TBM large-angle vertical curve direction adjusting system

Abstract

The invention discloses a TBM wide-angle vertical curve steering system, and relates to the technical field of TBM steering, and the technical scheme is characterized in that the TBM wide-angle vertical curve steering system comprises an inclination monitoring module used for monitoring initial inclination data of an anterior shield of a TBM; the data analysis control module is used for comparing and analyzing the initial inclination data with preset curve design parameters and formulating an adjustment instruction by applying a dynamic posture adjustment algorithm according to an analysis result; the auxiliary head-up execution module is used for driving a mechanical device in the auxiliary head-up system according to the adjustment instruction so as to adjust the inclination angle of the TBM anterior shield based on a PID control algorithm; and the feedback adjustment module is used for monitoring the real-time inclination data of the TBM anterior shield in real time, and replacing the initial inclination data with the real-time inclination data to realize dynamic update of the adjustment instruction. According to the method, the adjustment precision and the response speed can be remarkably improved, and a more accurate and rapid inclination adjustment decision is realized.

Description

Technical Field

[0001] This invention relates to the field of TBM steering technology, and more specifically, to a TBM large-angle vertical curve steering system. Background Art

[0002] During tunneling, a TBM (Tunnel Boring Machine) may deviate from its target axis due to factors such as its own weight, geological conditions, and human factors. In such cases, correction and reorientation are necessary to ensure construction quality. The purpose of correction and reorientation is to bring the TBM back to the target trajectory, avoid damage to the cutterhead and cutter head caused by over-adjustment, and reduce the risk of the TBM jamming.

[0003] In traditional TBM applications, especially when adjusting the orientation of TBMs on large-angle vertical curves, and particularly when the TBM needs to "raise its head," the TBM's own weight makes it difficult to effectively "raise its head," which affects construction efficiency, leads to insufficient construction accuracy, and causes safety risks and project delays.

[0004] Therefore, how to research and design a TBM large-angle vertical curve steering system that can overcome the above-mentioned defects is an urgent problem that we need to solve. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to provide a TBM large-angle vertical curve orientation system. This system uses a dynamic attitude adjustment algorithm based on a fusion of Fast Fourier Transform, Autoregressive Moving Average Model, Decision Tree, and Convolutional Neural Network to compare and analyze preset design parameters. Based on the analysis results, it determines whether the tilt angle of the TBM needs to be adjusted, which can significantly improve the accuracy and response speed of the adjustment, and achieve more accurate and rapid tilt adjustment decisions.

[0006] The above-mentioned technical objective of the present invention is achieved through the following technical solution: a TBM large-angle vertical curve steering system, comprising:

[0007] The tilt monitoring module is used to monitor the initial tilt data of the front shield of the TBM;

[0008] The data analysis and control module is used to compare and analyze the initial tilt data with the preset curve design parameters, and formulate adjustment instructions based on the analysis results using a dynamic attitude adjustment algorithm.

[0009] The auxiliary head-up execution module is used to drive the mechanical device in the auxiliary head-up system according to the adjustment command, so as to adjust the tilt angle of the TBM front shield based on the PID control algorithm.

[0010] The feedback adjustment module is used to monitor the real-time tilt data of the TBM front shield and dynamically update the adjustment command by replacing the initial tilt data with the real-time tilt data.

[0011] Furthermore, the tilt monitoring module includes:

[0012] The first sensor unit is used to measure the front shield data of the TBM, which includes initial tilt data and implementation effect data.

[0013] The data transmission unit is used to transmit the initial tilt data to the data analysis and control module via wireless communication technology.

[0014] Furthermore, the data analysis and control module includes:

[0015] The data analysis unit is used to receive initial tilt data sent from the tilt monitoring module and / or real-time tilt data sent from the feedback adjustment module, and to perform preliminary processing on the tilt data using statistical filtering and digital signal processing techniques.

[0016] The command decision unit is used to receive the pre-processed tilt data, compare and analyze it with the preset design parameters using a dynamic attitude adjustment algorithm, and decide whether the tilt angle of the TBM needs to be adjusted based on the analysis results, thus obtaining the adjustment command.

[0017] The feedback receiving unit is used to send the adjustment instructions generated by the instruction decision unit to the auxiliary head-up execution module, and at the same time receive the implementation effect data from the feedback adjustment module.

[0018] The implementation effect data includes posture data, hydraulic cylinder extension and retraction data, and environmental and equipment status data.

[0019] Furthermore, the dynamic attitude adjustment algorithm is based on a fusion of Fast Fourier Transform, Autoregressive Moving Average model, decision tree, and convolutional neural network. The process of comparing and analyzing preset design parameters in the dynamic attitude adjustment algorithm specifically involves:

[0020] We use Fast Fourier Transform to extract periodic variation features from tilted data and use an autoregressive moving average model to predict tilt trends and short-term tilt behavior.

[0021] An ensemble learning model is applied, combined with decision trees and convolutional neural networks, to compare and analyze the extracted features with preset design parameters, and to obtain the analysis results.

[0022] Based on the analysis results, a predictive model is used to dynamically adjust the control parameters to optimize the tilt angle adjustment of the TBM and generate adjustment commands.

[0023] Furthermore, the Fast Fourier Transform is used to convert the tilted data in the time domain into data in the frequency domain, and to obtain the period;

[0024] Furthermore, the autoregressive moving average model transforms the periodicity obtained through fast Fourier transform into time-dependent periodic features, and integrates the extracted periodic features as exogenous variables into the autoregressive moving average model.

[0025] Furthermore, the convolutional hierarchy in the convolutional neural network is used to extract and learn high-level features of the data, and then decision-making and classification of high-level features are performed through decision trees;

[0026] Furthermore, after the convolutional neural network undergoes convolution and pooling operations, the output is compressed and extracted into high-level features to form a feature vector.

[0027] Furthermore, the process of generating the adjustment instruction is as follows:

[0028] Develop a predictive model to describe how the tilt angle of the TBM responds to the control input;

[0029] Model predictive control strategy dynamically generates control commands by solving a rolling time-domain optimization problem, minimizing the cumulative cost within the future prediction horizon, which includes state error and control efficiency;

[0030] The sequential quadratic programming numerical optimization method is used to solve the optimization problem in real time. The first control input is applied to the TBM. In the next control cycle, the state is updated according to the latest tilt measurement, and the optimization problem is solved again.

[0031] Transform control inputs into adjustment commands.

[0032] Furthermore, the head-up assist module includes a mechanical device module and a control system module;

[0033] The mechanical device module includes a hydraulic cylinder and a connecting mechanism;

[0034] Additionally, the control system module uses the head-up assist system and data analysis control module to adjust the posture of the TBM front shield.

[0035] Furthermore, the process by which the control system module adjusts the TBM's front shield posture is specifically as follows:

[0036] The control system module receives adjustment commands from the data analysis and control module and extracts the front shield tilt angle from the adjustment commands;

[0037] Using a PID control algorithm, the difference between the target angle and the current angle is calculated based on the received adjustment command, an error signal is generated, and the control signal is calculated to adjust the driving voltage and current of the hydraulic cylinder to change the extension and retraction length of the cylinder.

[0038] The hydraulic cylinder is connected to the front shield via a connecting rod and a joint to adjust the tilt angle of the front shield.

[0039] Furthermore, the feedback adjustment module includes:

[0040] The second sensor unit is used to monitor the real-time tilt data and attitude of the TBM using a tilt sensor and a gyroscope.

[0041] The data feedback unit is used to transmit real-time tilt data to the data analysis and control module via wireless communication technology.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. The present invention provides a TBM large-angle vertical curve orientation system, which uses a dynamic attitude adjustment algorithm based on the fusion of fast Fourier transform, autoregressive moving average model, decision tree and convolutional neural network to compare and analyze with preset design parameters, and decides whether the tilt angle of the TBM needs to be adjusted based on the analysis results. This can significantly improve the accuracy and response speed of the adjustment, and achieve more accurate and rapid tilt adjustment decisions.

[0044] 2. This invention utilizes an auxiliary head-up system to adjust the posture and tilt of the TBM front shield, which can effectively adapt to complex and variable geological conditions, while reducing mechanical stress and potential operational risks. The use of the auxiliary head-up system makes the TBM more flexible and safer when handling large-angle vertical adjustments. Attached Figure Description

[0045] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0046] Figure 1 This is a system block diagram in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with examples and drawings. The exemplary embodiments of the present invention and their descriptions are only used to explain the present invention and are not intended to limit the present invention.

[0048] It should be noted that when a component is referred to as being "fixed to" or "set on" another component, it can be directly or indirectly attached to that other component. When a component is referred to as being "connected to" another component, it can be directly or indirectly connected to that other component.

[0049] It should be understood that the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention.

[0050] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0051] Example: A TBM large-angle vertical curve steering system, such as Figure 1 As shown, the module includes a tilt monitoring module, a data analysis and control module, an auxiliary head-up execution module, and a feedback adjustment module.

[0052] The tilt monitoring module is used to monitor the tilt angle of the front shield of the TBM in real time. It collects initial tilt data in real time through sensors and sends the initial tilt data to the data analysis and control module. The tilt monitoring module includes a first sensor unit and a data transmission unit; the first sensor unit is used to measure the front shield data of the TBM, including initial tilt data and implementation effect data; the data transmission unit is used to transmit the initial tilt data to the data analysis and control module through wireless communication technology.

[0053] The data analysis and control module is responsible for receiving tilt data from the tilt monitoring module and the feedback adjustment module, and comparing and analyzing this data with the preset curve design parameters; based on the analysis results, the data analysis and control module uses a dynamic attitude adjustment algorithm to formulate adjustment commands and send them to the auxiliary head-up execution module.

[0054] Specifically, the data analysis and control module includes a data analysis unit, a command decision unit, and a feedback unit. The data analysis unit receives tilt data from the tilt monitoring module and performs preliminary processing using statistical filtering and digital signal processing techniques. The command decision unit receives the pre-processed tilt data, compares and analyzes it with preset design parameters using a dynamic attitude adjustment algorithm, and determines whether the TBM's tilt angle needs adjustment based on the analysis results. The feedback unit sends the adjustment commands generated by the command decision unit to the auxiliary head-up execution module and simultaneously receives implementation effect data from the feedback adjustment module; this implementation effect data includes, but is not limited to, attitude data, cylinder extension / retraction data, and environmental and equipment status data.

[0055] The statistical filtering techniques in this invention are primarily used to extract useful information from noisy signals. These techniques can effectively remove or reduce random variations in data while preserving important signal characteristics, including Kalman filtering and moving averages.

[0056] Digital signal processing (DSC) is the technology for analyzing, modifying, and manipulating digital signals. It includes various methods for processing signals to improve signal quality or extract important information. In tilt monitoring systems, common DSC techniques include Fast Fourier Transform (FFT) and digital filtering.

[0057] The dynamic attitude adjustment algorithm is based on a fusion of Fast Fourier Transform, Autoregressive Moving Average model, decision tree, and convolutional neural network. The steps involved in comparing and analyzing the preset design parameters in the dynamic attitude adjustment algorithm are as follows:

[0058] S4.1. Use Fast Fourier Transform to extract periodic variation features from tilted data, and use an autoregressive moving average model to predict tilt trends and short-term tilt behavior.

[0059] S4.2 Apply an ensemble learning model, combining decision trees and convolutional neural networks, to compare and analyze the extracted features with the preset design parameters.

[0060] S4.3 Based on the preceding analysis results, the predictive model is used to dynamically adjust the control parameters to optimize the tilt angle adjustment of the TBM and generate specific adjustment instructions.

[0061] In step S4.1, the Fast Fourier Transform (FFT) is used to convert the time-domain skewed data into frequency-domain data and obtain the period T. The mathematical model involved in the FFT is as follows:

[0062]

[0063] Where X[k] represents the complex result at the k-th point in the frequency domain, which is the output after the time series is converted to the frequency domain, representing the amplitude and phase information of frequency k; x[n] represents the value at the n-th point in the time series, which is the input tilt angle data; k represents the frequency index; f represents the physical frequency; f s Indicates the sampling frequency; e -j2πkn / N This represents a complex exponential function used to transform a time series to the frequency domain in an FFT; T represents the period, which is the reciprocal of the frequency.

[0064] The autoregressive moving average model transforms the period T obtained from the fast Fourier transform into a periodic feature Z related to time t. t The extracted periodic features are integrated as exogenous variables into the autoregressive moving average model. The mathematical expression after integrating the periodic features into the autoregressive moving average model is as follows:

[0065]

[0066] Among them, X t Z represents the sequence value at time point t; c represents the constant term; Z t Represents periodicity; α represents the coefficient of the periodicity; ∈ t The white noise error term represents time t; The autoregressive part represents the influence of the p past values ​​of the sequence on the current value; The moving average portion represents the impact of the past q white noise error terms on the current value.

[0067] In step S4.2, the convolutional hierarchy in the convolutional neural network is used to extract and learn high-level features of the data, and then these high-level features are used to make decisions and classify them through a decision tree.

[0068] The mathematical expressions for the operations of the convolutional hierarchy in a convolutional neural network are as follows:

[0069] a (l) =σ(b (l) +k (l) *x (l) );

[0070] Among them, a (l) σ is the activation output of layer l; b is the activation function; (l) It is a bias term; k (l) It represents the convolution kernel; * indicates the convolution operation; x (l) It is the input of the l-th layer.

[0071] After convolutional neural networks undergo convolution and pooling operations, a (l) The output is compressed and extracted into high-level features, forming a feature vector v.

[0072] The decision tree takes input from the feature vector v of the convolutional neural network and performs decision-making and classification. The mathematical expression involved is as follows:

[0073]

[0074] Among them, D p It is the dataset of the parent node; IG(D) p f) represents the parent node D p The information gain obtained after segmentation using feature f; I(D) is the impurity measure function; D p D is the number of samples in the parent node's dataset; left and D right These are the left and right subsets of the dataset after splitting; |D left |and|D right | represents the number of samples in the left and right subsets.

[0075] In step S4.3, the steps involved in adjusting the control parameters of the predictive model and generating specific adjustment instructions are as follows:

[0076] S4.3.1 Establish an accurate prediction model f(x) t ,u t This describes how the tilt angle of the TBM responds to the control input;

[0077] S4.3.2, Model Predictive Control (MMCC) dynamically generates control commands by solving a rolling time-domain optimization problem, minimizing the cumulative cost within the future prediction horizon. The cumulative cost includes state error and control efficiency. The optimization problem can be expressed as:

[0078]

[0079] Where is the cost function at time k, including the state error x. k -x ref and control input u k The weights;

[0080] F(x t +N) is the predicted terminal cost at the end of the horizon; N is the predicted horizon length;

[0081] S4.3.3. Solve the above optimization problem in real time using the sequential quadratic programming numerical optimization method, taking the first control input u. t On the applied TBM, in the next control cycle, the state x is updated based on the latest tilt measurement. t And then resolve the optimization problem;

[0082] S4.3.4 Finally, control input u tConverted into an adjustment command.

[0083] In this embodiment, the auxiliary head-up execution module utilizes the auxiliary head-up system and drives the mechanical device according to the adjustment instructions of the data analysis and control module to adjust the tilt angle of the TBM front shield.

[0084] The auxiliary head-up system is based on a PID control algorithm and is used to calculate and implement mechanical adjustments to control the tilt angle of the front shield.

[0085] Specifically, the head-up assist module includes a mechanical device module and a control system module; the mechanical device module includes a hydraulic cylinder and a connecting mechanism; the control system module uses the head-up assist system to adjust the posture of the TBM front shield based on data analysis.

[0086] The procedure for adjusting the TBM front shield posture by the control system module is as follows:

[0087] S8.1 The control system module receives the adjustment command from the data analysis control module and extracts the front shield tilt angle θ(t) from the adjustment command;

[0088] S8.2. Using the PID control algorithm, calculate the difference between the target angle and the current angle based on the received adjustment command, generate the error signal e(t), and calculate the control signal u(t) to adjust the driving voltage and current of the hydraulic cylinder to change the extension length of the cylinder.

[0089] S8.3 The hydraulic cylinder is connected to the front shield through a series of connecting rods and joints to adjust the tilt angle of the front shield;

[0090] In step S8.2, the mathematical expression for calculating the control signal u(t) using the PID control algorithm is as follows:

[0091] e(t) = θ target -θ(t);

[0092]

[0093] Where u(t) is the control signal; K p , K i , K d These are the proportional, integral, and derivative gains of the PID control algorithm. This represents the integral of the error from the start to the current time t; It is the differential of the error.

[0094] In this embodiment, the feedback adjustment module includes a second sensor unit and a data feedback unit. The second sensor unit is used to monitor the real-time tilt data and attitude of the TBM using a tilt sensor and a gyroscope; the data feedback unit is used to transmit the real-time tilt data to the data analysis and control module via wireless communication technology.

[0095] It should be noted that the sensors in the first sensor unit are installed at the front end of the TBM (only the sensors at the front end are included) to obtain tilt data at the start of operation; while the sensors in the second sensor unit include not only the specific sensors installed at the front end of the TBM, but also those installed at different locations from the tilt monitoring module, so as to capture the adjusted tilt data from other angles, and to comprehensively evaluate the effect of the adjustment by integrating multi-angle tilt data.

[0096] Working principle: This invention uses a dynamic attitude adjustment algorithm based on the fusion of fast Fourier transform, autoregressive moving average model, decision tree and convolutional neural network to compare and analyze the preset design parameters, and based on the analysis results, decides whether the tilt angle of the TBM needs to be adjusted. This can significantly improve the accuracy and response speed of the adjustment, and achieve more accurate and rapid tilt adjustment decisions.

[0097] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A TBM large-angle vertical curve adjustment system, characterized in that: include: Tilt monitoring module, used to monitor the initial tilt data of the front shield of the TBM; A data analysis control module is used to compare and analyze the initial tilt data with the preset curve design parameters, and formulate adjustment instructions using a dynamic posture adjustment algorithm based on the analysis results; An auxiliary head-up execution module is used to drive the mechanical device in the auxiliary head-up system according to the adjustment instruction to adjust the tilt angle of the TBM front shield based on the PID control algorithm; The feedback adjustment module is used to monitor the real-time tilt data of the TBM front shield in real time, and replace the initial tilt data with the real-time tilt data to realize dynamic update of the adjustment instructions.

2. A TBM large-angle vertical curve steering system according to claim 1, characterized in that: The tilt monitoring module comprises: The first sensor unit is used to measure the front shield data of the TBM, the front shield data including initial tilt data and implementation effect data; The data transmission unit is used to transmit the initial tilt data to the data analysis control module through wireless communication technology.

3. A TBM large-angle vertical curve steering system according to claim 1, characterized in that: The data analysis control module includes: A data analysis unit, used to receive the initial tilt data sent from the tilt monitoring module and / or the real-time tilt data sent from the feedback adjustment module, and perform preliminary processing on the tilt data using statistical filtering technology and digital signal processing technology; The command decision unit is used to receive the tilt data after preliminary processing, and use the dynamic attitude adjustment algorithm to compare and analyze with the preset design parameters, and decide whether the tilt angle of the TBM needs to be adjusted based on the analysis results to obtain an adjustment command; A feedback receiving unit is used to send the adjustment instruction generated by the instruction decision unit to the auxiliary head-up execution module, and at the same time receive the implementation effect data from the feedback adjustment module; Among them, the implementation effect data includes posture data, cylinder extension and retraction data, and environment and equipment status data.

4. A TBM large-angle vertical curve steering system according to claim 1, characterized in that: The dynamic posture adjustment algorithm is implemented based on the fusion of fast Fourier transform, autoregressive moving average model, decision tree and convolutional neural network. The dynamic posture adjustment algorithm compares and analyzes the preset design parameters in the following process: Fast Fourier transform is used to extract periodic variation features from tilt data, and an autoregressive moving average model is used to predict tilt trends and short-term tilt behavior. Apply the integrated learning model, combined with decision tree and convolutional neural network, compare and analyze the extracted features with the preset design parameters to obtain the analysis results; Based on the analysis results, the prediction model is used to dynamically adjust the control parameters to optimize the TBM's tilt angle adjustment and generate adjustment instructions.

5. A TBM large-angle vertical curve steering system according to claim 4, characterized in that: The fast Fourier transform is used to convert the tilt data in the time domain into frequency domain data and obtain the period; Furthermore, the autoregressive moving average model transforms the period obtained by fast Fourier transform into a time-related periodic feature, and integrates the extracted periodic feature into the autoregressive moving average model as an exogenous variable.

6. A TBM large-angle vertical curve steering system according to claim 4, characterized in that: The convolution hierarchy in the convolutional neural network is used to extract and learn high-level features of the data, and then the high-level features are decided and classified through a decision tree; Furthermore, after the convolutional neural network undergoes convolution and pooling operations, the output is compressed and extracted into high-level features to form a feature vector.

7. A TBM large-angle vertical curve steering system according to claim 4, characterized in that: The generation process of the adjustment instruction is specifically as follows: Build a predictive model that describes how the TBM's tilt angle responds to control inputs; The model predictive control strategy dynamically generates control commands by solving a receding horizon optimization problem to minimize the cumulative cost within the future prediction horizon, which includes state error and control efficiency; The optimization problem is solved in real time using a sequential quadratic programming numerical optimization method. Take the TBM to which the first control input is applied, update the state based on the latest tilt measurement in the next control cycle, and solve the optimization problem again. Convert control input into adjustment instructions.

8. The TBM large-angle vertical curve steering system according to claim 1 is characterized in that: The auxiliary head-lifting execution module includes a mechanical device module and a control system module; Wherein, the mechanical device module includes a hydraulic cylinder and a connecting mechanism; And, the control system module adjusts the posture of the TBM front shield using the auxiliary head-up system and the data analysis control module.

9. A TBM large-angle vertical curve steering system according to claim 8, characterized in that: The process of adjusting the TBM front shield posture by the control system module is specifically as follows: The control system module receives the adjustment instruction from the data analysis control module and extracts the front shield tilt angle in the adjustment instruction; Using the PID control algorithm, the difference between the target angle and the current angle is calculated according to the received adjustment command, an error signal is generated, and a control signal is calculated to adjust the driving voltage and current of the hydraulic cylinder to change the telescopic length of the cylinder; The hydraulic cylinder is connected to the front shield through a connecting rod and a joint to adjust the tilt angle of the front shield.

10. A TBM large-angle vertical curve steering system according to claim 1, characterized in that: The feedback adjustment module comprises: The second sensor unit is used to monitor the real-time tilt data and posture of the TBM in real time by using a tilt sensor and a gyroscope; the data feedback unit is used to transmit the real-time tilt data to the data analysis control module through wireless communication technology.

Images