Tunnel construction management system and method

Abstract

To achieve accurate prediction of a spring state after blasting without depending on a builder while achieving control of crushed bedrock and reduction of the amount of explosives in tunnel excavation work.SOLUTION: Th present invention comprises a 3D scanner for acquiring 3D tunnel shape data (S1), a noise and vibration sensor (S2) that acquires noise and vibration data in tunnels, a water volume sensor that acquires spring water volume data in the tunnel (S3), a thermography camera (S4) that acquires heat distribution image data of a face, and a processor (102-107) for inputting a first drilling pattern and explosive amount employed in a blasting process, three-dimensional tunnel shape data, noise and vibration data, spring water volume data, and heat distribution image data, estimating spring water locations and water volume in the face from heat distribution image data, and using machine learning models (M1-M3) to predict at least a second drilling pattern and explosive mass to be employed in a next blasting process.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a construction management system for excavating tunnels for roads, railways, waterways, etc. [Background technology]

[0002] When excavating a tunnel, a construction method appropriate to the condition of the ground is used, but in mountainous areas with hard soil, the blasting method is used. In the blasting method, a drill jumbo is used to drill many charge holes in the tunnel face, and explosives with detonators attached are inserted into the charge holes and detonated to excavate the tunnel face, followed by excavating the tunnel wall and tunnel floor to the required shape, after which the tunnel face is excavated flat and the next charge hole is drilled. This process is repeated to proceed with the tunnel excavation work.

[0003] In the blasting process, the pattern of drilling charge holes in the tunnel face is important for controlling the rock fragmentation form, reducing the amount of explosives, and avoiding large amounts of water seepage. An example of a method for designing a drilling pattern is disclosed in, for example, Patent Document 1. According to Patent Document 1, the drilling pattern is designed by a designer interacting with a computer while inputting or selecting the necessary information.

[0004] In particular, if groundwater is present in front of the tunnel face, a large amount of water will spring up during the blasting process, so knowing the state of the groundwater in front of the face in advance is extremely important for safely carrying out tunnel construction. Patent Document 2 discloses a system that identifies areas where water will spring up and predicts or evaluates the state of groundwater from temperature distribution images of the face acquired using an infrared camera. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Special Publication No. 2010-513757 [Patent Document 2] JP 2010-101799 A Summary of the Invention [Problem to be solved by the invention]

[0006] However, in the system disclosed in Patent Document 1, the design of the drilling pattern is based on the designer's experience. In addition, the system disclosed in Patent Document 2 predicts or evaluates the state of groundwater from the temperature distribution of the face, and cannot accurately predict the state of water seepage after blasting. In this way, the background art cannot accurately predict the state of water seepage after blasting, and it is difficult to automate it, and variations due to designers or construction workers cannot be avoided.

[0007] The present invention has been devised in consideration of the above-mentioned circumstances, and an object of the present invention is to provide a tunnel construction management system and method that can control the rock crushing pattern and reduce the amount of explosives used in tunnel excavation work, while accurately predicting the state of water welling up after blasting without relying on the contractor. [Means for solving the problem]

[0008] In order to achieve the above-mentioned object, according to one embodiment of the present invention, there is provided a tunnel construction management system comprising: a three-dimensional scanner for acquiring three-dimensional tunnel shape data; a noise and vibration sensor for acquiring noise and vibration data in the tunnel; a water volume sensor for acquiring data on the amount of spring water in the tunnel; a thermographic camera for acquiring thermal distribution image data of a tunnel face; and at least one processor for inputting a first drilling pattern and an amount of explosives adopted in a blasting process, the three-dimensional tunnel shape data, the noise and vibration data, the spring water volume data, and the thermal distribution image data, and predicting at least a second drilling pattern and an amount of explosives to be adopted in a next blasting process, wherein the processor estimates the location and amount of spring water at the tunnel face from the thermal distribution image data, inputs the tunnel shape change acquired from the three-dimensional tunnel shape data, the geological conditions acquired from the noise and vibration data, the water volume change acquired from the spring water volume data, the location and amount of spring water at the tunnel face, the first drilling pattern, and the amount of explosives, and predicts the second drilling pattern and the amount of explosives using a machine learning model. Furthermore, according to one embodiment of the present invention, the processor can identify areas of the thermal distribution image data of the face that are displayed with a density equal to or higher than a predetermined reference level as spring water locations, and estimate the amount of water based on the density in the area. Moreover, according to one embodiment of the present invention, the system further includes a memory unit for storing time-series tunnel shape data obtained from the three-dimensional tunnel shape data, geological condition data obtained from the noise and vibration data, time-series water volume data obtained from the spring water volume data, and time-series data of the spring water location and water volume, and the time-series tunnel shape data, the geological condition data, the time-series water volume data, and the time-series data of the spring water location and water volume stored in the memory unit can be used for machine learning of the machine learning model. According to one embodiment of the present invention, the machine learning model may include a drilling pattern and explosive amount prediction model, a tunnel shape prediction model, and a spring water location and water amount prediction model. In order to achieve the above-mentioned object, according to one embodiment of the present invention, there is provided a construction management method for tunnel excavation work, comprising the steps of: a 3D scanner acquires 3D tunnel shape data; a noise and vibration sensor acquires noise and vibration data within the tunnel; a water volume sensor acquires spring water volume data within the tunnel; and a thermographic camera acquires thermal distribution image data of the tunnel face; and at least one processor estimates the location and volume of spring water at the tunnel face from the thermal distribution image data, and uses a machine learning model to predict a second drilling pattern and an amount of explosive to be adopted in at least the next blasting process based on the tunnel shape changes acquired from the 3D tunnel shape data, the geological conditions acquired from the noise and vibration data, the water volume changes acquired from the spring water volume data, the location and volume of spring water at the tunnel face, and the first drilling pattern and amount of explosive. Furthermore, according to one embodiment of the present invention, the processor can identify areas of the thermal distribution image data of the face that are displayed with a density equal to or higher than a predetermined reference level as spring water locations, and estimate the amount of water based on the density in the area. Furthermore, according to one embodiment of the present invention, time-series tunnel shape data obtained from the three-dimensional tunnel shape data, geological condition data obtained from the noise and vibration data, time-series water volume data obtained from the spring water volume data, and time-series data of the spring water location and water volume are stored in a memory unit, and the processor can use the time-series tunnel shape data, the geological condition data, the time-series water volume data, and the time-series data of the spring water location and water volume stored in the memory unit for machine learning of the machine learning model. According to one embodiment of the present invention, the machine learning model may include a drilling pattern and explosive amount prediction model, a tunnel shape prediction model, and a spring water location and water amount prediction model. Effect of the Invention

[0009] According to one embodiment of the present invention, changes in tunnel shape, geological conditions, changes in the amount of spring water, the location and amount of spring water at the face, the previous drilling pattern and amount of explosives are input, and the second drilling pattern and amount of explosives are predicted using a machine learning model, which is advantageous in achieving control of the rock fracture pattern and reduction in the amount of explosives in tunnel excavation work, while also achieving accurate prediction of the spring water condition after blasting without relying on the contractor. Furthermore, according to one embodiment of the present invention, the location and amount of water springing up are estimated from thermal distribution image data of the tunnel face obtained by a thermographic camera, which is advantageous in achieving accurate prediction of the water springing condition after blasting without relying on the contractor. Furthermore, according to one embodiment of the present invention, by accumulating time-series tunnel shape data, geological condition data, time-series water volume data, and time-series data on spring locations and water volumes, the data can be used for machine learning of a machine learning model, enabling quick and accurate prediction of drilling patterns and explosive amounts. In addition, according to one embodiment of the present invention, by using a tunnel shape prediction model, a water volume prediction model, and a spring water location and volume prediction model in addition to the drilling pattern and explosive amount prediction model as machine learning models, it is possible to predict tunnel shape, water volume, and spring water location and volume using the predicted drilling pattern and explosive amount, thereby improving prediction accuracy. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1A is a cross-sectional view of a tunnel showing the drilling process of a charge hole in a tunnel excavation work to which a tunnel construction management system according to one embodiment of the present invention is applied, and FIG. 1B is a cross-sectional view of a tunnel showing the shape measurement process after blasting. [Diagram 2] 1 is a block diagram showing an example of a functional configuration of a tunnel construction management system according to the present embodiment. FIG. [Diagram 3] 1 is a schematic diagram for explaining the function of predicting the location and water volume of spring water based on a thermal distribution image of the tunnel construction management system according to this embodiment. FIG. [Figure 4]2 is a flowchart showing an example of a tunnel construction management method according to the present embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, a tunnel construction management system according to one embodiment of the present invention will be described in detail, taking the tunnel excavation work shown in FIG. 1 as an example.

[0012] As shown in FIG. 1, a 3D scanner S1, a noise and vibration sensor S2, a water volume sensor S3, and a thermography camera S4 are provided in the tunnel 10. The 3D scanner S1 is mounted on a movable cart and can acquire a three-dimensional shape in the vicinity of the face 12. A plurality of noise and vibration sensors S2 are arranged at predetermined intervals from a position at a predetermined distance from the face 12, and each noise and vibration sensor S2 detects elastic waves during blasting or rock drilling, and the geological conditions in front of the face and on the side of the tunnel can be predicted from the detection signal (see, for example, JP 2014-106075 A). The water volume sensor S3 is provided in the drainage equipment 21 that discharges water that springs from the tunnel inner wall 11 and the face 12 to the outside of the tunnel, and measures the amount of discharged spring water. The thermography camera S4 is an infrared camera that uses infrared rays to acquire a thermal distribution image of the face 12, and is mounted on the same cart as the 3D scanner S1 here.

[0013] First, as shown in Fig. 1(A), the 3D scanner S1 obtains the three-dimensional shape near the tunnel face 12, and the thermographic camera S4 obtains a thermal distribution image of the tunnel face 12, and then they retract to the rear, after which the drill jumbo 22 drills a large number of charge holes 13 in the tunnel face 12 in a pattern to be described later. After drilling the charge holes, the drill jumbo 22 retracts to the rear, and explosives with detonators attached are inserted into each charge hole 13, which are then detonated to excavate the tunnel face 12.

[0014] After the blasting, a backhoe or a rock drill is used to excavate the tunnel wall into a predetermined cylindrical shape and the tunnel floor almost flat, and as shown in Fig. 1(B), a 3D scanner S1 captures the 3D shape of the tunnel face 12 and its vicinity after the blasting, and a thermographic camera S4 captures a thermal distribution image of the tunnel face 12 after the blasting. Thereafter, the tunnel 10 can be excavated by excavating the tunnel face 12 almost flat after the blasting, and repeating the above-mentioned process of drilling the charge holes, charging the charge, and blasting.

[0015] According to this embodiment, as described later, data acquired by each of the sensors S1 to S4 in the above-mentioned past processes is accumulated. Therefore, the next drilling pattern and explosive amount are determined using the previous drilling pattern and explosive amount, the time-dependent change in the three-dimensional shape detected by the 3D scanner S1, the geological exploration results detected by the noise and vibration sensor S2, the time-dependent change in the amount of spring water detected by the water volume sensor S3, and the time-dependent change in the thermal distribution image of the face 12 detected by the thermography camera S4, and the next blasting location, amount of spring water, and tunnel shape can be predicted.

[0016] The configuration and operation of the tunnel construction management system according to this embodiment will be described below with reference to Figures 2 to 4. The tunnel construction management system according to this embodiment can be built on a general-purpose server or a personal computer having one or more processors such as a CPU (Central Processing Unit) and an MPU (Micro Processor Unit).

[0017] 1. System Configuration As shown in Fig. 2, a tunnel construction management system 100 according to an embodiment of the present invention has an input unit 101 for inputting various data. The input unit 101 is connected by wire or wireless to a 3D scanner S1, a noise and vibration sensor S2, a water volume sensor S3, and a thermographic camera S4 installed in the tunnel, and inputs 3D scan data, noise and vibration data, spring water volume data, and thermal distribution image data, respectively. The input unit 101 is also connected to input devices such as a keyboard and a pointing device of the system, and in this embodiment, data such as a drilling pattern and explosive amount can be input.

[0018] The tunnel construction management system 100 has a control unit 102 that controls the operation of the entire system, and the control unit 102 is connected to an input unit 101, a face shape extraction unit 103, a tunnel side surface shape extraction unit 104, a geological exploration unit 105, a thermal distribution image analysis unit 106, an AI (Artificial Intelligence) analysis unit 107, a data storage unit (108 to 112), and an output unit 113. The face shape extraction unit 103 and the tunnel side surface shape extraction unit 104 extract the face shape and the tunnel side surface shape, respectively, from the 3D scan data acquired by the 3D scanner S1. The geological exploration unit 105 predicts the geological condition of the natural ground from the noise and vibration data acquired by the noise and vibration sensor S2.

[0019] The thermal distribution image analysis unit 106 analyzes the thermal distribution image data of the tunnel face 12 input from the thermographic camera S4, classifies the temperature area according to a predetermined temperature level standard, and estimates the location of spring water and its amount at the tunnel face 12 as described later. The thermal distribution image analysis unit 106 can also be configured using a machine learning model.

[0020] The data storage unit includes a shape data storage unit 108, a geological condition data storage unit 109, a water volume data storage unit 110, a spring water location and water volume estimation data storage unit 111, and a construction result storage unit 112. The shape data storage unit 108 stores the face shape and tunnel side shape data acquired by the face shape extraction unit 103 and the tunnel side shape extraction unit 104 in chronological order. The geological condition data storage unit 109 stores the geological condition data predicted by the geological exploration unit 105 based on the noise and vibration data acquired by the noise and vibration sensor S2 during blasting. The water volume data storage unit 110 stores the water volume data acquired by the water volume sensor S3 in chronological order. The spring water location and water volume estimation data storage unit 111 stores the spring water location and water volume estimation data acquired by the thermal distribution image analysis unit 106 by analyzing the thermal distribution image data of the face 12 in chronological order. The construction result storage unit 112 stores pre-construction predictions and construction results, and the stored data can be used to build a machine learning model in the AI ​​analysis unit 107.

[0021] The AI ​​analysis unit 107 inputs the drilling pattern and explosive amount data, the geological condition, time series shape, time series water volume, and time series estimated spring water location and water volume stored in the data storage units 108 to 112, and predicts the drilling pattern and explosive amount, tunnel shape, spring water location and water volume using the drilling pattern and explosive amount prediction model M1, tunnel shape prediction model M2, and spring water location and water volume prediction model M3, which are machine learning models. For example, the drilling pattern and explosive amount prediction model M1 can calculate the predicted value of the next drilling pattern and explosive amount using the previous drilling pattern and explosive amount, geological condition, tunnel shape, and spring water location and water volume change as input data. In addition, the tunnel shape prediction model M2 can predict the tunnel shape by the next blast using the current tunnel shape, geological condition, and next drilling pattern and explosive amount as input data. The water spring location and water volume prediction model M3 can predict the location and volume of water springs due to the next blasting using the next drilling pattern and explosive volume, the predicted tunnel shape, the geological conditions, and time-series data on the location and volume of water springs.

[0022] These machine learning models M1 to M3 can be constructed by, for example, deep learning. Deep learning technology can construct a machine learning model with a sufficiently small error by providing appropriate teacher data and adjusting neuron parameters even when a large number of data are mutually influencing each other. Each of the machine learning models M1 to M3 can be constructed using predetermined teacher data, but can also be constructed using data obtained in past blasting processes, namely, drilling pattern and explosive amount data, geological situation data, changes in tunnel shape over time, changes in spring water volume over time, and changes in spring water location and spring water volume over time.

[0023] The output unit 113 is a display unit such as a computer monitor, and displays the next drilling pattern and amount of explosives calculated by the AI ​​analysis unit 107, the tunnel shape resulting from the next blast, and the predicted location and amount of water springing.

[0024] The functions of the control unit 102, the face shape extraction unit 103, the tunnel side shape extraction unit 104, the geological exploration unit 105, the thermal distribution image analysis unit 106, and the AI ​​analysis unit 107 can be realized by executing programs stored in a memory (not shown) on a processor.

[0025] 2. Estimation and prediction of spring locations and spring volumes As shown in FIG. 3, the thermal distribution image analysis unit 106 analyzes the thermal distribution image IMG of the working face 12 input from the thermographic camera S4, and divides it into temperature regions according to a predetermined density level 106a. Here, it is assumed that the regions in the thermal distribution image IMG with a density higher than a predetermined reference level are identified as spring water regions 12a, 12b, and 12c. In this embodiment, it is assumed that the spring water from the working face 12 has a lower temperature than the surroundings, and the lower the temperature region is displayed in a darker blue color. Conversely, if the spring water is at a higher temperature than the surroundings, the higher the temperature region is displayed in a darker red color. In any case, it is possible to identify the high density region of the thermal distribution image as the spring water location.

[0026] The thermal distribution image analysis unit 106 further calculates the area of ​​the spring water area with a concentration higher than a predetermined reference level and the maximum concentration value in that area to estimate the amount of spring water. The amount of spring water is estimated to be larger as the area of ​​the spring water area is larger and the concentration is higher. The position information of the spring water area estimated in this way and the estimated water amount are stored in the spring water location / amount estimation data accumulation unit 111.

[0027] As described later, the spring water location / volume prediction model M3 of the AI ​​analysis unit 107 can estimate the amount of spring water from the tunnel face 12 after blasting based on geological condition data, changes in the tunnel shape over time, the amount of spring water and its changes over time, and the estimated spring water location and amount.

[0028] 3.Operation Hereinafter, the operation of the tunnel construction management system l00 in this embodiment will be described with reference to FIG.

[0029] As shown in Fig. 4, the control unit 102 judges whether or not it is the first blasting (operation 201). If it is the first blasting (YES in operation 201), the face shape and the tunnel side shape are extracted from the 3D scan data acquired from the 3D scanner S1 and stored in the shape data storage unit 108, and the spring water location and water volume are estimated from the thermal distribution image data of the face 12 acquired from the thermographic camera S4 and stored in the spring water location / water volume estimation data storage unit 111 (operation 202). The face shape and the tunnel side shape are extracted by the face shape extraction unit 103 and the tunnel side shape extraction unit 104. In addition, water volume data is acquired from the water volume sensor S3 and stored in the water volume data storage unit 109 (operation 203).

[0030] Next, the drilling pattern and the amount of explosives are set through the input device based on the results of the natural ground exploration carried out in advance (operation 204), and blasting is carried out (operation 205). Noise and vibration data during blasting are acquired by the noise and vibration sensor S2, and the geological exploration unit 105 predicts the geological conditions from the noise and vibration data, and the result is stored in the geological conditions data storage unit 109 (operation 206).

[0031] The control unit 102 extracts the tunnel face shape and the tunnel side shape from the 3D scan data acquired from the 3D scanner S1 installed in front of the tunnel face after blasting, and stores them in the shape data storage unit 108, and stores the spring water location and water volume estimated from the thermal distribution image data of the tunnel face 12 after blasting acquired from the thermographic camera S4 in the spring water location / water volume estimated data storage unit 111 (operation 207). In addition, water volume data is acquired from the water volume sensor S3 and stored in the water volume data storage unit 110 (operation 208). As described above, after the tunnel face 12 after blasting is excavated almost flat, the next blasting is performed as follows.

[0032] The AI ​​analysis unit 107 inputs the previous drilling pattern and explosive amount as well as various data from the data storage unit, and predicts the drilling pattern and explosive amount, the tunnel shape, and the spring water location and amount using the drilling pattern / explosive amount prediction model M1, the tunnel shape prediction model M2, and the spring water location / amount prediction model M3 (operation 209).

[0033] First, the AI ​​analysis unit 107 reads out time-series data on the tunnel face shape and tunnel side shape stored in the shape data storage unit 108, geological condition data stored in the geological condition data storage unit 109, time-series water volume data stored in the water volume data storage unit 110, and time-series data on the spring water location and water volume stored in the spring water location / water volume estimation data storage unit 111.

[0034] Next, the AI ​​analysis unit 107 calculates the changes over time in the tunnel face shape and the tunnel side shape from the time series data of the tunnel face shape and the tunnel side shape, and calculates the changes over time in the amount of spring water from the water volume data.

[0035] Finally, the AI ​​analysis unit 107 predicts the drilling pattern and explosive amount in the next blasting using the drilling pattern / explosive amount prediction model M1, predicts the tunnel face shape and tunnel side shape due to the next blasting using the tunnel shape prediction model M2, and further predicts the location and amount of water that will spring up due to the next blasting using the water spring location / water amount prediction model M3.

[0036] The control unit 102 displays the drilling pattern and explosive amount predicted by the AI ​​analysis unit 107, the predicted spring water location and spring water amount, and the predicted tunnel shape on the output unit 113 (operation 210).

[0037] Next, blasting is carried out with the displayed drilling pattern and explosive amount (operation 211). Noise and vibration data during blasting are acquired by the noise and vibration sensor S2, and the geological exploration unit 105 predicts the geological conditions from the noise and vibration data and stores the results in the geological conditions data storage unit 109 (operation 212).

[0038] The control unit 102 acquires 3D scan data and thermal distribution image data from the 3D scanner S1 and thermography camera S4 installed in front of the tunnel face after blasting, extracts the tunnel face shape and tunnel side shape from the 3D scan data after blasting, and stores them together with the error from the prediction in the shape data storage unit 108, and stores the estimated values ​​of the spring water location and water volume from the thermal distribution image data after blasting and the error from the predicted value in the spring water location / water volume estimated data storage unit 111 (operation 213). Also, water volume data is acquired from the water volume sensor S3, and stored in the water volume data storage unit 110 together with the error from the prediction (operation 214).

[0039] In this way, the blasting process is repeated (NO in operation 215) according to operations 209 to 215 until construction is completed (YES in operation 215). That is, the constructor drills the charge holes 13 in the face 12 using the drill jumbo 22 according to the drilling pattern displayed on the display unit 113, and charges each charge hole 13 with the displayed amount of explosives to perform blasting. The tunnel construction management system according to this embodiment can measure the tunnel shape and spring water volume after blasting and reflect them in the prediction for the next blasting process. Alternatively, if the error from the prediction is larger than a predetermined value, the model can be automatically adjusted by the learning process. [Explanation of symbols]

[0040] 10. Tunnel 11 Tunnel Inner Wall 12 Cutting Edge 13 Charge hole 21 Drainage equipment 22 Drill Jumbo 100 Tunnel construction management system 101 Input section 102 Control section 103 Face shape extraction section 104 Tunnel side shape extraction section 105 Geological Exploration Department 106 Heat distribution image analysis section 107 AI analysis department 108 Shape data storage unit 109 Geological Condition Data Storage Department 110 Water volume data storage unit 111 Thermal distribution image data storage unit 112 Construction result storage section 113 Output section M1 Perforation Pattern and Explosive Amount Prediction Model M2 Tunnel shape prediction model M3 Spring location and water volume prediction model S1 3D Scanner S2 Noise and Vibration Sensor S3 Water Sensor S4 Thermal Imaging Camera

Claims

1. A tunnel construction management system, a three-dimensional scanner for acquiring three-dimensional tunnel shape data; A noise and vibration sensor that collects noise and vibration data inside the tunnel; A water volume sensor that acquires data on the amount of spring water in the tunnel; A thermography camera that captures thermal distribution image data of the tunnel face; At least one processor that inputs a first drilling pattern and an amount of explosives used in a blasting process, the three-dimensional tunnel shape data, the noise and vibration data, the water volume data, and the thermal distribution image data, and predicts a second drilling pattern and an amount of explosives to be used in at least a next blasting process; having The processor, Estimating the location and amount of water at the face from the thermal distribution image data; inputting the tunnel shape change obtained from the three-dimensional tunnel shape data, the geological conditions obtained from the noise and vibration data, the water volume change obtained from the spring water volume data, the spring water location and water volume at the face, and the first drilling pattern and explosive amount, and predicting the second drilling pattern and explosive amount using a machine learning model; A tunnel construction management system characterized by:

2. The tunnel construction management system described in claim 1, characterized in that the processor identifies areas of the thermal distribution image data of the face that are displayed with a density above a predetermined standard level as spring water locations, and estimates the amount of water based on the density in the area.

3. The system further includes a memory unit for storing time-series tunnel shape data obtained from the three-dimensional tunnel shape data, geological condition data obtained from the noise and vibration data, time-series water volume data obtained from the spring water volume data, and time-series data on the spring water location and water volume, A tunnel construction management system as described in claim 1 or 2, characterized in that the time-series tunnel shape data, the geological condition data, the time-series water volume data, and the time-series data of the spring location and water volume stored in the memory unit are used for machine learning of the machine learning model.

4. The tunnel construction management system according to any one of claims 1 to 3, characterized in that the machine learning model comprises a drilling pattern and explosive amount prediction model, a tunnel shape prediction model, and a spring water location and water volume prediction model.

5. A construction management method for tunnel excavation work, comprising: A 3D scanner acquires 3D tunnel shape data, Noise and vibration sensors collect noise and vibration data inside the tunnel, The water level sensor collects data on the amount of water that springs up inside the tunnel. A thermographic camera captures thermal distribution image data of the tunnel face, At least one processor Estimating the location and amount of water at the face from the thermal distribution image data; Using a machine learning model, predict a second drilling pattern and an amount of explosive to be adopted in at least a next blasting process based on the tunnel shape change obtained from the three-dimensional tunnel shape data, the geological conditions obtained from the noise and vibration data, the water volume change obtained from the spring water volume data, the spring water location and water volume at the face, and the first drilling pattern and explosive amount; A tunnel construction management method comprising:

6. The tunnel construction management method described in claim 5, characterized in that the processor identifies areas of the thermal distribution image data of the face that are displayed with a concentration above a predetermined standard level as spring water locations, and estimates the amount of water based on the concentration in the area.

7. storing, in a storage unit, time-series tunnel shape data acquired from the three-dimensional tunnel shape data, geological condition data acquired from the noise and vibration data, time-series water volume data acquired from the spring water volume data, and time-series data on the spring water location and water volume; The tunnel construction management method described in claim 5 or 6, characterized in that the processor uses the time-series tunnel shape data, the geological condition data, the time-series water volume data, and the time-series data of the spring location and water volume stored in the memory unit for machine learning of the machine learning model.

8. The tunnel construction management method according to any one of claims 5 to 7, characterized in that the machine learning model comprises a drilling pattern and explosive amount prediction model, a tunnel shape prediction model, and a spring water location and water volume prediction model.

Images