KR20220147393A - Prediction method for thrust parameters of tbm and excavation monitoring system providing the same - Google Patents

Abstract

The present invention relates to a prediction method for thrust parameters of the tunnel boring machine (referred to as TBM hereinafter) and an excavation monitoring system providing the method thereof, which can predict the thrust parameter with respect to a unit excavation distance which is to be subsequently excavated by the TBM, each time the TBM excavates by a unit excavation distance. To this end, the excavation monitoring system of the present invention can predict the thrust parameter corresponding to the N+1^th unit excavation distance when the excavation with respect to the Nth unit excavation distance is completed. The present invention comprises a pretreatment system which assumes the ground parameter corresponding to the Nth unit excavation distance based on a first aggravated parameter which corresponds to the distance from the location of the Nth unit excavation distance to the front boring hole, a second aggravated parameter which corresponds to the distance from the location of the Nth unit excavation distance to the rear boring hole, a front ground parameter including the geographical information of the front boring hole location, and a rear ground parameter including the geographical information of the rear boring hole location. The pretreatment system produces an excavation data set including the thrust parameter and the ground parameter with respect to each of the first unit excavation distance and the Nth unit excavation distance. Therefore, the thrust parameters of the TBM may be precisely predicted.

Description

PREDICTION METHOD FOR THRUST PARAMETERS OF TBM AND EXCAVATION MONITORING SYSTEM PROVIDING THE SAME

In the present invention, whenever the tunnel boring machine (hereinafter referred to as 'TBM') excavates by the unit excavation distance, the thrust parameter (Thrust) for the unit excavation distance to be excavated next by the tunnel boring machine (TBM) Parameter) relates to a method for predicting thrust parameters of TBM and an excavation monitoring system providing the method.

The TBM (Tunnel Boring Machine) method is a mechanical excavation method in which a tunnel is excavated by rotating a cutter head in the front of the excavator, and the tunnel wall (segment lining) is assembled in advance and excavated.

The TBM method can shorten the construction period by showing an excavation rate 3 to 5 times higher than the NATM (New Austrain Tunneling Method) method, which has been widely used in the past, and has fewer civil complaints due to low vibration and low noise. It has the advantage of being able to protect nature, such as preventing the water level from dropping. Accordingly, the application of the TBM method, which is an eco-friendly and economical tunnel construction method, is expanding worldwide in urban tunnels, undersea tunnels, and mountain tunnels.

Meanwhile, in the design stage of tunnel excavation, the contractor predicts the construction period, establishes a construction plan based on the predicted construction period, and calculates the budget. However, in actual construction sites, the construction period is frequently exceeded due to unexpected geographical conditions (rock crushing, faults, cracks, high water pressure in the case of undersea tunnels, etc.) are doing

In order to solve this problem, it is possible to predict the thrust parameters of the tunnel boring machine (TBM) in real time according to the geographical conditions, and adjust the actual thrust of the tunnel boring machine (TBM) in consideration of the predicted thrust parameters and the construction period. there should be For example, if the construction period exceeds one month when excavation is carried out with the thrust corresponding to the predicted thrust parameter, the tunnel boring machine (TBM) is operated with a higher thrust than the thrust corresponding to the predicted thrust parameter and the construction period can be adjusted to match.

Recently, a cyclic neural network-based model that predicts thrust parameters in real time has been developed at home and abroad. However, these models reflect the past thrust parameters, but do not take into account the geographic conditions that affect the actual thrust, that is, the geographic parameters, and thus have a low precision of prediction. For example, it is suitable to operate a tunnel boring machine (TBM) with a relatively high thrust for rock with high strength or no cracks, and for ground with severe weathering or cracks (closed section) with relatively low thrust. It is appropriate to operate a tunnel boring machine (TBM), but this geographic condition is not reflected in the prediction of thrust parameters.

Accordingly, the thrust parameter of the tunnel boring machine (TBM) for the excavation distance to be excavated after considering all ground data as well as the thrust data collected in real time during excavation construction (e.g., every time a unit excavation distance is excavated) ( There is a great need for a system that predicts thrust parameters with high precision.

Japanese registered patent 3260629 Japanese Patent 3721486 Japanese Patent 3821538

The present invention is to provide an excavation monitoring system that provides a method and method for predicting thrust parameters of a TBM capable of constructing an excavation data set corresponding to each of the unit excavation distances from the starting position where the excavation is performed to the current position.

An object of the present invention is to provide a TBM thrust parameter prediction method for predicting the thrust parameter of a tunnel boring machine (TBM) with high precision, and an excavation monitoring system providing the method.

The excavation monitoring system according to one aspect of the present invention is an excavation monitoring system for predicting a thrust parameter corresponding to an N+1-th unit excavation distance when excavation for an N-th unit excavation distance is completed, the N-th unit excavation a first weighting parameter corresponding to the distance from the position of the distance to the front borehole, a second weighting parameter corresponding to the distance from the position of the N-th unit excavation distance to the rear borehole, and a front comprising geographic information of the position of the forward borehole A preprocessing system for estimating a ground parameter corresponding to the N-th unit excavation distance based on a ground parameter and a rear ground parameter including geographic information of the location of the rear borehole, wherein the pre-treatment system includes a first unit excavation distance To each of the N-th unit excavation distance, an excavation data set including the thrust parameter and the ground parameter is generated.

When the excavation for the N-th unit excavation distance according to another feature of the present invention is completed, in the excavation monitoring system for predicting a thrust parameter corresponding to the N+1-th unit excavation distance, a thrust prediction model for predicting the thrust parameter A model generating unit generating N pieces, a model selecting unit selecting a thrust prediction model showing the best performance among the N thrust prediction models, and a thrust corresponding to the N+1th unit excavation distance through the selected thrust prediction model a prediction system including a thrust prediction unit for predicting parameters, wherein the model generation unit determines N corresponding to the number of generated thrust prediction models by random search, and whenever the thrust prediction model is generated, the thrust prediction The hyperparameter that determines the performance of the model is determined by random search.

In the method of predicting the thrust parameter of TBM according to another feature of the present invention, when the excavation for the N-th unit of excavation distance is completed, the method of predicting the thrust parameter corresponding to the N+1-th unit of excavation distance, the pre-processing system , estimating a ground parameter corresponding to the N-th unit excavation distance, and generating an excavation data set including the thrust parameter and the ground parameter for each of the first unit excavation distance to the N-th unit excavation distance, prediction generating, by the system, a thrust prediction model for predicting a thrust parameter corresponding to the N+1-th unit excavation distance, and by substituting, by the prediction system, input data selected from the excavation data set, into the thrust prediction model and predicting a thrust parameter corresponding to an N+1th unit excavation distance.

According to the present invention, it is possible to accurately estimate the ground parameters corresponding to the excavation section, which could not be directly collected through a conventional borehole.

According to the present invention, the thrust parameter corresponding to the unit excavation distance to be excavated next can be accurately predicted by reflecting the ground parameter.

According to the present invention, the driver (manager) of the tunnel boring machine TBM can efficiently manage (operate) the tunnel boring machine TBM based on the predicted thrust parameter.

The present invention has the effect of improving the excavation performance of the tunnel boring machine (TBM) to prevent the extension of the construction period and the occurrence of additional costs.

1 is an exemplary view for explaining a site in which excavation is carried out according to an embodiment.
FIG. 2 is a view for explaining the structure of the segment lining 2 shown in FIG. 1 .
3 is a block diagram illustrating an excavation monitoring system according to an embodiment.
FIG. 4 is a conceptual diagram for explaining the function of the preprocessing system of FIG. 3 .
5 is a flowchart illustrating a method of predicting a thrust parameter of a TBM according to an embodiment.
6 is a flowchart illustrating in detail the thrust prediction model generation step ( S200 ) of FIG. 5 .

Hereinafter, the embodiments disclosed in the present specification will be described in detail with reference to the accompanying drawings, but identical or similar components are given the same and similar reference numerals, and overlapping descriptions thereof will be omitted. The suffixes “module” and/or “part” for the components used in the following description are given or mixed in consideration of only the ease of writing the specification, and do not have distinct meanings or roles by themselves. In addition, in describing the embodiments disclosed in the present specification, if it is determined that detailed descriptions of related known technologies may obscure the gist of the embodiments disclosed in the present specification, the detailed description thereof will be omitted. In addition, the accompanying drawings are only for easy understanding of the embodiments disclosed in the present specification, and the technical spirit disclosed herein is not limited by the accompanying drawings, and all changes included in the spirit and scope of the present invention , should be understood to include equivalents or substitutes.

Terms including ordinal numbers such as first, second, etc. may be used to describe various elements, but the elements are not limited by the terms. The above terms are used only for the purpose of distinguishing one component from another.

When a component is referred to as being “connected” or “connected” to another component, it may be directly connected or connected to the other component, but it is understood that other components may exist in between. it should be On the other hand, when it is said that a certain element is "directly connected" or "directly connected" to another element, it should be understood that the other element does not exist in the middle.

In the present application, terms such as “comprises” or “have” are intended to designate that a feature, number, step, operation, component, part, or combination thereof described in the specification exists, but one or more other features It is to be understood that this does not preclude the possibility of the presence or addition of numbers, steps, operations, components, parts, or combinations thereof.

1 is an exemplary view for explaining a site in which excavation is performed according to an embodiment, and FIG. 2 is a view for explaining the structure of the segment lining 2 shown in FIG. 1 .

Referring to Figure 1, the TBM (Tunnel Boring Machine) method, the tunnel boring machine (Tunnel Boring Machine, hereinafter referred to as 'TBM') (1) every time a tunnel is excavated by a unit excavation distance, formed of an annular block It is a mechanical excavation method in which the tunnel wall, that is, the segment lining (2) is assembled into the tunnel and excavated.

Referring to FIG. 2 , the segment lining 2 may be configured as an annular ring having a predetermined width. If a plurality of segment linings 1, 2, ..., N-2, N-1, N are installed in the tunnel, the tunnel can maintain its shape. According to an embodiment, when the tunnel boring machine (TBM) excavates by a unit excavation distance, the process of installing the unit segment lining 2 may be repeated. In this case, the unit excavation distance may correspond to a predetermined width of the unit segment lining (2).

Thereafter, with reference to FIGS. 3 to 6 , an excavation data set (thrust parameter) corresponding to each of the unit excavation distances from the start position where the excavation is performed to the current position (1, 2, ..., N-2, N-1, N) and the ground parameter) and the second embodiment for predicting the thrust parameter corresponding to the unit excavation distance (N+1) to be excavated next, every time the tunnel boring machine (TBM) excavates by the unit excavation distance It can be performed repeatedly.

3 is a block diagram illustrating an excavation monitoring system according to an embodiment, and FIG. 4 is a conceptual diagram illustrating a function of the pretreatment system of FIG. 3 .

Referring to FIG. 3 , the excavation monitoring system includes a pretreatment system 100 , and a prediction system 200 .

When the excavation for the N-th unit excavation distance is completed, the pre-processing system 100 estimates the ground parameter corresponding to the N-th unit excavation distance, and for each of the first unit excavation distance to the N-th unit excavation distance, the thrust parameter and an excavation data set including ground parameters.

The thrust parameter includes information on the thrust of the tunnel boring machine (TBM), and the ground parameters include geographic conditions (uniaxial compressive strength, rock quality index, rock classification, deformation coefficient, absorption rate, P and S wave velocity). etc.) are included.

Referring to FIG. 3 , it is assumed that the tunnel boring machine TBM starts excavation in the first borehole H1 and completes the excavation for a unit excavation distance corresponding to the current N-th unit segment lining position. That is, the N-th unit excavation distance may correspond to a current position where the tunnel boring machine (TBM) has completed excavation. In addition, the N+1th unit excavation distance (not shown) may correspond to a future position where the tunnel boring machine TBM will perform the next excavation in the direction of the second borehole H2.

The ground parameter can be obtained by analyzing the ground obtained by drilling a borehole (H) in the ground surface (L). For example, in the tunnel construction planning stage, ground parameters can be acquired through a plurality of boreholes (H1-HN), but this is limited, and there is a practical limit to directly acquiring the ground parameters corresponding to each unit excavation distance. . In comparison with this, the thrust parameter can be directly acquired every time a tunnel boring machine (TBM) excavates a unit excavation distance. Accordingly, the thrust parameter corresponding to each unit excavation distance can be directly acquired, but the ground data corresponding to each unit excavation distance cannot be directly acquired in many cases.

According to an embodiment, the pre-processing system 100 includes a front ground parameter (LDH1) of the front borehole (H1) closest to the current location, a rear ground parameter (LDH2) of the rear borehole (H2), and the front and rear boreholes ( A ground parameter corresponding to the current location is estimated based on the first and second weighting parameters W1 and W2 corresponding to respective distances up to H1 and H2. For example, the pre-processing system 100 may estimate the ground parameter corresponding to the current location as a value corresponding to the ratio of the distance to the ground parameter (LDH1, LDH2) of each of the two boreholes (H1, H2).

According to one embodiment, the pretreatment system 100, when the excavation for the N-th unit excavation distance is completed, for each of the first unit excavation distance to the N-th unit excavation distance, the N-th including a thrust parameter and a ground parameter Create an excavation data set.

For example, the pre-processing system 100, when the excavation for the first unit excavation distance is completed, estimates the first ground data, and a first excavation data set (first thrust data and the first excavation data set for the first unit excavation distance) 1 ground data).

For another example, the pre-processing system 100, when the excavation for the second unit excavation distance is completed, estimates the second ground data, and generates the second excavation data (the second thrust data and the second ground data) . And, the pre-processing system 100 generates a second excavation data set (first excavation data and second excavation data) including a thrust parameter and a ground parameter for each of the first unit excavation distance and the second unit excavation distance. can

For another example, the pre-processing system 100, when the excavation for the third unit excavation distance is completed, estimates the third ground data, and generates the third excavation data (the third thrust data and the third ground data) do. And, the pre-processing system 100, a third excavation data set (first excavation data, second excavation data, and third excavation data).

The prediction system 200 includes a model generation unit 210 , a model selection unit 230 , and a thrust prediction unit 250 , and a thrust parameter corresponding to an N+1th unit excavation distance based on the excavation data set. predict

The model generator 210 may generate N thrust prediction models for predicting a thrust parameter. The model generation unit 210 may determine N corresponding to the number of generated thrust prediction models by random search, and determine hyperparameters that determine the performance of the thrust prediction model each time a thrust prediction model is generated by random search. . Then, a plurality of (N) thrust prediction models having different values of the hyperparameters (p, d, q, r) to be described below may be generated.

According to an embodiment, the thrust prediction model may be configured as an ARIMAX model, which is one of various time series analysis techniques. For example, the thrust prediction model may be composed of an ARIMAX model corresponding to Equation (1) below.

- 식(1)

- Equation (1)

Here, each of X and t is the exogenous parameter and the order of the exogenous parameter, and Y and Ψ are the target parameter and the order of the target parameter, respectively. [epsilon] and [theta] are the order of error and error, respectively. P corresponds to the input number of the target parameter, d corresponds to the differential value of the target parameter, q corresponds to the number of errors (term), and r corresponds to the input number of the exogenous parameter. Here, p, d, q, and r are hyperparameters that are determined by the user's setting, and the prediction performance of the ARIMAX model varies according to the setting. According to an embodiment, the exogenous parameter (X) may correspond to a ground parameter, and the target parameter (Y) may correspond to a thrust parameter, respectively.

For example, if p=3 and r=2, three thrust parameters (N-th thrust parameter, N-1 thrust parameter, N-2 thrust parameter) and two ground parameters (N-th ground parameter and N-th thrust parameter) -1 ground parameter) value can be input data of ARIMAX model. In this case, the output data may be a thrust parameter value corresponding to the N+1th unit excavation distance.

The model selection unit 230 may select a thrust prediction model showing the best performance among the N thrust prediction models. For example, the model selection unit 230 may select a thrust prediction model corresponding to the lowest AIC value among Akaike Information Criterion (AIC) values of each of the N thrust prediction models. AIC is a relative indicator that can compare the fit between statistical models, and the smaller the value, the higher the prediction performance.

The thrust prediction unit 250 selects input data from the excavation data set based on the hyperparameter of the thrust prediction model selected as the best prediction performance, and substitutes the selected input data into the selected thrust prediction model to the N+1th unit excavation distance It is possible to predict the thrust parameter corresponding to .

For example, if p=2 and r=1 of the hyperparameters of the thrust prediction model (ARIMAX model) selected as the best predictive performance, the Nth thrust parameter, the N-1th thrust parameter, and the Nth ground parameter value are the thrust prediction model It can be input data of (ARIMAX model). In this case, the output data may be a thrust parameter value corresponding to the N+1th unit excavation distance. For another example, if p = 1 and r = 1 of the hyperparameter of the thrust prediction model (ARIMAX model) selected as the best prediction performance, the Nth thrust parameter and the Nth ground parameter value are the input of the thrust prediction model (ARIMAX model) data can be In this case, the output data may be a thrust parameter value corresponding to the N+1th unit excavation distance.

5 is a flowchart illustrating a method of predicting a thrust parameter of a TBM according to an embodiment, and FIG. 6 is a flowchart illustrating in detail the thrust prediction model generation step (S200) of FIG. 5 .

Hereinafter, with reference to FIGS. 3 to 6, a method for predicting a thrust parameter of a TBM and an excavation monitoring system providing the method will be described.

5, the pre-processing system 100 estimates the ground parameter corresponding to the N-th unit excavation distance, and includes a thrust parameter and a ground parameter for each of the first unit excavation distance to the N-th unit excavation distance Create an excavation data set (S100).

The preprocessing system 100 includes a first weighting parameter W1 corresponding to the distance from the position of the N-th unit excavation distance to the front borehole H1, and the distance from the position of the N-th unit excavation distance to the rear borehole H2. The Nth unit excavation is based on the corresponding second weighted parameter W2, the front ground parameter including geographic information of the front borehole H1 location, and the rear ground parameter including the geographic information of the rear borehole H2 location. A ground parameter corresponding to the distance may be estimated. For example, the preprocessing system 100 may estimate the ground parameter corresponding to the current location as a value corresponding to the ratio of the distance to the ground parameter value of each of the boreholes H1 and H2.

According to one embodiment, the pretreatment system 100, when the excavation for the N-th unit excavation distance is completed, for each of the first unit excavation distance to the N-th unit excavation distance, the N-th including a thrust parameter and a ground parameter Create an excavation data set.

Next, the prediction system 200 generates a thrust prediction model for predicting a thrust parameter corresponding to the N+1th unit excavation distance (S200).

Referring to FIG. 6 , in step S200, the prediction system 200 determines N corresponding to the number of generated thrust prediction models (S210), and determines the hyperparameter for determining the performance of the thrust prediction model by random search. (S230).

In step S200, the prediction system 200 generates a thrust prediction model corresponding to the determined hyperparameter (S250).

In step S200, the prediction system 200 determines whether the total number of the generated thrust prediction models coincides with the N (S270).

If the determination result does not match (S270, No), the prediction system 200 continues to generate the thrust prediction model by repeatedly performing from step S230. If the determination result agrees (S270, Yes), the prediction system 200 selects the thrust prediction model showing the best performance among the N thrust prediction models (S290).

For example, the prediction system 200 may select a thrust prediction model corresponding to the lowest AIC value among Akaike Information Criterion (AIC) values of each of the N thrust prediction models. AIC is a relative indicator that can compare the fit between statistical models, and the smaller the value, the higher the prediction performance.

Next, the prediction system 200 predicts a thrust parameter corresponding to the N+1th unit excavation distance by substituting the input data selected from the excavation data set into the selected thrust prediction model (S200).

The prediction system 200 selects input data from the excavation data set based on the hyperparameter of the selected thrust prediction model, and substitutes the selected input data into the selected thrust prediction model, a thrust parameter corresponding to the N+1th unit excavation distance can be predicted

According to an embodiment, when the excavation for the N-th unit excavation distance is completed, the pre-processing system 100 generates an N-th data set (including the N-th ground data and the N-th thrust data) for the N-th unit excavation distance. And, the prediction system 200 may predict the thrust parameter corresponding to the N+1th unit excavation distance based on the Nth data set through the thrust prediction model generated at the position corresponding to the Nth unit excavation distance. Then, when the excavation for the N+1-th unit excavation distance is completed, the pre-processing system 100 provides an N+1-th data set (N+1th ground data and N+th ground data) for the N+1-th unit excavation distance. 1, including thrust data), and the prediction system 200 excavates the N+2th unit based on the N+1th data set through the thrust prediction model generated at the position corresponding to the N+1th unit excavation distance. The thrust parameter corresponding to the distance can be predicted.

That is, in one embodiment, the excavation monitoring system generates a thrust prediction model every time a unit excavation distance is excavated, and repeats the process of predicting only the thrust parameters corresponding to the immediately next excavation distance through the generated thrust prediction model.

According to another embodiment, when the excavation for the N-th unit excavation distance is completed, the pre-processing system 100 generates an N-th data set (including the N-th ground data and the N-th thrust data), and the prediction system 200 may predict a thrust parameter corresponding to the N+1th unit excavation distance based on the Nth data set.

Then, the pre-processing system 100 estimates the ground parameter corresponding to the N+1th unit excavation distance, even if there is no excavation for the N+1th unit excavation distance, and estimates the N+1th unit through the thrust prediction model. An N+1-th data set (including an estimated ground parameter and a predicted thrust parameter corresponding to the N+1-th unit excavation distance) may be generated based on the thrust parameter value corresponding to the unit excavation distance. Then, the prediction system 200 estimates the thrust parameter corresponding to the N + 2 unit excavation distance through the thrust prediction model generated at the position corresponding to the N th unit excavation distance based on the N + 1 th data set. can In the same way, the prediction system 200 corresponds to the thrust parameter corresponding to the N+3th unit excavation distance, and the N+4th unit excavation distance through the thrust prediction model generated at the position corresponding to the Nth unit excavation distance. The thrust parameters can be estimated.

That is, in another embodiment, the excavation monitoring system generates a thrust prediction model at the position where the N-th unit excavation distance is excavated, and the N+1th excavation distance as well as the N+2th excavation distance immediately immediately through the generated thrust prediction model. It is possible to estimate a plurality of thrust parameters for a plurality of future excavation distances such as the excavation distance, the N+3th excavation distance, and the N+4th excavation distance. In this case, the thrust parameter, which is output data of the thrust prediction model, may be used as input data of the thrust prediction model for estimating the thrust parameter for the next excavation distance.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements by those of ordinary skill in the field to which the present invention belongs are also entitled to the rights of the present invention. belong to the scope

Claims (15)

In the excavation monitoring system for predicting the thrust parameter corresponding to the N+1th unit excavation distance when the excavation for the Nth unit excavation distance is completed,
A first weighting parameter corresponding to the distance from the position of the Nth unit drilling distance to the front borehole, a second weighting parameter corresponding to the distance from the position of the Nth unit drilling distance to the rear borehole, and a geographic location of the front borehole location. A preprocessing system for estimating a ground parameter corresponding to the N-th unit excavation distance based on a front ground parameter including information and a back ground parameter including geographic information of the location of the rear borehole,
The pretreatment system is
For each of the first unit excavation distance to the N-th unit excavation distance, an excavation monitoring system for generating an excavation data set including the thrust parameter and the ground parameter. According to claim 1,
Further comprising a prediction system for predicting a thrust parameter corresponding to the N+1-th unit excavation distance based on the excavation data set,
The prediction system is
A model generator for generating N thrust prediction models for predicting the thrust parameters;
A model selection unit for selecting a thrust prediction model showing the best performance among the N thrust prediction models, and
Excavation monitoring system, including a thrust prediction unit for predicting a thrust parameter corresponding to the N+1-th unit excavation distance through the selected thrust prediction model. 3. The method of claim 2,
The model generation unit,
Determine N corresponding to the number of generated thrust prediction models by random search,
Whenever the thrust prediction model is generated, a hyperparameter for determining the performance of the thrust prediction model is determined by random search. 4. The method of claim 3,
The model generation unit,
An excavation monitoring system for generating the N thrust prediction models corresponding to each of the hyper parameters. 3. The method of claim 2,
The model selection unit,
An excavation monitoring system for selecting a thrust prediction model corresponding to the lowest AIC value among the Akaike Information Criterion (AIC) values of each of the N thrust prediction models. 5. The method of claim 4,
The thrust prediction unit,
selecting input data from the excavation data set based on the hyperparameter of the selected thrust prediction model;
An excavation monitoring system for predicting a thrust parameter corresponding to the N+1th unit excavation distance by substituting the selected input data into the selected thrust prediction model. In the excavation monitoring system for predicting the thrust parameter corresponding to the N+1th unit excavation distance when the excavation for the Nth unit excavation distance is completed,
A model generation unit for generating N thrust prediction models for predicting the thrust parameters, a model selection unit for selecting a thrust prediction model showing the best performance among the N thrust prediction models, and the first through the selected thrust prediction model A prediction system including a thrust prediction unit for predicting a thrust parameter corresponding to an N+1 unit excavation distance,
The model generation unit,
Determine N corresponding to the number of generated thrust prediction models by random search,
Whenever the thrust prediction model is generated, a hyperparameter for determining the performance of the thrust prediction model is determined by random search. 8. The method of claim 7,
The model generation unit,
An excavation monitoring system for generating the N thrust prediction models corresponding to each of the hyper parameters. 8. The method of claim 7,
The model selection unit,
An excavation monitoring system for selecting a thrust prediction model corresponding to the lowest AIC value among the Akaike Information Criterion (AIC) values of each of the N thrust prediction models. In the method of predicting a thrust parameter corresponding to the N+1th unit excavation distance when the excavation for the Nth unit excavation distance is completed,
The pre-processing system estimates the ground parameter corresponding to the N-th unit excavation distance, and generates an excavation data set including the thrust parameter and the ground parameter for each of the first unit excavation distance to the N-th unit excavation distance step,
generating, by the prediction system, a thrust prediction model for predicting a thrust parameter corresponding to the N+1-th unit excavation distance; and
The prediction system, including the step of predicting a thrust parameter corresponding to the N+1-th unit excavation distance by substituting the input data selected from the excavation data set into the thrust prediction model, TBM thrust parameter prediction method. 11. The method of claim 10,
The step of generating the excavation data set comprises:
A first weighting parameter corresponding to the distance from the position of the Nth unit drilling distance to the front borehole, a second weighting parameter corresponding to the distance from the position of the Nth unit drilling distance to the rear borehole, and a geographic location of the front borehole location. A method of estimating a ground parameter corresponding to the N-th unit excavation distance based on a front ground parameter including information and a rear ground parameter including geographic information of the location of the rear borehole. 12. The method of claim 11,
The generating of the thrust prediction model comprises:
determining N corresponding to the number of generated thrust prediction models through random search;
determining the hyperparameter for determining the performance of the thrust prediction model through random search;
generating a thrust prediction model corresponding to the determined hyper parameter;
Determining whether the total number of the generated thrust prediction model coincides with the N, and
If the determination result agrees, the TBM thrust parameter prediction method comprising the step of selecting a thrust prediction model showing the best performance among the N thrust prediction models. 13. The method of claim 12,
If the determination result does not match, the TBM thrust parameter prediction method proceeds to the step of determining the hyperparameter through a random search. 13. The method of claim 12,
The step of selecting the thrust prediction model showing the best performance is,
A thrust parameter prediction method of TBM for selecting a thrust prediction model corresponding to the lowest AIC value among the Akaike Information Criterion (AIC) values of each of the N thrust prediction models. 14. The method of claim 13,
Predicting the thrust parameter corresponding to the N+1th unit excavation distance comprises:
selecting input data from the excavation data set based on the hyperparameter of the selected thrust prediction model;
By substituting the selected input data into the selected thrust prediction model to predict a thrust parameter corresponding to the N+1-th unit excavation distance, the TBM thrust parameter prediction method.

Images