JP2022080756A - Automatic groundwater environment prediction system, and automatic groundwater environment prediction method - Google Patents

Abstract

To provide an automatic groundwater environment prediction system that can promptly and automatically predict the groundwater environment during tunnel construction.SOLUTION: The inventive automatic groundwater environment prediction system comprises an estimated spring water amount acquisition unit that acquires the estimated spring water amount calculated based on reproduction conditions by an osmotic flow analysis system, a reproduction analysis instruction unit that causes the osmotic flow analysis system to perform osmotic flow analysis using different permeability coefficients when the estimated spring water amount is not within a predetermined range based on the spring water amount measurement value, and a learning unit that uses the permeability coefficients used as the reproduction condition and the estimated spring water amount obtained by using the water permeability coefficients as teacher data to learn the relationship between permeability coefficients and an estimated spring water amount, thus generating a trained model. The predicted position permeability coefficients obtained by using the trained model are set in the osmotic flow analysis system to perform osmotic flow analysis, thus obtaining the spring water amount to be generated when excavation is performed at the excavation position to be predicted.SELECTED DRAWING: Figure 1

Description

The present invention relates to a groundwater environment automatic prediction system and a groundwater environment automatic prediction method.

When constructing an underground cavity represented by a mountain tunnel, it will affect the groundwater environment in no small measure. For this reason, it is common to carry out numerical analysis (simulation) that predicts the groundwater environment according to the importance of the structure before and during the construction. For example, in tunnel construction, various measurements such as the amount of spring water from the face and the condition of the bedrock (ground) are obtained to obtain this information, and the above-mentioned numerical analysis is performed based on this information. A construction method (groundwater informatization construction) that will be useful for the next excavation work at the site will be applied.
Patent Document 1 discloses a technique for automatically estimating the amount of spring water in the ground when excavating a daily tunnel.

Japanese Patent No. 4196279

However, predicting the groundwater environment requires advanced knowledge and skills of analysis specialists (for example, technical evaluation of reproducible analysis / predictive analysis results, review method of analysis conditions, etc.), and much more. It takes a lot of time (for example, several days). In particular, when the groundwater environment is predicted during construction and the excavation of the next face is proceeded based on the prediction result, the prediction simulation may not catch up with the progress of the construction. This may affect the progress of construction.

The present invention has been made in view of such circumstances, and an object of the present invention is to provide a groundwater environment automatic prediction system and a groundwater environment automatic prediction method capable of promptly predicting the groundwater environment during tunnel construction. It is in.

In order to solve the above-mentioned problems, one aspect of the present invention includes a water permeation coefficient with respect to a permeation flow analysis system (10) for obtaining an estimated spring water amount which is an estimated spring water amount by performing a permeation flow analysis. The reproduction condition setting unit (301) for setting the reproduction condition, the estimated spring water amount acquisition unit (302) for acquiring the estimated spring water amount calculated based on the reproduction condition by the permeation flow analysis system (10), and the permeation. The spring water amount measurement value acquisition unit (303) that acquires the spring water amount measurement value that is the result of measuring the spring water amount generated at the face of the tunnel to be analyzed, and the estimated spring water amount are based on the spring water amount measurement value. A determination unit (304) that determines whether or not the data is within the predetermined range, a water permeability coefficient generation unit (306) that generates a different water permeability coefficient, and a predetermined range in which the estimated spring water amount is based on the spring water amount measurement value. If not, the reproduction analysis instruction unit (307) that causes the permeation flow analysis system (10) to perform permeation flow analysis under the reproduction conditions including the different water permeation coefficients, the water permeation coefficient used as the reproduction conditions, and the said. The learning unit (305) that generates a trained model by learning the relationship between the water permeability coefficient and the estimated spring water amount using the estimated spring water amount obtained by using the water permeation coefficient as teacher data, and the learning unit. By inputting the spring water volume measurement value measured at the wellhead side of the drilling position of the predicted target for which the spring water volume is to be predicted into the completed model, the predicted position water permeability coefficient, which is the water permeability coefficient at the drilling position of the predicted target, is obtained. By setting the coefficient acquisition unit (308) and the prediction conditions including the water permeability coefficient at the drilling position of the prediction target in the permeation flow analysis system (10) and performing the permeation flow analysis, the drilling position of the prediction target is subjected to. It has a predictive analysis unit (309) that obtains a predicted amount of spring water, which is the amount of spring water generated when excavation is performed.

Further, in one aspect of the present invention, the water permeability coefficient is relative to the permeation flow analysis system (10) in which the reproduction condition setting unit (301) obtains the estimated spring water amount, which is the spring water amount estimated by performing the permeation flow analysis. The reproduction condition including the above is set, and the estimated spring water amount acquisition unit (302) acquires the estimated spring water amount calculated based on the reproduction condition by the permeation flow analysis system (10), and the spring water amount measurement value acquisition unit ( 303) acquires the spring water amount measurement value which is the result of measuring the spring water amount generated in the face of the tunnel to be analyzed for the permeation flow analysis, and the determination unit (304) determines that the estimated spring water amount is the spring water amount measurement value. The reproduction analysis instruction unit (307) includes a different water permeation coefficient when the estimated spring water amount is not within the predetermined range based on the spring water amount measurement value. The permeation flow analysis as the reproduction condition was performed by the permeation flow analysis system (10), and the learning unit (305) used the water permeation coefficient as the reproduction condition and the estimated spring water amount obtained by using the water permeation coefficient. By learning the relationship between the water permeability coefficient and the estimated spring water amount, a trained model is generated, and the water permeability coefficient acquisition unit (308) predicts the spring water amount in the trained model. By inputting the spring water volume measurement value measured on the wellhead side of the drilling position of the predicted target, the predicted position water permeability coefficient, which is the water permeability coefficient at the drilling position of the predicted target, is obtained, and the prediction analysis unit (309) obtains the predicted position water permeability coefficient. , It occurs when excavation is performed at the excavation position of the prediction target by setting the prediction condition including the water permeability coefficient at the excavation position of the prediction target in the permeation flow analysis system (10) and causing the permeation flow analysis to be performed. This is an automatic groundwater environment prediction method that obtains the predicted amount of spring water.

As described above, according to the present invention, it is possible to promptly and automatically predict the groundwater environment during tunnel construction.

It is a figure explaining the whole flow of the groundwater environment prediction by one Embodiment of this invention. It is a schematic block diagram which shows the structure of the groundwater environment automatic prediction system 1. It is a flowchart explaining the operation in the groundwater environment automatic prediction system 1 when the reproduction analysis loop is executed. It is a flowchart explaining the operation in the groundwater environment automatic prediction system 1 when the prediction analysis loop is executed. It is a figure which shows the distribution state of the total head considering the predicted amount of spring water obtained by the predictive analysis loop with respect to the 3D geological structure model which represented the 3D shape of the ground to be constructed.

Hereinafter, the groundwater environment automatic prediction system according to the embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram illustrating an overall flow of groundwater environment prediction according to an embodiment of the present invention.
In the present embodiment, the groundwater environment prediction predicts the groundwater environment by performing the seepage flow analysis using the seepage flow analysis system and obtaining the amount of spring water.
When the permeation flow analysis system is used in the present embodiment, it is roughly divided into two processes (analysis loop) of "reproduction analysis loop" and "predictive analysis loop".

<< Reproduction analysis loop >>
The reproduction analysis loop is a process of reproducing the excavation status of the tunnel to be constructed on a simulation. In the reproduction analysis loop, for the infiltration flow analysis system, (1) infiltration characteristics of rock (ground), (2) rainfall, groundwater level, etc., (3) geological structure, (4) water stoppage / drainage measures, etc. Enter the reproduction conditions for the four. The osmotic flow analysis system performs osmotic flow analysis based on this reproduction condition. By performing osmotic flow analysis based on the reproduction conditions, it is possible to obtain an estimated value of the amount of spring water that is estimated at the face of the tunnel to be constructed. In this reproduction analysis loop, the reproduction conditions that roughly match the amount of spring water actually measured at the face of the tunnel to be constructed are obtained.
In determining the reproduction conditions, some preliminary surveys and observation sites will be used for (2) rainfall, groundwater level, etc., (3) geological structure, and (4) water stoppage / drainage measures. (1) The infiltration characteristics of the bedrock (ground) are determined by considering the type of bedrock (ground) in the face, the size of the crack, the direction of the crack, and the distribution on the surface of the face. It is necessary and not easy to determine. Therefore, in order to set the infiltration characteristics of the bedrock (ground), a high degree of judgment by an analysis specialist is usually required. In this embodiment, by obtaining the infiltration characteristics (particularly the permeability coefficient) of the bedrock (ground) using the groundwater environment automatic prediction device 30, the analysis expert does not have to judge the infiltration characteristics, and the prediction analysis is performed. It is possible to reduce the time required for analysis and the labor of analysis specialists.

《Predictive analytics loop》
The predictive analysis loop is used when the next excavation is carried out (for example, when excavation is performed for 10 m) using the reproduction conditions (especially the water permeability coefficient) in which a certain reproducibility is obtained for the tunnel to be constructed by the reproduction analysis loop. It is a process to predict the amount of spring water generated using a permeation flow analysis system.
In this way, by executing the reproduction analysis loop, a highly reproducible permeability coefficient is obtained, and by executing the prediction analysis loop using this obtained permeability coefficient, the amount of spring water when the position of the prediction target is excavated. Can be predicted accurately.

The predicted value of the amount of spring water obtained by the predictive analysis loop is displayed on the field terminal installed in the field office or the station. The predicted value of the amount of spring water is calculated and displayed at a timing such as once a day. As a result, on-site technicians and workers can grasp the predicted value of the amount of spring water that is sequentially updated according to the progress of excavation of the tunnel at shorter intervals than in the past. This can be utilized for danger prediction activities (KY activities) and review of drainage plans.

Next, FIG. 2 is a schematic block diagram showing the configuration of the groundwater environment automatic prediction system 1.
The groundwater environment automatic prediction system 1 includes a permeation flow analysis system 10, a reproduction condition storage unit 20, a groundwater environment automatic prediction device 30, and a field terminal 40.

The osmotic flow analysis system 10 stores a numerical analysis model for performing osmotic flow analysis. The permeation flow analysis system 10 inputs the above-mentioned reproduction conditions to this numerical analysis model, and performs permeation flow analysis for predicting the behavior of groundwater due to tunnel excavation or the like for the ground when excavating the tunnel to be constructed. By doing so, the amount of spring water is calculated. Here, the above-mentioned (3) geological structure is represented by a three-dimensional geological model. In the infiltration flow analysis system 10, the elements corresponding to each layer in this three-dimensional geological model include (1) physical properties corresponding to this layer (for example, physical properties such as permeability coefficient) based on the infiltration characteristics of the bedrock (ground). Value).
Here, the constituents constituting the stratum have a hydraulic conductivity according to the type of the constituents. It is known that such a hydraulic conductivity has a range of typical values for each type of constituent. The infiltration flow analysis system 10 corresponds to each layer in the three-dimensional geological model based on the range of typical values of the permeability coefficient based on (1) the infiltration characteristics of the rock (ground) input as the reproduction conditions. Assign a hydraulic conductivity to the element. Then, the seepage flow analysis system 10 can obtain the amount of spring water by excavating a tunnel on the numerical analysis model. The osmotic flow analysis system 10 can obtain the amount of spring water with high accuracy by appropriately assigning the osmotic characteristics (particularly the hydraulic conductivity). The existing system (software) generally used can be applied to the permeation flow analysis system 10.

The reproduction condition storage unit 20 stores the reproduction condition data to be reproduced as a model representing the ground of the target for which the infiltration flow analysis is performed. As mentioned above, the items included in the reproduction condition data include (1) infiltration characteristics of rock (ground), (2) rainfall, groundwater level, etc., (3) geological structure, and (4) water stoppage / drainage measures. And so on.
(1) The infiltration characteristics of the bedrock (ground) are data showing the characteristics of how much water the bedrock (ground) infiltrates according to the type. Generally, a technician who specializes in analysis sets the infiltration characteristics of rock (ground) based on knowledge, skills, etc., but in this embodiment, it is possible to set by the groundwater environment automatic prediction device 30. It has become.
(2) The amount of rainfall, the groundwater level, etc. are data indicating how much rainfall there is in the area to be analyzed, how high the groundwater level is, etc., and are set by the engineer. (3) The geological structure is a three-dimensional geological model that shows how the geological structure of the area where the tunnel is constructed is distributed three-dimensionally. It is set by the engineer based on the geological information obtained in the tunnel excavation process and the geological survey results as a result of conducting the survey in advance. (4) Water stoppage / drainage measures, etc. are based on the construction conditions of the tunnel, and if measures are taken to prevent spring water from occurring, the methods and conditions of the measures are shown. This water stoppage / drainage measures, etc. are data set by engineers.

The reproduction condition storage unit 20 is a storage medium, for example, an HDD (Hard Disk Drive), a flash memory, an EEPROM (Electrically Erasable Programmable ReadOnly Memory), a RAM (Random Access read / maid) It is composed of any combination of these storage media. For the reproduction condition storage unit 20, for example, a non-volatile memory can be used.

The groundwater environment automatic prediction device 30 includes a reproduction condition setting unit 301, an estimated spring water amount acquisition unit 302, a spring water amount measurement value acquisition unit 303, a determination unit 304, a learning unit 305, a permeability coefficient generation unit 306, a reproduction analysis instruction unit 307, and water permeability. It has a coefficient acquisition unit 308, a prediction analysis unit 309, and a prediction information providing unit 310.

The reproduction condition setting unit 301 reads the reproduction condition data including the hydraulic conductivity stored in the reproduction condition storage unit 20, and outputs the read reproduction condition data to the permeation flow analysis system 10.

The estimated spring water amount acquisition unit 302 acquires the spring water amount obtained from the infiltration flow analysis system 10 as the estimated spring water amount based on the reproduction conditions set by the reproduction condition setting unit 301. That is, the estimated spring water amount acquisition unit 302 acquires the spring water amount obtained by the infiltration flow analysis system 10 in the reproduction analysis loop.

The spring water amount measurement value acquisition unit 303 acquires the spring water amount measurement value which is the result of measuring the spring water amount generated at the face of the tunnel at the construction site to be subjected to the infiltration flow analysis. The spring water amount measurement value acquisition unit 303 acquires the spring water amount measurement value input based on the operation of the operator via an input device such as a keyboard or a mouse.
Further, the spring water amount measurement value acquisition unit 303 can also acquire the result measured by a sensor or the like instead of the operation input by the operator. For example, the spring water generated in the face is once stored in a large container (for example, a kettle) and then discharged to the outside of the mine by a pump. The discharge amount at this time is measured as a spring water amount by a sensor, and this measurement result can be acquired by the spring water amount measurement value acquisition unit 303. This makes it possible to automatically acquire the amount of spring water in the face.

The determination unit 304 determines whether or not the estimated spring water amount is within a predetermined range based on the spring water amount measurement value. The predetermined range may be predetermined.
When the estimated amount of spring water is within a predetermined range based on the measured value of the amount of spring water, it can be said that the estimated amount of spring water and the amount of spring water generated in the actual face are almost the same (high degree of similarity). In this case, it can be said that the set reproduction condition data has high reproducibility with respect to the actual situation of the region around the face, and therefore can be used in the predictive analysis loop. In this case, the hydraulic conductivity is also considered to be highly reproducible.
On the other hand, when the estimated amount of spring water is not within the predetermined range based on the measured value of the amount of spring water, the estimated amount of spring water and the amount of spring water generated in the actual face deviate from each other. In this case, the learning unit 305, which will be described later, resets / changes the hydraulic conductivity by learning so that the hydraulic conductivity is within a predetermined range based on the measured spring water volume.

The learning unit 305 uses the hydraulic conductivity and the estimated spring water amount as teacher data, and learns the relationship between the hydraulic conductivity and the estimated spring water volume so that the estimated spring water volume falls within a predetermined range based on the spring water volume measurement value. Generate a trained model that gives a good hydraulic conductivity.
For example, the learning unit 305 uses an ideal tunnel model in advance to discuss (1) rock (ground) infiltration characteristics, (2) rainfall, groundwater level, (3) geological structure, and (4) water stoppage / drainage measures. It has been learned. Then, the learning unit 305 performs additional learning based on the teacher data obtained in the construction process of the tunnel to be constructed, in addition to the trained model learned in advance. The learning unit 305 proceeds with learning so that the estimated spring water amount falls within a predetermined range based on the spring water amount measurement value by performing additional learning, and optimizes the trained model.
By optimizing the trained model, it is possible to obtain a trained model capable of obtaining a hydraulic conductivity such that the estimated spring water amount falls within a predetermined range based on the spring water amount measurement value.

The learning unit 305 repeatedly learns until the determination unit 304 determines that the estimated spring water amount is within a predetermined range based on the spring water amount measurement value.

The learning performed by the learning unit 305 is learning using AI (artificial intelligence), and may be any of machine learning, deep learning, and deep reinforcement learning. Further, it may be a rule-based AI such as an expert system.

The permeability coefficient generation unit 306 generates a permeability coefficient different from the permeability coefficient that has been set in the permeation flow analysis system 10 by the reproduction condition setting unit 301. When the hydraulic conductivity generation unit 306 generates the hydraulic conductivity, the water permeability coefficient generation unit 306 considers that the estimated spring water amount becomes close to the spring water amount measurement value according to the difference between the estimated spring water amount based on the learning status of the learning unit 305 and the spring water amount measurement value. It may be made to generate a hydraulic conductivity to be obtained.

When the estimated spring water amount is not within a predetermined range based on the spring water amount measurement value, the reproduction analysis instruction unit 307 writes the water permeability coefficient generated by the water permeability coefficient generation unit 306 to the reproduction condition storage unit 20 as a reproduction condition. , Have the permeability analysis system perform permeability analysis. Here, when a new water permeability coefficient is written in the reproduction condition storage unit 20 by the reproduction condition setting unit 301, it is assumed that the reproduction condition setting unit 301 has been instructed to perform reproduction analysis, and the reproduction condition data including the new water permeability coefficient is included. Can be input to the osmotic flow analysis system 10 to perform osmotic flow analysis.

The hydraulic conductivity acquisition unit 308 wants to predict the spring water volume in a trained model optimized so that the hydraulic conductivity can be obtained so that the estimated spring water volume falls within a predetermined range based on the spring water volume measurement value. By inputting the measured value of the amount of spring water measured in the vicinity) or outside the mine, the predicted hydraulic conductivity at the drilling position to be predicted is obtained.
Here, the position where the amount of spring water is to be predicted is the position where excavation is advanced by a certain distance from the current face. This distance is, for example, the distance traveled in the construction process of one excavation, and if the rock (ground) condition is the same to some extent, the permeability coefficient at the current face and the permeability of the rock (ground) at the desired position The coefficients are considered to be in good agreement.

The prediction analysis unit 309 sets the prediction conditions including the prediction position permeability coefficient in the seepage flow analysis system 10 and causes the seepage flow analysis to perform the infiltration flow analysis, so that the amount of spring water generated when excavation is performed at the excavation position to be predicted. Get the predicted amount of spring water. The predictive analysis unit 309 uses the trained model trained in the learning unit 305 until the determination unit 304 determines that the estimated spring water volume is within a predetermined range based on the spring water volume measurement value, and the predicted position permeability coefficient. Is acquired, and the permeation flow analysis is performed by outputting the predicted position permeability coefficient to the permeation flow analysis system 10 as a prediction condition.
For example, the predictive analysis unit 309 outputs the predicted position permeability coefficient to the seepage flow analysis system 10, and the amount of spring water when excavated to a certain distance from the current face position at the construction site (for example, when excavated 10 m). Is obtained as the predicted amount of spring water.

The prediction information providing unit 310 outputs the predicted spring water amount obtained from the permeation flow analysis system 10 by the prediction analysis unit 309 to the on-site terminal 40.

The site terminal 40 is provided at either a station where a tunnel is constructed, a site office, or the like, and is used by engineers, workers, and the like. The number of field terminals 40 may be plurality. As the field terminal 40, at least one of a computer, a tablet, a smartphone, and the like may be used.

Next, the operation of the groundwater environment automatic prediction system 1 will be described.
FIG. 3 is a flowchart illustrating the operation in the groundwater environment automatic prediction system 1 when the reproduction analysis loop is executed.
First, the groundwater environment automatic prediction system 1 performs a reproduction analysis loop process. In the reproduction analysis loop processing, a rock (ground) model for excavating a mountain tunnel is created (step S101), and the created rock (ground) model is stored in the reproduction condition storage unit 20 as (3) a geological structure. Further, the parameters related to the infiltration characteristics (for example, the permeability coefficient) are stored in the reproduction condition storage unit 20 as (1) the infiltration characteristics of the bedrock (ground) (step S102). A plurality of types of parameters related to the permeation characteristics may be prepared in advance and stored in the reproduction condition storage unit 20. In addition to this, (2) rainfall, groundwater level, etc., (3) geological structure, (4) water stoppage / drainage measures, etc. are also stored in the reproduction condition storage unit 20.
When various parameters are stored in the reproduction condition storage unit 20, the reproduction condition setting unit 301 sets the reproduction condition data stored in the reproduction condition 20 in the permeation flow analysis system 10. When the reproduction condition data is set, the permeation flow analysis system 10 performs permeation flow analysis (forward analysis) (step S103).

The estimated spring water amount acquisition unit 302 acquires the spring water amount obtained by the osmotic flow analysis as the estimated spring water amount (step S104). The determination unit 304 compares the estimated spring water amount with the spring water amount measurement value acquired by the spring water amount measurement value acquisition unit 303 (step S105).
When the deviation between the estimated spring water amount and the spring water amount measurement value is large (the estimated spring water amount is not within the predetermined range based on the spring water amount measurement value), the learning unit 305 sets the estimated spring water amount obtained by the permeation flow analysis. , The water permeability coefficient is learned as teacher data (step S106), and the trained model is optimized. This learning proceeds to step S102 until it is determined that the discrepancy between the estimated spring water amount and the spring water amount measurement value is not large (the estimated spring water amount is within a predetermined range based on the spring water amount measurement value), and is different. Osmotic flow analysis is performed using the permeability coefficient.
On the other hand, in step S105, when the determination unit 304 determines that the discrepancy between the estimated spring water amount and the spring water amount measurement value is not large (the estimated spring water amount is within a predetermined range based on the spring water amount measurement value). The process proceeds to the execution of the predictive analysis loop (step S107).
Learning can be performed by repeating the above-mentioned reproduction analysis loop during tunnel construction, learning effects corresponding to the site are accumulated, and more accurate judgment can be made as excavation progresses. You can build a model.

Here, as the excavation of the tunnel proceeds, the (3) geological structure stored as a reproduction condition may not match the geological structure observed in the current face. In such a case, (3) the geological structure will be modified based on the results of observing the current face or the results of the geological survey obtained during tunnel construction such as forward exploration of the face. As a result, the 3D geological model is also updated. Then, a reproduction analysis loop can be executed based on the updated geological structure.

FIG. 4 is a flowchart illustrating the operation in the groundwater environment automatic prediction system 1 when the prediction analysis loop is executed.
In the predictive analysis loop, the water permeability coefficient acquisition unit 308 inputs the spring water amount measurement value measured at the face site and obtained by the spring water amount measurement value acquisition unit 303 into the optimized trained model (step S201). The water permeability coefficient corresponding to the infiltration characteristics of the bedrock (ground) from which the measured value of the amount of spring water can be obtained is obtained (step S202).
The predictive analysis unit 309 inputs the obtained permeability coefficient to the permeation flow analysis system 10 to perform permeation flow analysis (step S203), and is the amount of spring water generated when excavation is performed at the excavation position to be predicted. Obtain the predicted amount of spring water (step S204).
The prediction information providing unit 310 outputs the predicted spring water amount to the on-site terminal 40 (step S205).

FIG. 5 is a diagram showing the distribution of all heads in consideration of the predicted amount of spring water obtained by the predictive analysis loop for a three-dimensional geological structure model that three-dimensionally represents the topography and geological structure of the ground to be constructed. Is. For example, the result shown in this figure is output to the field terminal 40. When the result shown in this figure is obtained, for example, by excavating a tunnel, it can be grasped that the groundwater level is lowered in the traveling direction of excavation with reference to the face. By conducting such analysis on a daily basis, it is possible to predict the amount of spring water that will occur in the near future and carry out construction safely.

In the above-described embodiment, the measured value of the amount of spring water cannot be obtained because the excavation is not performed at the stage before the construction starts (the first infiltration flow analysis). That is, there is no information obtained from learning. In such a case, based on the rock type / permeability coefficient database stored in advance by associating the relationship between the rock type and the permeability coefficient, the reproduction analysis loop is not executed, only the predictive analysis loop is executed, and the groundwater is grounded. Predict the groundwater environment inside. Then, the amount of spring water may be obtained from the result of the predictive analysis loop.

According to the embodiment described above, the following effects can be obtained.
(A) Originally, the judgments made by analysis specialists regarding the water permeability coefficient can be ruled, indexed, and implemented in the system as AI technology to automate the analytical judgments required for reproduction analysis and predictive analysis. be able to. This allows the entire system to be fully automated.
(B) Since it is fully automated, it can be used for simulation work even outside business hours when analysis specialists cannot perform normal work, and real-time prediction is possible.
(C) In the "reproduction analysis loop", the reproduction conditions (1) to (4) (especially (1) infiltration characteristics of the bedrock (ground)), which are the analysis conditions, are sequentially updated based on the judgment of AI, which is the core technology. Introduce a mechanism to do. This not only realizes highly accurate reproduction analysis, but also makes it possible to output highly accurate prediction results.
(D) By fully automating, it is possible to significantly reduce the amount of work involved in operating groundwater computerized construction such as predictive analysis.
(E) The number of analysis specialists can be reduced, and the specialists can focus on more specialized work.
(F) Because the groundwater environment automatic prediction device 30 can determine the optimum hydraulic conductivity of the bedrock (ground) according to the amount of spring water measured by an analysis expert without making an empirical judgment. , It is always possible to predict. As a result, real-time performance can be improved, and countermeasures can be promptly planned and discussed for spring water that will occur in the near future.
(G) It is possible to reduce tunnel face disasters caused by spring water.
(H) It is possible to carry out a more rational drainage plan than before.

The groundwater environment automatic prediction device 30 in the above-described embodiment may be realized by a computer. In that case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read by a computer system and executed. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. Further, the "computer-readable recording medium" refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, or a CD-ROM, and a storage device such as a hard disk built in a computer system. Further, a "computer-readable recording medium" is a communication line for transmitting a program via a network such as the Internet or a communication line such as a telephone line, and dynamically holds the program for a short period of time. It may also include a program that holds a program for a certain period of time, such as a volatile memory inside a computer system that is a server or a client in that case. Further, the above program may be for realizing a part of the above-mentioned functions, and may be further realized for realizing the above-mentioned functions in combination with a program already recorded in the computer system. It may be realized by using a programmable logic device such as FPGA (Field Programmable Gate Array).

Although the embodiments of the present invention have been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and includes designs and the like within a range that does not deviate from the gist of the present invention.

1 ... Groundwater environment automatic prediction system, 10 ... Permeability analysis system, 20 ... Reproduction condition storage unit, 30 ... Groundwater environment automatic prediction device, 40 ... On-site terminal, 301 ... Reproduction condition setting unit, 302 ... Estimated spring water amount acquisition unit, 303 ... Spring water volume measurement value acquisition unit, 304 ... Judgment unit, 305 ... Learning unit, 306 ... Water permeability coefficient generation unit, 307 ... Reproduction analysis instruction unit, 308 ... Water permeability coefficient acquisition unit, 309 ... Prediction analysis unit, 310 ... Prediction information Providing department

Claims (4)

A reproduction condition setting unit that sets reproduction conditions including the permeability coefficient for an osmotic flow analysis system that obtains an estimated spring water amount that is an estimated spring water amount by performing osmotic flow analysis.
An estimated spring water amount acquisition unit that acquires an estimated spring water amount calculated based on the reproduction conditions by the osmotic flow analysis system, and an estimated spring water amount acquisition unit.
A spring water volume measurement value acquisition unit that acquires a spring water volume measurement value that is the result of measuring the spring water volume generated at the face of the tunnel to be subjected to the osmotic flow analysis, and a spring water volume measurement value acquisition unit.
A determination unit for determining whether or not the estimated spring water amount is within a predetermined range based on the spring water amount measurement value, and a determination unit.
A permeability coefficient generator that generates different permeability coefficients,
A reproduction analysis instruction unit that causes the osmotic flow analysis system to perform osmotic flow analysis under the reproduction conditions including the different permeability coefficients when the estimated spring water amount is not within a predetermined range based on the spring water amount measurement value.
A trained model by learning the relationship between the hydraulic conductivity and the estimated hydraulic conductivity by using the hydraulic conductivity used as the reproduction condition and the estimated hydraulic conductivity obtained by using the hydraulic conductivity as teacher data. And the learning department to generate
By inputting the spring water volume measurement value measured at the wellhead side of the excavation position of the prediction target for which the spring water amount is to be predicted into the trained model, the predicted position permeability coefficient, which is the permeability coefficient at the excavation position of the prediction target, can be obtained. The hydraulic conductivity acquisition unit to be obtained and
It is the amount of spring water generated when excavation is performed at the excavation position of the prediction target by setting the prediction condition including the water permeability coefficient at the excavation position of the prediction target in the permeation flow analysis system and performing the permeation flow analysis. An automatic groundwater environment prediction system with a prediction analysis unit that obtains the predicted amount of spring water. The groundwater environment automatic prediction system according to claim 1, wherein the learning unit repeatedly learns until the determination unit determines that the estimated spring water amount is within a predetermined range based on the spring water amount measurement value. The groundwater environment automatic prediction system according to claim 1 or 2, further comprising a prediction information providing unit that outputs the predicted spring water amount obtained by the prediction analysis unit to a site terminal. The reproduction condition setting unit sets the reproduction conditions including the permeability coefficient for the osmotic flow analysis system that obtains the estimated spring water amount, which is the estimated spring water amount by performing the osmotic flow analysis.
The estimated spring water amount acquisition unit acquires the estimated spring water amount calculated based on the reproduction conditions by the osmotic flow analysis system.
The spring water amount measurement value acquisition unit acquires the spring water amount measurement value which is the result of measuring the spring water amount generated at the face of the tunnel to be analyzed for the infiltration flow.
The determination unit determines whether or not the estimated spring water amount is within a predetermined range based on the spring water amount measurement value, and determines whether or not the estimated spring water amount is within a predetermined range.
The reproduction analysis instruction unit causes the osmotic flow analysis system to perform osmotic flow analysis under the reproduction conditions including different hydraulic conductivity when the estimated spring water amount is not within a predetermined range based on the spring water amount measurement value.
The learning unit uses the permeability coefficient used as the reproduction condition and the estimated spring water amount obtained by using the water permeability coefficient as teacher data, and learns the relationship between the permeability coefficient and the estimated spring amount. , Generate a trained model,
The permeability coefficient acquisition unit inputs the measured value of the spring water amount measured at the wellhead side of the excavation position of the prediction target for which the spring water amount is to be predicted into the trained model, so that the permeability coefficient at the excavation position of the prediction target is obtained. Obtaining a certain predicted position permeability coefficient,
When the prediction analysis unit sets the prediction conditions including the hydraulic conductivity at the excavation position of the prediction target in the permeation flow analysis system and causes the permeation flow analysis to perform excavation at the excavation position of the prediction target. An automatic groundwater environment prediction method that obtains the predicted amount of spring water that will be generated.

Images