Steering control device for shielded TBM and steering control method using the same

Abstract

The present invention provides a steering control device and control method for a shield tunneling machine (TBM) that can perform precise steering control along the designed excavation path without relying on human judgment. [Solution] The steering control method using the steering control device for a shield TBM according to the present invention individually controls the stroke length of at least four folding jacks that are provided between a forward shield equipped with a cutter head and an intermediate shield and are spaced radially at equal intervals, and the steering control device includes the steps of: extracting the reference point coordinates of the cutter head at the current location and the reference point coordinates of the cutter head at the target location based on one cycle of operation of the shield jacks; calculating the stroke length of each folding jack based on the reference point coordinates of the current location and the reference point coordinates of the target location using an artificial intelligence calculation model; and simultaneously operating each folding jack according to the calculated stroke length.

Description

[Technical Field]

[0001] The present invention relates to a steering control device for a shield tunneling machine (TBM) and a steering control method using the same. More specifically, the present invention relates to a steering control device for a shield tunneling machine (TBM) and a steering control method using the same, wherein the steering control device uses an artificial intelligence calculation model to calculate the stroke length of each articulated jack based on the reference point coordinates of the current location and the reference point coordinates of the target location, and operates each articulated jack simultaneously according to the calculated stroke length, thereby ensuring high accuracy even when excavating curves. [Background technology]

[0002] Unlike the NATM method which uses explosives, the Shield TBM method involves deploying a cylindrical shield machine into a temporary work tunnel. The cutter head attached to the front of the machine is rotated to excavate the tunnel face, and at the same time, segment rings are assembled inside, thereby achieving early ground stabilization and high quality.

[0003] As shown in Figure 1a, the cylindrical shield machine includes a front shield, a middle shield, a shield jack connecting them, and an articulation jack for excavation. The shield jack is used to operate the front shield directly or indirectly for straight or nearly straight excavation, moving the cutter head forward by a depth corresponding to the width of the segment ring during the excavation process. When the tunnel alignment consists of straight lines and sharp curves, an articulation jack located between the front shield and the middle shield is used in conjunction with the shield jack. In this case, the articulation jack is changed to correspond to a predetermined articulation angle, and then the shield jack is extended to perform excavation.

[0004] Furthermore, as shown in Figure 1b, the folding system can be divided into V-type and X-type depending on the position of the folding point by the folding jack. (a) In the X-type, the rotating pin is located at the folding point and connects the forward shield and the intermediate shield, so the folding angle is small, but it is advantageous for folding seal and waterproofing. (b) In the V-type, the folding point is located on the shield skin plate, so the folding angle is large, but it has the characteristic of being less effective at preventing the inflow of excavated soil and water.

[0005] Korean Published Patent Publication No. 2022-0033323, "TBM Operating System" (published March 16, 2022, hereinafter referred to as "Prior Art Document") discloses a technical configuration for monitoring and controlling the status of a TBM in real time. In this configuration, when a user inputs an operating signal for the TBM via an operation panel including a controller, a control unit controls the TBM based on the input operating signal and outputs the status of the TBM via a display unit.

[0006] However, conventional TBM control devices, including those in prior art literature, require the user to directly control the rotation of the cutter, the operation of the main jacks and articulated jacks. This makes accurate excavation difficult depending on the user's skill level. In particular, curved excavation is difficult because it must be performed along a three-dimensional curve that has curvature not only in the xy plane but also in the z-axis direction. Articulated jacks, which are placed between the shields, must be arranged in multiples around the shield and each articulated jack must be contracted or expanded in different lengths and directions. This limits the system, as errors are inevitable when users directly control it based on their own judgment.

[0007] In particular, conventional TBM control systems are concentrated in surveying systems that use gyroscopes, accelerometers, and magnetometers for directional control, and it is recognized that they are mathematically incapable of solving the multiple-solution problem resulting from the combined drive of articulated jacks and shielded jacks.

[0008] Furthermore, as shown in Figure 2, even with precise operational control by the user, there are limitations due to the ground characteristics and machine weight, which inevitably cause pitching, yawing, and rolling rotations during the TBM's advancement. Therefore, even with a skilled user controlling the TBM based on provided drawing information, real-time errors make accurate excavation difficult, and these problems become even more severe in the case of curved excavation. [Prior art documents] [Patent Documents]

[0009] [Patent Document 1] Korean Published Patent Gazette No. 1997-0044027 [Patent Document 2] Korean Published Patent Gazette No. 2022-0033323 [Overview of the Initiative] [Problems that the invention aims to solve]

[0010] The present invention was made to solve the problems of the conventional technology described above, and its purpose is to provide a steering control device for a shield TBM and a steering control method using the same that can realize accurate steering control along a pre-designed excavation path without relying on the user's human judgment, and in particular when controlling multiple articulated jacks in curved excavation, continuous control rather than discrete control is possible by accurately calculating the stroke length of each articulated jack using an artificial intelligence calculation model, and even if some errors occur due to pitching, yawing, and rolling during the excavation process, by performing calculations in real time each time the shield jack is operated, deviation from the excavation path due to cumulative errors is prevented, and it is possible to re-verify whether the excavation path predicted by the calculation matches the target. [Means for solving the problem]

[0011] To achieve the above object, a steering control method CM using the steering control device of the shield TBM of the present invention is such that the steering control device SC of the shield TBM individually controls the stroke lengths of at least four intermediate folding jacks 3 provided between the front shield 1 having a cutter head 1a and the intermediate shield 2 along the excavation path and radially spaced apart at equal intervals. The steering control device SC has the reference point coordinates T of the cutter head 1a at the current location c and the reference point coordinates T of the cutter head 1a at the target location by one cycle operation of the shield jack 4 t and extracting step S10, and the steering control device SC uses an artificial intelligence calculation model to calculate the reference point coordinates T at the current location c and the reference point coordinates T at the target location t Based on this, the stroke length l of each intermediate folding jack 3 is calculated as a continuous real number in step S20, and the steering control device SC simultaneously operates each intermediate folding jack 3 according to the calculated stroke length l in step S40. It is characterized by including

[0012] Further, the steering control device SC may further include a step S00 in which coordinates based on a pre-designed three-dimensional tunnel path are input.

[0013] Further, the artificial intelligence calculation model inputs the diameter D of the cutter head 1a, the reference point coordinates T at the current location and the target location c 、T t According to this, the intermediate folding angles α and β can be input to calculate the stroke length l of each intermediate folding jack 3 within the operating range.

[0014] Further, the artificial intelligence calculation model includes a step S21A of generating a random number for the stroke length l within the operating range of each intermediate folding jack 3, and the intermediate folding angles α according to the stroke length l of each intermediate folding jack 3 based on the generated random number cal 、β cal Calculating step S22A, and the calculated intermediate folding angles α cal 、β calThe reference point coordinates T of the current location and the target location. c , T t The corresponding folding angle α in , β in Step S23A compares with the calculated bending angle α cal , β cal The predictive model can be a repeating loop that includes step S24A, which, if the value is within the error range, extracts the generated random number as the effective stroke length l of each folding jack 3, and returns to step S21A to generate the random number if the value is outside the error range.

[0015] Furthermore, the step S21A for generating the random numbers can restrict the range of random number generation to several points, centering on the effective stroke length l of each folding jack 3 extracted in the previous cycle.

[0016] Furthermore, the artificial intelligence computation model can be a machine running model for regression calculations generated by learning a dataset that includes diameter D and bending angles α and β as input variables, and stroke length l as an output variable.

[0017] Furthermore, the steering control device SC may further include step S25 of outputting the calculated stroke length l of each folding jack 3 to a display.

[0018] Then, the steering control device SC, when each folding jack 3 operates according to the calculated stroke length l, predicts the reference point coordinates T of the cutter head 1a. p and the reference point coordinates T of the target location t The step S30 may further include verifying whether it is operational by comparing it with the other.

[0019] The steering control device SC of the shield TBM of the present invention includes a memory containing an application program for providing a steering control method for individually controlling the stroke length of at least four folding jacks 3 that are provided between a forward shield 1 equipped with a cutter head 1a along the excavation path and an intermediate shield 2 and are spaced radially at equal intervals, and a processor that executes the application program stored in the memory, wherein the processor executes the application program to determine the reference point coordinates T of the cutter head 1a at the current location. c And the reference point coordinate T of the cutter head 1a at the target location due to one cycle of operation of the shield jack 4. t The following are extracted, and an artificial intelligence computation model is used to determine the reference point coordinates T of the current location. c and the coordinates T of the reference point of the target location t Based on this, the stroke length l of each folding jack 3 is calculated, and the predicted reference point coordinates T of the cutter head 1a are determined when each folding jack 3 operates according to the calculated stroke length l. p and the reference point coordinates T of the target location t The steering control device SC verifies the feasibility of operation by comparing the two, and then simultaneously operates each folding jack 3 according to the calculated stroke length l.

[0020] Furthermore, the four folding jacks 3 are spaced apart around a rotating pin located in the center of the shield, and the stroke length l of the upper folding jack and the lower folding jack u ,l d The sum of the stroke lengths of the left and right folding jacks. l ,l r The sum can always be controlled to remain constant. [Effects of the Invention]

[0021] The steering control device SC according to the present invention calculates the stroke length of each folding jack based on the reference point coordinates of the current location and the reference point coordinates of the target location using an artificial intelligence calculation model, thereby enabling accurate steering control along a pre-designed excavation path without relying on the user's manual judgment.

[0022] In particular, even in situations requiring high precision, such as curved excavation, the artificial intelligence computation model allows for continuous and accurate control of multiple articulated jacks over the stroke length, rather than discrete control, enabling curved excavation to be performed within the margin of error.

[0023] Furthermore, even if some errors occur between the actual excavation path and the designed excavation path due to shield pitching, yawing, and rolling during the excavation process, calculations can be performed in real time each time the shield jack operates in a cycle, thereby preventing deviation from the excavation path due to cumulative errors.

[0024] Furthermore, the stroke length of each folding jack calculated by the artificial intelligence calculation model is immediately output to the display, allowing the user to select whether or not to operate the folding jack. Depending on the embodiment, if each folding jack is to be operated according to the calculated stroke length, the predicted reference point coordinates can be compared with the target reference point coordinates to verify whether or not to operate. [Brief explanation of the drawing]

[0025] [Figure 1a] This is a schematic diagram showing the structure of a typical shielded TBM. [Figure 1b] This is a conceptual diagram showing a different folding system for a standard shielded TBM. [Figure 2] This is a conceptual diagram illustrating the pitching, yawing, and rolling phenomena that occur in a shield. [Figure 3] This is a flowchart illustrating a steering control method according to one embodiment of the present invention in chronological order. [Figure 4]This is a schematic diagram illustrating a shielded TBM structure for explaining a steering control method according to one embodiment of the present invention. [Figure 5] This is a simulation image showing the excavation path formed when excavation is carried out continuously according to the angle of inclination. [Figure 6] Figure 4 is a schematic diagram illustrating the steering of the cutter head by the operation of the folding jack in the shield TBM structure. [Figure 7] Figure 4 is a schematic diagram illustrating the excavation process using a shield TBM structure and the operation of shield jacks. [Figure 8a] This is a simulation image showing the excavation path using the steering control method based on the artificial intelligence model of the present invention. [Figure 8b] This is a simulation image showing the excavation path using the steering control method based on the artificial intelligence model of the present invention. [Figure 8c] This is a simulation image showing the excavation path using the steering control method based on the artificial intelligence model of the present invention. [Figure 8d] This is a simulation image showing the excavation path using the steering control method based on the artificial intelligence model of the present invention. [Modes for carrying out the invention]

[0026] The following describes preferred embodiments of the present invention in detail based on the drawings. However, if a specific description of a related known function or configuration is deemed likely to unnecessarily obscure the gist of the present invention, such detailed description will be omitted.

[0027] The shield TBM to be controlled by the present invention uses an X-shaped folding system and includes a forward shield 1 equipped with a cutter head 1a and at least four folding jacks 3 spaced radially at equal intervals between it and an intermediate shield 2. It also includes a plurality of shield jacks 4 that are supported by a segment ring 5 directly or indirectly connected to the forward shield 1 and installed continuously on the excavation surface, thereby applying thrust. At this time, a further rear shield may be provided behind the intermediate shield 2.

[0028] The steering control device SC for the shield TBM of the present invention includes a memory storing an application program for providing a steering control method for the shield TBM, and a processor that executes the application program stored in the memory, wherein an artificial intelligence calculation model determines the reference point coordinates T of the current location. c and the coordinates T of the reference point of the target location t Based on this, the stroke length l of each folding jack 3 is calculated, and in accordance with the calculated stroke length l, each folding jack 3, which is spaced radially apart, is operated simultaneously and continuously, enabling accurate excavation within the margin of error even when the pre-designed excavation path includes curves.

[0029] Next, we will explain in detail the steering control method CM performed by the steering control device SC. Figure 3 is a flowchart showing the steering control method CM according to one embodiment of the present invention in chronological order.

[0030] First, the steering control device SC receives coordinates based on a pre-designed three-dimensional tunnel path (S00). It is desirable that the coordinates based on the tunnel path be specified so as to match the reference point coordinates T of the cutter head 1a of the shield TBM, which will be described later. The tunnel path coordinates are in the form of fixed data and can be stored in memory and retrieved, or input from an external storage device.

[0031] The steering control device SC uses the coordinates of the three-dimensional tunnel to determine the reference point coordinates T of the cutter head 1a at the current location. c And the reference point coordinate T of the cutter head 1a at the target location due to one cycle of operation of the shield jack 4. t Extract and (S10).

[0032] The aforementioned reference point coordinates T preferably include the coordinates of the front reference point of the cutter head 1a. Depending on the embodiment, as shown in Figure 4, the rear-left point and rear-right point of the front shield 1 may be further included to reflect the attitude of the front shield 1 due to pitching and yawing. Also, the reference point coordinates T of the cutter head 1a at the current location. c This is the reference point coordinate T predicted by calculation in the previous cycle. p While this can be used as is, the real-time coordinates of the cutter head 1a relative to the reference point may be measured using a gyro sensor and an acceleration sensor so as to reflect errors caused by pitching, yawing, and rolling that occur during the excavation process.

[0033] At this time, the reference point coordinates T of the current location and the target location c , T t It is desirable that the corresponding bending angles α and β are extracted together. The bending angle consists of a horizontal angle α and a vertical angle β, where the horizontal angle α represents the angle in the xy-plane and the vertical angle β represents the angle in the yz-plane. That is, the reference point coordinates T of the current point and the target point. c , T t The bending angles α and β can be extracted together by connecting them with a vector.

[0034] Figure 5 is a simulation image illustrating, for ease of understanding, the excavation path formed when excavation is carried out continuously according to the bending angles α and β, where (a) is when the bending angles α and β are 0°, (b) is when the vertical angle β is 0°, (c) is when the horizontal angle α is 0°, and (d) is when neither the horizontal angle α nor the vertical angle β is 0°.

[0035] Next, the steering control device SC uses an artificial intelligence calculation model to determine the reference point coordinates T of the current location. c and the coordinates T of the reference point of the target location t Based on this, the stroke length l of each folding jack 3 is calculated (S20). The artificial intelligence calculation model uses the diameter D of the cutter head 1a, and the reference point coordinates T of the current location and the target location. c , T t The system takes the corresponding folding angles α and β as input and calculates and outputs the stroke length l of each folding jack 3 as a continuous real value within the operating range.

[0036] On the other hand, the artificial intelligence calculation model can be implemented in various ways to calculate the stroke length l of the folding jack 3, but the present invention proposes two forms of calculation models, and each artificial intelligence calculation model will be described in more detail below.

[0037] In the two embodiments described below, the description is based on a configuration in which four folding jacks 3 are arranged radially at a distance from each other, and the symbol l is used for the stroke length. u ,l d ,l l ,l r l represents the stroke length of the upper folding jack, lower folding jack, left folding jack, and right folding jack, respectively. That is, as shown in Figure 6, the stroke length l of each folding jack for curved excavation. u ,l d ,l l ,l r This must be determined, and as shown in Figure 7, the reference point coordinates T of the current location when the shield jack 4 is activated must be determined. c The reference point coordinates T of the target locationt The stroke length l approximates this. u ,l d ,l l ,l r This must be controlled as a continuous real value.

[0038] <Predictive model using repeating loops> One embodiment of the artificial intelligence computation model is a predictive model using a repeating loop, and in order to find the optimal stroke length l for the input variables diameter D, bending angle α, and β, it can be formulated as the following optimization problem.

[0039] MINIMIZE[α cal -α in ] and MINIMIZE[β cal -β in ] subject to 0≦l u ,l d ,l l ,l r ≤200(mm) (Here, α cal and β cal This is the angle of inclination based on the calculation result, and α in and β in This is the target angle of inclination.

[0040] More specifically, the artificial intelligence calculation model calculates the stroke length l within the operating range of each folding jack 3. u ,l d ,l l ,l r Random numbers are generated for each (S21A). At this time, the generated random numbers are limited to positive numbers within the range of 200 mm, which is the operable length of each folding jack 3, and are generated as continuous real values.

[0041] Subsequently, the artificial intelligence calculation model calculates each generated random number over the stroke length l of each folding jack 3. u ,l d ,l l ,l rand calculate the corresponding folding angles α cal and β cal (S22A), and compare the calculated folding angles α cal and β cal with the reference point coordinates T c and T t of the current point and the target point, and compare with the corresponding folding angles α in and β in (S23A).

[0042] At this time, if the calculated folding angles α cal and β cal and the target folding angles α in and β in are within the error range set by the user in advance, extract the generated random numbers as the effective stroke lengths l u of each folding jack 3, l d l l l r ; if they are outside the error range, return to the step S21A of generating random numbers and repeat the above process (S24A).

[0043] On the other hand, in the prediction model using the above-mentioned iterative loop, since random numbers are generated to calculate the folding angles α cal and β cal and compare with the target folding angles α in and β in , there is a risk of delay in the calculation time. To solve this problem, when compiling the prediction model using the above iterative loop, at the time of actually executing the program, perform JIT (Just-In-Time) compilation that translates it into machine language, and it is desirable to implement parallel operations in the random search based on each random number generated by implementing multi-processing.

[0044] Also, when generating random numbers, when applying a probabilistic optimization method, which is a kind of Monte Carlo method, to extract continuous real values, quickly obtain the effective stroke lengths l u of l d l l l rBy extracting certain values ​​and restricting the random number generation range to several points, centered around the effective stroke length extracted in the previous cycle, high accuracy can be achieved while reducing computation time.

[0045] In particular, utilizing predictive models with repeating loops eliminates the need for pre-processing, such as collecting a large amount of data in advance for specific conditions and repeatedly training on that collected data. This offers the advantage of enabling the rapid implementation of computational modules compared to other deep learning and machine learning models.

[0046] <Random Forest Model> Another embodiment of the artificial intelligence computation model is a Random Forest model for regression calculations, which predicts the stroke length l as a continuous real value for input variables: diameter D, bending angle α, and β.

[0047] More specifically, the input variables include diameter D and bending angles α and β, and the corresponding output variable is the stroke length l for each of the four bending jacks 3. u ,l d ,l l ,l r After collecting a large dataset containing the data, the dataset is split into a training dataset and a test dataset, and a decision tree is constructed by training the test set.

[0048] Furthermore, when evaluating a model using a test set, the mean squared error (MSE) and R are used. 2 The scores can be used to train the model to improve prediction accuracy and model fit. Furthermore, K-fold cross-validation can be used to further improve the generalization performance of the model for model optimization, or grid search or random search can be used to find the optimal hyperparameters for the random forest model.

[0049] On the other hand, the machine-running models for the regression calculation can, of course, be implemented using polynomial regression models, support vector models, gradient boosting models, artificial neural network (ANN) models, Gaussian process regression (GPR) models, and so on.

[0050] As described above, once the steering control device SC calculates the stroke length l of each folding jack 3 using an artificial intelligence calculation model, the steering control device SC can output the calculated stroke length l of each folding jack 3 via a display so that the user can confirm it (S25). At this time, along with the stroke length l, the error rate and prediction accuracy according to the artificial intelligence calculation model are also output, allowing the user to choose whether or not to operate the folding jack 3 with its stroke length l.

[0051] Furthermore, the steering control device SC predicts the reference point coordinates T of the cutter head 1a when each folding jack 3 operates according to the calculated stroke length l. p and the coordinates T of the reference point at the target location t The feasibility of operation can be verified by comparing it with the other (S30).

[0052] This makes it difficult to determine whether the folding jack 3 can operate based simply on numerical values ​​such as stroke length l, error rate, and prediction accuracy, as the predicted reference point coordinates T during operation are used. p and the coordinates T of the reference point of the target location t It can be intuitively compared using as a reference. At this time, the reference point coordinates T are automatically calculated by the steering control device SC. p , T t The comparison results can also be output via the display.

[0053] Finally, when the user issues an operation command, the steering control device SC simultaneously operates each of the folding jacks 3 according to the calculated stroke length l (S40), and thereafter, the shield jack 4 operates in conjunction with the rotation of the cutter head 1a, completing one cycle of excavation. At this time, since the operating speed of each folding jack 3 is the same, the stroke length l of each folding jack 3 is proportional to the operating time.

[0054] Figures 8a to 8d are simulation images illustrating the excavation path obtained by performing the steering control method of the present invention using an artificial intelligence model that utilizes a prediction model (iteration) with an error range of 0.5% and a random forest model (machine learning). In this case, the diameter D of the reflected cutter head 1a is 8,000 mm, and the length of the forward shield is l F The simulation was performed by operating the system for 100 cycles, with a length of 3,000 mm, a segment ring 5 length and a shield jack 4 stroke length of 1,500 mm, and a shield jack speed of 10 mm / sec.

[0055] Specifically, Figure 8a shows the results of an excavation simulation for a shield TBM tunnel with an upward 0.2° inclination (α=0°, β=0.2°), Figure 8b shows the results of an excavation simulation for a shield TBM tunnel with a leftward 0.2° inclination (α=0.2°, β=0°), Figure 8c shows the results of an excavation simulation for a shield TBM tunnel with a rightward upward inclination (α=0.5°, β=0.5°), and Figure 8d shows the results of an excavation simulation for a shield TBM tunnel with a leftward upward inclination (α=1.0°, β=0.5°). It was found that these results closely matched the path of the designed 3D tunnel.

[0056] On the other hand, the steering control device SC for the shielded TBM of the present invention is a computing device that performs the steering control method CM for the shielded TBM described above, and details that overlap with the matters described for the steering control method CM will be omitted.

[0057] The present invention provides a steering control method CM for individually controlling the stroke length of at least four folding jacks 3 that are provided between a forward shield 1 equipped with a cutter head 1a along a pre-designed excavation path and an intermediate shield 2, and are spaced radially at equal intervals, and includes a memory storing an application program and a processor that executes the application program stored in the memory.

[0058] The processor of the steering control device SC receives coordinates based on a pre-designed 3D tunnel path by executing an application program, and the reference point coordinates T of the cutter head 1a at the current location. c And the reference point coordinate T of the cutter head 1a at the target location due to one cycle of operation of the shield jack 4. t Extract and.

[0059] The aforementioned processor uses an artificial intelligence computation model linked to the application program to determine the reference point coordinates T of the current location. c and the coordinates T of the reference point of the target location t Based on this, the stroke length l of each folding jack 3 is calculated as a continuous real value, and each folding jack 3 is operated simultaneously according to the calculated stroke length l.

[0060] At this time, the processor predicts the reference point coordinates T of the cutter head 1a when each folding jack 3 operates according to the calculated stroke length l. p And the reference point coordinates T of the target location t By comparing them, it is possible to verify whether or not it can operate.

[0061] Furthermore, if the shield TBM is an X-shaped folding system, the four folding jacks 3 are arranged spaced apart from a rotating pin located in the center of the shield.

[0062] At this time, the stroke length l of the upper folding jack and the lower folding jack u ,l dThe sum of the stroke lengths of the left and right folding jacks. l ,l r The sum of these values ​​can be controlled to remain constant so that it always achieves the maximum stroke length. For example, if the operating range of each folding jack 3 is 200 mm, the aforementioned sum can be kept constant at 200 mm. In this case, by reducing the amount of computation performed by the artificial intelligence calculation model, simpler and more accurate calculations can be achieved.

[0063] The steering control device SC for a shielded TBM and the steering control method CM using the same, as described above, can be implemented in other specific forms by a person with ordinary skill in the art to which the present invention belongs, without changing the technical idea or essence of the present invention.

[0064] Therefore, these embodiments described above should be understood to be illustrative in all respects and not limiting, and the scope of the present invention is defined more by the claims described below than by the detailed description above, and all modified or altered forms derived from the meaning and scope of the claims, as well as equivalent concepts, should be interpreted as being included within the scope of the present invention. [Explanation of Symbols]

[0065] 1: Front Shield 2: Mid-shield 3: Folding jack 4: Shield Jack 5: Segment Ring SC: Steering control unit

Claims

1. A steering control method (CM) for a shield TBM, wherein the steering control device (SC) of the shield TBM individually controls the stroke length of at least four folding jacks (3) that are provided between a forward shield (1) equipped with a cutter head (1a) along the excavation path and an intermediate shield (2), and are spaced radially at equal intervals, The steering control device (SC) determines the reference point coordinates (T) of the cutter head (1a) at the current location. ｃ ) and the reference point coordinates (T) of the cutter head (1a) at the target location due to one cycle of operation of the shield jack (4) ｔ The steps include extracting (S10) and The steering control device (SC) uses an artificial intelligence calculation model to determine the reference point coordinates (T) of the current location. ｃ ) and the reference point coordinates (T ｔ Step (S20) is to calculate the stroke length (l) of each folding jack (3) as a continuous real value based on ), The steering control device (SC) simultaneously operates each folding jack (3) according to the calculated stroke length (l) (S40), A steering control method using a steering control device for a shielded TBM, characterized by including the following:

2. The steering control method using the steering control device of a shield TBM according to claim 1, further comprising the step (S00) of inputting coordinates based on a pre-designed three-dimensional tunnel path to the steering control device (SC).

3. The R artificial intelligence computation model is Diameter D of the cutter head (1a), reference point coordinates (T) of the current location and target location. ｃ , T ｔ A steering control method using a steering control device for a shield TBM according to claim 1, characterized in that the folding angle (α, β) corresponding to the ) is input and the stroke length (l) of each folding jack (3) is calculated within the operating range.

4. The R artificial intelligence computation model is Step (S21A) of generating a random number for the stroke length (l) within the operating range of each folding jack (3), Based on the generated random numbers, the folding angle (α) of each folding jack (3) is determined according to the stroke length (l). ｃａｌ , β ｃａｌ The steps include: (S22A) calculating ) and Calculated intermediate folding angle (α ｃａｌ , β ｃａｌ ), and compare it with the intermediate folding angle (α ｃ , β ｔ ) corresponding to the reference point coordinates (T ｉｎ , β ｉｎ ) of the current point and the target point. Step (S23A), and Calculated bending angle (α ｃａｌ , β ｃａｌ If the value is within the error range, the generated random number is extracted as the effective stroke length (l) of each folding jack (3). If it is outside the error range, the process returns to the step of generating random numbers (S21A) (S24A). A steering control method using a steering control device for a shielded TBM according to claim 3, characterized in that it is a predictive model using a repeating loop that includes a predictive model.

5. The step of generating the aforementioned random numbers (S21A) is: A steering control method using a steering control device for a shielded TBM according to claim 4, characterized in that the range of random number generation is limited to several points, centered around the effective stroke length (l) of each folding jack (3) extracted in a previous cycle.

6. The steering control method using a steering control device for a shielded TBM according to claim 3, characterized in that the artificial intelligence computation model is a machine running model for regression calculations generated by learning a dataset that includes diameter (D) and bending angle (α, β) as input variables and stroke length (l) as an output variable.

7. The steering control method using the steering control device of a shielded TBM according to claim 3, further comprising the step (S25) of outputting the calculated stroke length (l) of each folding jack (3) to a display, wherein the steering control device (SC) further comprises the step of outputting the calculated stroke length (l) of each folding jack (3) to a display.

8. When the steering control device (SC) operates each folding jack (3) according to the calculated stroke length (l), the predicted reference point coordinates (T) of the cutter head (1a) ｐ ) and the reference point coordinates (T ｔ A steering control method using a steering control device for a shielded TBM according to claim 3, further comprising the step (S30) of verifying whether it can be operated by comparing it with the other.

9. A steering control device (SC) for a shield TBM includes a memory containing an application program for providing a steering control method for individually controlling the stroke length of at least four folding jacks (3) that are provided between a forward shield (1) equipped with a cutter head (1a) along the excavation path and an intermediate shield (2) and are spaced radially at equal intervals, and a processor for executing the application program stored in the memory, The processor executes an application program to determine the reference point coordinates (T) of the cutter head (1a) at the current location. ｃ ) and the reference point coordinates (T) of the cutter head (1a) at the target location due to one cycle of operation of the shield jack (4) ｔ ) and extract, Using an artificial intelligence computation model, determine the reference point coordinates (T) of the current location. ｃ ) and the reference point coordinates (T ｔ Based on this, the stroke length (l) of each folding jack (3) is calculated as a continuous real value. When each folding jack (3) operates according to the calculated stroke length (l), the predicted reference point coordinates (T) of the cutter head (1a) ｐ ) and the reference point coordinates (T ｔ ) Compare it with the other to verify whether it works, A steering control device for TBM, characterized by simultaneously operating each folding jack (3) according to the calculated stroke length (l).

10. The aforementioned folding jacks (3) consist of four jacks spaced apart from a rotating pin located in the center of the shield, and the stroke length (l) of the upper folding jack and the lower folding jack is ｕ , l ｄ The sum of ( ) and the stroke length ( l of the left and right folding jacks) ｌ , l ｒ The steering control device for a shielded TBM according to claim 9, characterized in that the sum of ) is always controlled to be constant.

Images