 worker terminal 21 judgment data acquisition unit 22 estimation processing Department 31 Learning data acquisition unit 32 Learning processing unit

Claims (8)

It is a ground-based situation learning device that learns the situation of the tunnel face equipped with a learning data acquisition unit and a learning processing unit.
The learning data acquisition unit acquires a set of the survey data before construction at the position of the tunnel face, the construction data obtained at the time of construction, and the ground grade as learning data.
The learning processing unit has a learning device, and the learning device is machine-learned so as to output the ground grade by inputting the survey data and the construction data.
A ground-based situation learning device characterized by this. The survey data includes any one of the elastic wave velocity, overburden and rock species of the ground at the location of the tunnel face.
The ground-based situation learning device according to claim 1. The construction data includes any one of the extrusion displacement amount of the face and the rockfall frequency.
The ground-based situation learning device according to claim 1 or 2, wherein the device is characterized by the above. The elastic wave velocity and the overburden are reference values or statistical values of a predetermined section divided in the tunnel axial direction.
The rock type is a representative rock type of a predetermined section divided in the tunnel axial direction.
The ground-based situation learning device according to claim 2. The rockfall frequency is the number of rockfalls that occur at the tunnel face within a predetermined time after the excavation scraping work is completed.
The extrusion displacement amount of the face is a measured value obtained by measuring the tunnel face within a predetermined time after the completion of the excavation slip-out work.
The ground-based situation learning device according to claim 3. It is a ground condition determination device that determines the condition of the tunnel face equipped with a determination data acquisition unit and an estimation processing unit.
The determination data acquisition unit acquires the survey data before construction at the position of the tunnel face and the construction data obtained at the time of construction.
The estimation processing unit has a trained learner that has been machine-learned to output a ground grade by inputting the survey data and the construction data.
A ground condition judgment device characterized by this. It is a ground-based situation learning method for learning the situation of the tunnel face.
A learning data acquisition step for acquiring a set of pre-construction survey data at the position of the tunnel face, construction data obtained at the time of construction, and a ground grade as learning data.
It has a learning process step of causing the learner to perform machine learning so as to output the ground grade by inputting the survey data and the construction data into the learner.
A method of learning the ground situation that is characterized by this. It is a method for determining the condition of the ground, which determines the condition of the tunnel face.
Judgment data acquisition step to acquire the survey data before construction at the position of the tunnel face and the construction data obtained at the time of construction,
An estimation processing step for estimating the ground grade from data acquired using a trained learner that has been machine-learned to output the ground grade by inputting the survey data and the construction data. Have,
A method for determining the condition of the ground, which is characterized by the fact that.

Images