Intelligent control method and system for cantilever type tunneling

Abstract

The invention discloses an intelligent control method and system for cantilever type tunneling. The invention relates to the technical field of the full-section tunneling process. The coding layer receives an input sequence S, calculates forward and backward candidate hidden states through multiple groups of forward and backward GRU (gated loop) units, forms a hidden state sequence H, inputs the hidden state sequence H into the fuzzy processing layer for fuzzy operation processing, and generates a corrected hidden state sequence H'into the decoding layer; according to the method, various tunneling parameters such as total thrust, cutterhead power and torque are comprehensively considered, a rich input sequence is formed through heat coding operation, comprehensive information is provided for the model, and therefore the accuracy of tunneling speed prediction is improved. The bidirectional Seq-Bi-GRU neural network can capture time sequence dependence and long-term memory effect in data, and compared with a traditional BPNN neural network, the method has higher generalization ability and prediction precision when processing complex sequence data.

Description

Technical Field

[0001] The present invention relates to the technical field of full-face tunneling process, and in particular to a prediction and control technology of a PR value (penetration rate) of a full-face hard rock tunnel boring machine (TBM), belonging to the direction of model predictive control (MPC), and in particular to an intelligent control method and system for cantilever tunneling. Background Art

[0002] With its significant advantages such as convenience, high efficiency, economy and environmental friendliness, TBM has become an indispensable and important technical means in many fields such as urban transportation construction, coal mine tunnel excavation, water conservancy and hydropower tunnel construction. Especially in areas with significant terrain height differences such as water network branch line construction, the use of small-diameter TBM for construction has become an inevitable choice.

[0003] However, the construction performance of TBM is extremely sensitive to construction geological conditions, especially during excavation in composite strata (composite strata refer to a combination of two or more strata with significant differences in geotechnical properties, engineering geological characteristics and hydrogeological conditions within the tunnel excavation section or along the excavation extension direction).

[0004] When geological conditions do not match TBM excavation parameters, it will directly lead to a significant decrease in excavation efficiency, which will in turn cause a series of problems such as rising construction costs and extended construction periods. For example, in the construction of small-diameter tunnels, the setting of TBM excavation parameters is more difficult due to the narrow tunnel space. This not only increases the uncertainty of construction, but also significantly increases the fluctuation range of the PR value. Therefore, as a key indicator for measuring TBM construction efficiency, accurate prediction of excavation speed is of great significance for TBM parameter control, construction risk control, and construction period estimation.

[0005] To this end, many existing technologies attempt to achieve this goal, such as: (1) A TBM cutterhead speed adaptive control method is disclosed (publication date ), which adjusts the cutterhead speed according to real-time parameters by establishing a relevant database and random forest algorithm optimization model, realizes adaptive control of the cutterhead speed, and can improve the efficiency of construction advancement and the accuracy of TBM operation. However, in the process of excavation in composite formations, intelligent control of the cutterhead speed is not equivalent to improving the overall construction efficiency of the TBM, because the independent variables of TBM excavation parameters include total thrust, cutterhead power, cutterhead torque, cutterhead penetration, FPI (field penetration), TPI (torque penetration index) and current excavation speed in addition to the cutterhead speed.

[0006] (2) A TBM excavation speed intelligent control method based on XGBoost algorithm was disclosed (publication date ), an intelligent speed control model is established based on the XGBoost algorithm, and the recommended excavation speed of the TBM is pushed in real time. The recommended parameters are written through the PLC control system to realize the intelligent control of the TBM excavation speed. However, the model of this solution is a regression task based on excavation environment parameters such as historical rock parameters. If the current construction environment does not have data that can perform the regression task or the deviation is large, the confidence of the recommended excavation speed is not high.

[0007] (3) A multi-objective optimization method for shield construction parameters based on machine learning and improved genetic algorithm was published (publication date: ), which establishes two SVM prediction models of shield construction parameters and shield horizontal and vertical postures, and an LGBM prediction model of shield construction parameters for final surface settlement, obtains the Pareto optimal solution set based on the improved NSGA-III algorithm, and determines the control range of shield construction parameters that meet the shield vertical and horizontal posture correction errors and surface settlement deformation limits. However, this scheme is more inclined to the optimal solution of TBM excavation efficiency at different times; from a macro perspective, this scheme that focuses on the optimal solution at every moment is essentially trapped in the defects of the local optimal solution of the genetic algorithm, and is not a truly optimal solution for the overall construction and its excavation speed.

[0008] (4) A mechanism- and data-driven method for predicting shield machine excavation speed and surface settlement was disclosed (published on ), which is used to predict the speed of the shield machine and the surface settlement during the construction of the shield machine. It includes starting from geotechnical mechanics after geological exploration, considering the mechanical properties of the soil, establishing a force model for the cutter to cut the soil, and further deriving the calculation formula for the horizontal cutting force of the cutter; further introducing the cutting force as an input parameter into the dual-drive prediction model as an input parameter of the dual-drive prediction model, establishing a tunneling speed and surface settlement prediction model of the support vector machine optimized by the improved particle swarm algorithm, realizing the prediction of tunneling speed and surface settlement, and greatly improving the accuracy of subsequent tunneling speed and surface settlement prediction. Although this scheme is based on the force model of the cutter cutting the soil, and derives the horizontal cutting force of the cutter for optimization at every moment, this scheme is only applicable to the single-input and single-output data-driven mode of "horizontal cutting force-surface settlement prediction". When considering multiple independent variables, the model does not support situations where the input and output sequences are of unequal length, and is therefore not robust enough to support more complex data-driven processing and more accurate data prediction. In the actual tunnel excavation process, the dimensions and quantity of input data at different time points may vary due to changes in geological conditions and adjustments to the tunnel boring machine's operating strategy. The model fails to effectively handle the mapping relationship of such unequal-length sequences, limiting its application in complex and changing construction environments.

[0009] (5) The document "Yang Yaohong, Han Xingzhong, Zhang Zhixiao, et al. Prediction of the speed interval of a small-diameter tunnel boring machine in composite strata [J]. Science, Technology and Engineering, 2023, 23(34): 14638-14650." discloses a BPNN neural network, which establishes a speed interval prediction model based on the Bootsurap resampling method to predict the PR value. However, this scheme also does not support the case where the input and output sequences are not of equal length. Although this scheme adopts the bootstrap sampling technology combined with random distribution to solve the defects of the regression model in the above-mentioned traditional technology, the Bootstrap resampling technology constructs multiple sub-sample sets by repeatedly extracting the original data set to increase the robustness of the model and reduce the risk of overfitting. However, this method is essentially a random resampling technology based on finite samples. When the sample size is limited or the sample distribution does not represent the overall distribution, the stability and representativeness of the results obtained by Bootstrap sampling will be questioned. Moreover, when combined with random distribution for statistical inference, although this method can generate estimates of the prediction interval, these estimates are often based on the empirical distribution of the sample data rather than the true overall distribution. Therefore, the obtained statistical results (such as confidence intervals) may not accurately reflect the true distribution characteristics of the actual PR value of the tunneling speed, resulting in a significant reduction in the accuracy and reliability of statistical inference. Furthermore, the combination of Bootstrap sampling and random distribution cannot directly provide an accurate description of the entire confidence interval distribution, and the model faces challenges when making statistical inferences. This limitation is particularly prominent when it is necessary to accurately evaluate the uncertainty of the PR value of the tunneling speed, conduct risk analysis or make decisions.

[0010] To this end, the present invention proposes an intelligent control method and system for cantilever tunneling. Summary of the invention

[0011] In view of this, the present invention hopes to provide an intelligent control method and system for cantilever tunneling to solve or alleviate the technical problems existing in the prior art, namely, how to consider the support of multi-independent variable data drive (i.e., support input and output sequences of unequal length) in the execution of tunneling tasks by TBM, and further improve the confidence of the estimated BP value (i.e., provide an accurate description of the entire confidence interval distribution) under the premise of avoiding falling into the local optimal solution, and further improve the control effect of TBM. The technical solution of the present invention is implemented as follows: First, intelligent control method for cantilever tunneling: 1. Overview: The present invention aims to predict and control the tunneling speed of TBM through a data-driven method. First, the tunneling parameters are collected as independent variables X using the TBM sensor system to form an input sequence S, and the future tunneling speed PR is set as the dependent variable Y. Then, a bidirectional Seq-Bi-GRU neural network model is constructed, and the Seq2Seq architecture is combined to handle the problem of unequal input and output lengths. A fuzzy processing layer is added to the model to reduce data noise and generate a modified hidden state sequence; the decoding layer uses the DS evidence theory algorithm to fuse multiple evidences to improve the prediction confidence, and finally outputs the most credible tunneling speed prediction value Y'. Finally, according to the difference between the predicted value Y' and the actual tunneling speed PR', the control electrical signal is calculated by the PID controller to achieve precise control of the TBM tunneling speed. The integration of deep learning, fuzzy processing and evidence theory effectively improves the accuracy of tunneling speed prediction and control precision, providing strong support for efficient and safe tunneling of TBM.

[0012] (II) Technical solution: To achieve the above objectives, the present invention selects to perform the following operation steps.

[0013] 2.1 Step S1, Data Engineering: At the current time step t, the independent variables X(t) of the TBM excavation parameters are collected through the TBM's own sensor system or control system, including the total thrust G composed by heat coding operation. t (unit kN), cutter head power C p (Unit: kW), cutter head torque C t (unit kN·m), cutter penetration C h (Unit: mm·r -1 ), Cutter speed C s (Unit: min·r -1 ), field penetration FPI (unit kN·cutter -1 ), torque penetration index TPI [unit (kN·m)·(mm·r -1 ) -1 ] and the current excavation speed PR' (in m·h -1 ); let the future excavation speed PR be the dependent variable Y; and let the independent variable X form the input sequence S.

[0014] 2.1.1 Step S100, construct independent variable X(t): At the current time step t, the independent variable X(t) collected by the TBM’s own sensor system or control system is: X(t)=[G t ,C p ,C t ,C h ,C s ,FPI,TPI,PR′].

[0015] 2.1.2 Step S101, construct an input sequence S: The independent variable X(t) constitutes the input sequence S, which is represented as a set of independent variables at a series of time steps: S=[X(1),X(2),…,X(t),…,X(T)]; Where T represents the total time step of the input sequence. The independent variable X(t) at each time step t contains the eight tunneling parameters mentioned above.

[0016] 2.2 Step S2, bidirectional Seq-Bi-GRU-DS fuzzy neural network operation: In order to support the problem of unequal length of input and output between input sequence S and dependent variable Y, a bidirectional Seq-Bi-GRU neural network is constructed by combining the Seq2Seq model with the bidirectional Bi-GRU neural network to realize the "encoding layer-decoding layer" processing to support unequal length of input and output sequences; at the same time, a fuzzy processing layer is further added between the encoding layer and the decoding layer to perform fuzzy operation on the hidden state sequence output by the encoding layer, forming a bidirectional Seq-Bi-GRU fuzzy neural network, which reduces the noise and uncertainty in the data through fuzzy processing and generates a modified hidden state sequence. In the decoding layer, the DS evidence theory algorithm is used to calculate the joint trust function, which can fuse multiple evidences (hidden states of the decoding layer) and calculate a comprehensive trust, thereby improving the confidence of the prediction.

[0017] Therefore, the model of this step is a bidirectional Seq-Bi-GRU-DS fuzzy neural network as a whole. The process includes: the encoding layer receives the input sequence S, calculates the candidate hidden states of the forward and backward directions through multiple groups of GRU (gated recurrent) units, and forms a hidden state sequence H, which is input to the fuzzy processing layer for fuzzy operation processing, and generates a modified hidden state sequence H' to the decoding layer; at the same time, a backtracking layer is added between the "encoding layer-decoding layer" to modify the hidden state sequence H' of the previous time step. t-1 , perform residual backtracking through the least squares method, use historical prediction errors for compensation, and generate an intermediate vector C; use the intermediate vector C as the input of the decoding layer; The decoding layer is based on the intermediate vector C and the modified hidden state sequence H'. It also calculates the forward and backward candidate hidden states through multiple groups of forward and backward GRU units to form a prediction sequence S' of a set of dependent variables Y; the confidence of different dependent variables Y in the prediction sequence S' is calculated through the DS evidence theory algorithm, and the most credible final dependent variable Y' is further extracted.

[0018] 2.2.1 Step S200, coding layer processing: For each time step t, the forward GRU unit receives the input sequence S and the hidden state of the previous time step , calculate the forward candidate hidden state for the current time step and the forward hidden state ; Then similarly, for each time step t (from T to 1), the backward GRU unit receives the input sequence S and the hidden state of the next time step , calculate the backward candidate hidden state of the current time step and the backward hidden state ; Finally, for each time step t, the forward hidden state and the backward hidden state Merge to form the final hidden state h(t).

[0019] 2.2.2.1 Forward calculation, backward calculation and their splicing: (1) Forward hidden state sequence : ; (2) Backward hidden state : ; (3) Hidden state sequence H: ; in, , [·;·] represents vector concatenation operation.

[0020] 2.2.2.2 The internal data of the GRU unit is driven by: (1) Update gate: ; in, represents the sigmoid activation function, W z and U z is the weight matrix, b z is the bias vector; z t is the output of the update gate.

[0021] (2) Reset gate: ; Among them, W r and U r is the weight matrix, b r is the bias vector; r t is the output of the reset gate.

[0022] (3) Candidate hidden states ; Among them, tanh represents the hyperbolic tangent activation function, ⊙ represents element-wise multiplication (Hadamard product), and W h and Uh is the weight matrix, b h is the bias vector.

[0023] (4) Actual hidden state .

[0024] 2.2.2 Step S201, fuzzy processing layer processing: (1) Based on N fuzzy rules, each rule consists of a premise part (fuzzy set) and a conclusion part (linear transformation function); a fuzzy set is a set of "IF-THEN" rule statements: ; Among them, A i is a fuzzy set, F i is a linear transformation function.

[0025] (2) Fuzzy set A i is a fuzzy set defined on the hidden state space, including Gaussian fuzzy sets or triangular fuzzy sets. Gaussian fuzzy sets are: ; Among them, μ i and are the mean and standard deviation of the Gaussian fuzzy set respectively.

[0026] (3) Linear transformation: F i (h(t)) = W i h(t) + b i ; Among them, W i is the weight matrix, b i is the bias vector.

[0027] (4) Fuzzy operation processing: For each time step t and each fuzzy rule R i , calculate h(t) belongs to the fuzzy set A i The membership degree α i (t): ; Then, using the membership degree α i (t) for the linear transformation F i (h(t)) is weighted averaged to obtain the corrected hidden state ; Modify the hidden state sequence H' = [h'(1), h'(2), ..., h'(T)].

[0028] 2.2.3 Step S202, backtracking layer processing: (1) Calculate the residual: Let E t−1 is the prediction error of the previous time step, that is, the actual value With the predicted value The difference between: ; (2) Least squares backtracking: Use the least squares method to fit the residuals and obtain the residual backtracking matrix W e , so that: ; Where W is the backtracking matrix, ∥⋅∥ 2 is the square of the 2-norm of the vector.

[0029] Solving for W e Available through: ; Among them, ┳ represents the transpose operation.

[0030] (3) Using the obtained residual backtracking matrix W e And the modified hidden state sequence H′ t−1 , generate the intermediate vector C: .

[0031] (4) Take the intermediate vector C as the input of the decoding layer: sum the intermediate vector C with the encoding layer output H at the current time step t Combined as the input of the decoding layer. Here, the combination method can be concatenation method: H′′ t =f(H t ,C); Where f represents the concatenation operation function: H′′t=[H t ;C].

[0032] 2.2.4 Step S203, decoding layer processing: (1) Similar to step S200, it also includes the calculation of the forward GRU unit, the calculation of the backward GRU, and the splicing operation. The difference is that this step is a decoding operation, while S200 is an encoding operation: ; ; ; in, represents the candidate hidden state computed by the forward GRU unit at time step t. → indicates that this is a forward (i.e., from left to right, or in chronological order) computation. h′ is the representation of the hidden state. f indicates that it is the output of the forward GRU. t is the index of the time step. GRU f Represents the data-driven operation of the forward GRU unit (see step S200 for details). represents the candidate hidden state calculated by the forward GRU unit at time step t−1. It is the hidden state of the previous time step of the current time step t and is used to transfer information in the forward GRU. represents the candidate hidden state calculated by the backward GRU unit at time step t. b Represents the data-driven operation of the backward GRU unit (see step S200 for details). Represents the candidate hidden state calculated by the backward GRU unit at time step t+1.

[0033] (2) Use hidden state to represent h' t And the intermediate vector C generates the prediction sequence S' of the dependent variable Y. A fully connected layer NN can be used: ; S' is a sequence containing the predicted values ​​of multiple dependent variables Y: S' = [Y'1, Y'2, ..., Y' N ].

[0034] (3) Use the DS evidence theory algorithm to calculate the confidence of different dependent variables Y in the prediction sequence S': that is, for each dependent variable Y' i , calculate its confidence Bel(Y' i ): (3.1) Identification framework Θ and its propositions: The identification framework Θ is defined as the set of all possible excavation speed ranges. For example, the excavation speed is divided into several intervals, each of which represents a proposition: Let Θ = {Y1, Y2, ..., Y n}, where Y i The proposition that the excavation speed falls within the i-th interval.

[0035] (3.2) Execute the basic probability assignment mechanism: a way to assign confidence to each proposition based on the available evidence. Here, we can assign basic probabilities based on the information about tunneling speed provided by the TBM sensor data, that is, calculate each proposition Y i The basic probability distribution m(Y i ). This basic probability is expressed as the likelihood or confidence that the excavation speed falls within this interval: Considering the existence of subset structure, and our essential idea is to calculate the confidence more accurately, we need to consider all the sets containing Y i The basic probability distribution of the subset of ′. Let P(Y i ) represents the proposition Y i The set of all subsets of (including Y i itself), then the confidence Bel(Yi ′ ) is calculated as: ; Based on this idea, it can be further expanded within the framework of the DS evidence theory algorithm: not limited to the information about the excavation speed provided by the TBM sensor data, but also based on the external detection of the TBM excavation speed (such as underground radar) to obtain two sources of evidence, then: the Dempster combination rule is needed to allow us to combine the basic probability distributions from different evidence sources to obtain a new, comprehensive basic probability distribution. This comprehensive basic probability distribution can then be used to calculate the confidence ; Among them, m(Y i ) is the basic probability distribution after combination. m1(A) and m2(B) are the basic probability distributions of the two evidence sources (the sensor data of the TBM and the excavation speed of the external detection TBM), respectively. A and B are propositions in the identification framework Θ, which come from the two evidence sources, respectively. K is a normalization constant used to ensure that the sum of the basic probability distributions after combination is 1: ; Among them, Ø represents the empty set; represents the sum of the products of the basic probability assignments of all pairs of propositions that have no intersection.

[0036] Once the combined basic probability distribution m(Y i ), we can use it to calculate the confidence Bel(Y i ′). If the predicted value Y i ′ directly corresponds to a proposition Y i (depends on the number of propositions, that is, if conditions permit, the more propositions the better), then the confidence Bel(Y i ′) can be simply represented by m(Y i ) gives: Bel(Y' i ) = m(Y i ); If the predicted value Y i ' does not correspond to a proposition Y i , then it must also be in Y i-1 and Y i+1 In the interval, we can determine whether it belongs to Y according to its trend. i-1 or Y i+1 .

[0037] Furthermore, by combining information from different sources, we can reduce the errors or uncertainties that may exist in a single source of evidence, thereby improving the accuracy of the advance rate prediction. In complex or uncertain environments, a single source of evidence may not provide sufficiently reliable information. By combining multiple sources of evidence, we can make the system more robust and better able to cope with various uncertainties and changes. Accurate advance rate prediction is essential for the operation and maintenance of TBMs. By combining information from multiple sources of evidence, we can provide decision makers with more comprehensive and accurate data support to help them make more informed decisions.

[0038] (4) Based on the confidence level, extract the most credible final dependent variable Y', that is, the dependent variable with the highest confidence level: .

[0039] 2.3 Step S3, execution control: Based on the difference and direction between the final dependent variable Y' and the current advance speed PR', a PID controller system is used to calculate to obtain an electrical signal for controlling the advance speed of the TBM and control the TBM.

[0040] Error e(t) = Y' set -PR'; Integral term: I(t) = I(t-1) + e(t)·△t; Differential term: ; Proportional term: P(t) = K p ·e(t); Control electrical signal u(t) = P(t) +K i I(t) + K d ·D(t); Among them, K p is the proportional gain; K i is the integral gain; K d is the differential gain; t is the time gain. The set value Y' is the target excavation speed; I(t) is the integral term value at the current moment; I(t-1) is the integral term value at the previous sampling moment; e(t-1) is the error at the previous sampling moment; P(t) and D(t) are the differential term value and proportional term value at the current moment.

[0041] 3. Mechanisms for solving technical problems: 3.1 Avoid inaccurate estimation when bootstrap sampling technology is combined with random distribution to generate prediction intervals: The traditional bootstrap sampling technique combined with random distribution to generate prediction intervals often relies on a large number of repeated sampling and simulations to estimate the distribution of predicted values. However, this method may not be able to accurately capture the true distribution of data when faced with high-dimensional, nonlinear or time-dependent data, resulting in inaccurate prediction interval estimates. The present invention constructs a bidirectional Seq-Bi-GRU-DS fuzzy neural network to directly learn the complex mapping relationship between the input sequence and the output sequence without repeated sampling and simulation. At the same time, the DS evidence theory algorithm is used to fuse multiple evidences and calculate the joint trust function, which can more accurately describe the uncertainty of the predicted value, thereby generating a more reliable prediction interval.

[0042] 3.2 Overcoming the limitations of BPNN neural network in processing complex sequence data: When processing complex sequence data, BPNN (back propagation neural network) often has difficulty capturing the temporal dependencies and long-term memory effects in the data. The bidirectional Seq-Bi-GRU-DS fuzzy neural network used in the present invention can simultaneously consider past and future information through forward and backward GRU units, effectively capturing the temporal characteristics in the data. In addition, the introduction of the fuzzy processing layer further reduces the noise and uncertainty in the data, improving the robustness of the model and the prediction accuracy.

[0043] 3.3 Avoid local optimization and improve the confidence of the estimated BP value: In order to prevent the model from falling into a local optimal solution, the present invention adds a backtracking layer between the "encoding layer-decoding layer". The backtracking layer uses the least squares method to perform residual backtracking, compensates the current prediction based on the historical prediction error, and generates a more accurate intermediate vector as the input of the decoding layer. This design enables the model to dynamically adjust the prediction strategy to avoid converging to the local optimal solution too early during the training process. At the same time, the application of the DS evidence theory algorithm also provides a higher confidence level for inferring BP values, because it can fuse multiple pieces of evidence and calculate a comprehensive degree of trust, thereby more comprehensively evaluating the uncertainty of the prediction results.

[0044] 3.4 Supports multi-variable data drive and input / output sequences of unequal length: The present invention naturally supports the problem of unequal lengths of input and output sequences by constructing a bidirectional Seq-Bi-GRU-DS fuzzy neural network. The design of the encoding layer and the decoding layer enables the model to flexibly process input sequences of different lengths and generate output sequences of corresponding lengths. In addition, the data-driven design of multiple independent variables enables the model to comprehensively consider the impact of multiple excavation parameters on the excavation speed, thereby improving the accuracy and generalization ability of the prediction.

[0045] 3.5 Further improve the control effect of TBM: Based on the excavation speed predicted by the above model, the present invention accurately controls the TBM through a PID controller system. The PID controller calculates the electrical signal for controlling the excavation speed of the TBM according to the difference and direction between the predicted value and the actual value, and adjusts the excavation parameters in real time to achieve the desired excavation speed. This prediction-based control strategy not only improves the excavation efficiency, but also enhances the stability and safety of the excavation process.

[0046] Second, intelligent control system for cantilever tunneling: The system includes a processor and a memory connected to the processor, wherein program instructions are stored in the memory. When the program instructions are executed by the processor, the processor executes the intelligent control method as described above; after the control electrical signal u(t) is obtained, it is input into the TBM control system, and the TBM control system controls the actuator of the TBM to perform corresponding operations.

[0047] Compared with the prior art, the present invention has the following beneficial effects: 1. Improve prediction accuracy: This invention comprehensively considers a variety of tunneling parameters, such as total thrust, cutterhead power, torque, etc., and forms a rich input sequence through heat encoding operations, providing comprehensive information for the model, thereby improving the accuracy of tunneling speed prediction. The bidirectional Seq-Bi-GRU neural network can capture the temporal dependencies and long-term memory effects in the data. Compared with the traditional BPNN neural network, it has stronger generalization ability and prediction accuracy when processing complex sequence data. The fuzzy processing layer is used to reduce data noise, and the DS evidence theory algorithm fuses multiple evidences to calculate the joint trust, which further improves the confidence and reliability of the prediction results.

[0048] 2. Avoid local optimization and improve robustness: The backtracking layer of the present invention uses the least squares method to perform residual backtracking, and compensates the current prediction based on the historical prediction error, avoiding the model from falling into the local optimal solution and improving the robustness and stability of the prediction. The hidden state sequence output by the coding layer is processed by fuzzy operation, which reduces the uncertainty and noise in the data and enhances the model's ability to handle abnormal data.

[0049] 3. Accurately control the TBM excavation speed: Based on the difference between the predicted value and the actual excavation speed, the present invention adjusts the excavation parameters in real time through the PID controller, thus achieving accurate control of the TBM excavation speed and improving excavation efficiency and safety. The PID controller can dynamically adjust the control strategy according to the prediction results, so that the TBM can maintain a stable excavation speed under different geological conditions.

[0050] 4. Support for unequal lengths of input / output sequences: The combination of the Seq2Seq architecture and the bidirectional Bi-GRU neural network adopted in the present invention naturally supports the problem of unequal lengths of input and output sequences, allowing the model to flexibly handle tunneling parameter sequences and tunneling speed sequences of different lengths.

[0051] 5. Improve construction efficiency: The present invention reduces unnecessary downtime and energy consumption during excavation by accurately predicting and controlling the excavation speed, thereby improving excavation efficiency. Accurate prediction and control help workers avoid unexpected situations during excavation, such as machine jams and over-excavation, and reduce excavation costs and safety risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0053] Figure 1 It is a schematic diagram of the method flow of the present invention; Figure 2 Schematic diagram of the composition of the input sequence S of the present invention; Figure 3 It is a schematic diagram of the coding layer architecture of the present invention; Figure 4 It is a schematic diagram of a GRU unit of the present invention; Figure 5 It is a schematic diagram of the fuzzy processing layer architecture of the present invention; Figure 6 It is a schematic diagram of the backtracking layer architecture of the present invention; Figure 7 It is a schematic diagram of the decoding layer architecture of the present invention; Figure 8 A schematic diagram of a data-driven method of the DS evidence theory algorithm of the present invention; Fig. 9 It is a schematic diagram of the system composition of the present invention; Fig.10 It is a comparison diagram of the test examples of the present invention; Fig.11 It is a schematic diagram of the finite element site model of the test example of the present invention; Fig.12 It is a comparative schematic diagram of the first finite element simulation of the test example of the present invention; Fig.13 It is a comparative schematic diagram of the second finite element simulation of the test example of the present invention. DETAILED DESCRIPTION

[0054] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below; It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0055] Embodiment 1: Figure 1 As shown, this embodiment discloses an intelligent control method for cantilever tunnel excavation, which specifically includes the following steps S1 to S3.

[0056] In this embodiment, if Figure 2 As shown, regarding step S1, data engineering: This step aims to collect and process key parameters in the TBM excavation process to build a model that supports multi-independent variable data-driven, thereby achieving precise control. At the current time step t, through the TBM's own sensor system or control system, collect the independent variables X(t) of the TBM excavation parameters, including the total thrust G composed by heat encoding operation t (unit kN), cutter head power C p (Unit: kW), cutter head torque C t (unit kN·m), cutter penetration C h (Unit: mm·r -1 ), Cutter speed C s (Unit: min·r -1 ), Field Penetration FPI (unit kN·cutter -1 ), torque penetration index TPI [unit (kN·m)·(mm·r -1 ) -1 ] and the current excavation speed PR' (in m·h -1 ); let the future excavation speed PR be the dependent variable Y; let the independent variable X form the input sequence S; Specifically, step S100, constructing the independent variable X(t): at the current time step t, the independent variable X(t) collected by the TBM's own sensor system or control system is: X(t)=[G t ,C p ,C t ,Ch ,C s ,FPI,TPI,PR′]; Specifically, step S101, constructing an input sequence S: constructing an input sequence S from independent variables X(t), expressed as a set of independent variables at a series of time steps: S=[X(1),X(2),…,X(t),…,X(T)]; Where T represents the total time step of the input sequence. The independent variable X(t) at each time step t contains the eight tunneling parameters mentioned above.

[0057] It can be understood that in this step, by collecting multiple independent variables that are closely related to excavation efficiency, we construct a comprehensive input feature space, which provides a rich source of information for the model. This helps the model to more accurately capture the complex relationships in the excavation process and improve the accuracy of prediction. Arranging the independent variables in time step order to form an input sequence S enables the model to learn the temporal dependency in the excavation process. This dependency is crucial for predicting future excavation speed PR because it reflects the trend of excavation status over time. A longer input sequence S can contain more historical information, which helps the model capture the long-term memory effect in the excavation process. It helps the model consider more distant past states when predicting and improves the robustness of the prediction.

[0058] It should be pointed out that in practical applications, the length of the input sequence S (i.e., the time step T) can be determined according to specific needs and data availability. Longer input sequences can contain more historical information, which helps the model to more accurately capture the temporal dependency and long-term memory effect of the data; however, it will also increase the complexity and computational complexity of the model. Therefore, a trade-off needs to be made between the accuracy of the model and computational efficiency.

[0059] In this embodiment, regarding step S2, this step aims to build a neural network model that can handle the problem of unequal lengths of input and output sequences, while improving the confidence of prediction. To this end, we combine the Seq2Seq model with the bidirectional Bi-GRU neural network, and introduce the fuzzy processing layer and DS evidence theory algorithm to construct a bidirectional Seq-Bi-GRU-DS fuzzy neural network.

[0060] In this scheme, regarding the construction of a bidirectional Seq-Bi-GRU neural network: (1) Encoding layer: The encoding layer receives the input sequence S and calculates the candidate hidden states for the forward and backward directions through multiple groups of GRU (gated recurrent units) units. These candidate hidden states comprehensively consider the historical information and future information in the sequence to form a hidden state sequence H containing rich temporal features.

[0061] (2) Decoding layer: The decoding layer makes predictions based on the hidden state sequence H output by the encoding layer. Similarly, multiple groups of forward and backward GRU units are used to calculate candidate hidden states and form a set of prediction sequences S' of the dependent variable Y.

[0062] (3) Fuzzy processing layer: A fuzzy processing layer is introduced between the encoding layer and the decoding layer to perform fuzzy operation on the hidden state sequence H output by the encoding layer. Fuzzy processing can reduce the noise and uncertainty in the data and generate a more robust modified hidden state sequence H'. This step helps to improve the model's ability to handle complex data and improve the accuracy of prediction.

[0063] (4) Backtracking layer and residual backtracking: A backtracking layer is added between the encoding layer and the decoding layer to correct the hidden state sequence H' at the previous time step t-1 Perform residual backtracking. The historical prediction error is calculated by the least squares method, and the error is used to compensate the corrected hidden state of the previous time step to generate an intermediate vector C. This step can use historical information to correct the current prediction and further improve the accuracy of the prediction.

[0064] (5) Application of DS Evidence Theory Algorithm: In the decoding layer, the DS Evidence Theory algorithm is used to calculate the confidence of different dependent variables Y in the prediction sequence S'. DS Evidence Theory can integrate multiple pieces of evidence (i.e., the hidden states of the decoding layer) and calculate a comprehensive confidence level. Based on the confidence level, the most credible final dependent variable Y' is extracted from the prediction sequence S'. This step ensures that the prediction results output by the model have a high degree of confidence and reliability.

[0065] In general, this step realizes the "encoding layer-decoding layer" processing by constructing a bidirectional Seq-Bi-GRU neural network, which effectively supports the problem of unequal lengths of input and output sequences. This enables the model to handle complex data in the actual excavation process more flexibly, improving the applicability and generalization ability of the model. The fuzzy processing layer can reduce the noise and uncertainty in the data and generate a more robust correction of hidden state sequences. At the same time, the application of the DS evidence theory algorithm can integrate multiple evidences and calculate a comprehensive trust, further improving the confidence of the prediction. The prediction results output by the model are more reliable and help guide actual excavation operations. The introduction of the backtracking layer enables the model to use historical prediction errors to correct the current prediction. This step can reduce the prediction deviation caused by the accumulation of historical errors and improve the accuracy and stability of the prediction.

[0066] Specifically, Figure 3~4As shown, regarding step S200, the encoding layer processes: the encoding layer is responsible for converting the input sequence S into a hidden state sequence H, providing a basis for subsequent calculations. This step achieves comprehensive encoding of the input sequence through forward and backward GRU units.

[0067] For each time step t, the forward GRU unit receives the input sequence S and the hidden state of the previous time step , calculate the forward candidate hidden state for the current time step and the forward hidden state ; Then similarly, for each time step t (from T to 1), the backward GRU unit receives the input sequence S and the hidden state of the next time step , calculate the backward candidate hidden state of the current time step and the backward hidden state ; Finally, for each time step t, the forward hidden state and the backward hidden state Merge to form the final hidden state h(t).

[0068] Specifically, regarding the forward calculation, backward calculation and their concatenation in step S200: (1) Forward hidden state sequence : For each time step t, the forward GRU unit receives the current element in the input sequence S and the hidden state of the previous time step, and calculates the forward candidate hidden state and forward hidden state of the current time step through the internal data-driven mechanism (including update gate, reset gate, calculation of candidate hidden state and actual hidden state): ; (2) Backward hidden state : Similar to the forward calculation, but in the opposite direction. For each time step t (from T to 1), the backward GRU unit receives the current element in the input sequence S and the hidden state of the next time step, performs the same internal data-driven calculation, and obtains the backward candidate hidden state and backward hidden state of the current time step: ; (3) Hidden state sequence H: ; in, , [·;·] represents the vector concatenation operation. T is the length of the input sequence.

[0069] Furthermore, the internal data of the GRU unit is driven by: (1) Update gate: Calculate the update gate output z for the current time step t , which determines the current hidden state h tHow much of the previous time step hidden state h should be retained t-1 , and how many candidate hidden states of the current time step should be received : ; in, represents the sigmoid activation function, W z and U z is the weight matrix, b z is the bias vector.

[0070] (2) Reset gate: Calculate the reset gate output r of the current time step t , which determines how much the hidden state ht-1 of the previous time step should be forgotten at the current time step: ; Among them, W r and U r is the weight matrix, b r is the bias vector.

[0071] (3) Candidate hidden states: ; Among them, tanh represents the hyperbolic tangent activation function, ⊙ represents element-wise multiplication (Hadamard product), and W h and U h is the weight matrix, b h is the bias vector.

[0072] (4) Actual hidden state: According to the output z of the update gate t , the hidden state h of the previous time step t-1 and the candidate hidden state at the current time step Perform linear combination to get the actual hidden state h t : ; It can be understood that through the forward and backward GRU units, the encoding layer can fully capture the temporal features in the input sequence S and realize multi-independent variable data drive. This enables the model to more accurately understand the complex relationship between the input sequence and the output sequence, providing a solid foundation for subsequent prediction and control. The combination of the bidirectional encoding mechanism (forward and backward) in this step enables the model to consider both historical information and future information at the same time, thereby more accurately inferring the BP value (or other related variables). This comprehensive information fusion helps to improve the confidence of the prediction, allowing the model to make more reliable decisions when facing complex excavation environments. The update gate and reset gate mechanism inside the GRU unit enables the model to dynamically adjust the update and forgetting degree of the hidden state. This flexibility helps the model avoid falling into the local optimal solution during training, thereby finding a better global solution.

[0073] Specifically, Figure 5 As shown, regarding step S201, fuzzy processing layer processing: In the intelligent control method for cantilever tunneling, the fuzzy processing layer further extracts and transforms information by fuzzy processing the hidden state sequence H output by the encoding layer to improve the confidence and accuracy of the inferred BP value.

[0074] (1) Define N fuzzy rules. Each rule consists of a premise part (fuzzy set A i ) and the conclusion section (linear transformation function F i ). The fuzzy set Ai describes a certain fuzzy state that the hidden state h(t) may belong to, and the linear transformation function F i It defines how to transform h(t) to obtain the modified hidden state h'(t) when h(t) belongs to this fuzzy state. A fuzzy set is a set of "IF-THEN" rule statements: ; Among them, A i is a fuzzy set, F i is a linear transformation function.

[0075] (2) Fuzzy set A i is a fuzzy set defined on the hidden state space, which is a Gaussian fuzzy set: ; Among them, μ i and are the mean and standard deviation of the Gaussian fuzzy set respectively.

[0076] (3) Linear transformation: F i (h(t)) = W i h(t) + b i ; Among them, W i is the weight matrix, b i is the bias vector.

[0077] The linear transformation is a scaling, translation, or rotation operation on the original hidden state, depending on the weight matrix W. i and the bias vector b i The value of .

[0078] (4) Fuzzy operation processing: For each time step t and each fuzzy rule R i , calculate h(t) belongs to the fuzzy set A i The membership degree α i (t): ; Membership degree α i (t) represents the hidden state h(t) and the fuzzy set Ai The similarity or matching degree between them. i When (t) is close to 1, it means that h(t) belongs to the fuzzy set A to a large extent. i ; When αi(t) is close to 0, it means that h(t) is consistent with the fuzzy set A i Quite a difference.

[0079] Then, using the membership degree α i (t) for the linear transformation F i (h(t)) is weighted averaged to obtain the corrected hidden state ; Modify the hidden state sequence H' = [h'(1), h'(2), ..., h'(T)].

[0080] The modified hidden state sequence H' is a further refinement and transformation of the original hidden state sequence H. It contains more fuzzy information and nonlinear features, providing a richer and more accurate information basis for subsequent prediction or control modules.

[0081] It can be understood that the fuzzy processing layer provides a more detailed and comprehensive description of the hidden state by introducing fuzzy sets and linear transformation functions. This description method can capture the fuzzy relationship and nonlinear characteristics between hidden states, thereby improving the confidence of the inferred BP value (or other related variables). At the same time, the weighted average processing method also reduces the error and uncertainty that may be caused by a single rule. The fuzzy processing layer can process hidden state sequences H of any length from the encoding layer, and each fuzzy rule can be defined and adjusted according to actual needs. This enables the model to flexibly adapt to the multi-independent variable data-driven requirements under different excavation environments and task requirements. The fuzzy processing layer increases the diversity and robustness of the model by introducing multiple fuzzy rules and weighted average processing methods. This design helps the model avoid falling into local optimal solutions during training, thereby finding a better global solution. The modified hidden state sequence H' contains more fuzzy information and nonlinear characteristics, which are essential for guiding TBM (tunnel boring machine) to perform precise and stable excavation operations. Therefore, the introduction of the fuzzy processing layer helps to improve the control effect of TBM and make the excavation process more efficient and safe.

[0082] Specifically, Figure 6 As shown, regarding step S202, backtracking layer processing: This step aims to support multi-independent variable data driving and improve the confidence of the estimated BP value by calculating residuals, backtracking fitting, generating intermediate vectors, and combining the vectors with the encoding layer output. It includes: (1) Calculate the residual: Let E t−1 is the prediction error of the previous time step, that is, the actual value With the predicted value The difference between: ; The residual is the difference between the actual value and the predicted value, which reflects the accuracy of the model prediction. During the excavation process, due to the influence of various factors such as geological conditions and excavation parameters, the actual excavation effect often deviates from the predicted value. Therefore, calculating the residual is the basis for evaluating the model prediction effect and conducting subsequent backtracking processing. By calculating the residual, the deviation in the model prediction can be discovered in time, providing an accurate data basis for subsequent backtracking processing. At the same time, the size of the residual can also be used as one of the important indicators for evaluating model performance.

[0083] (2) Least squares backtracking: Use the least squares method to fit the residuals and obtain the residual backtracking matrix W e , so that: ; Where W is the backtracking matrix, ∥⋅∥ 2 is the square of the 2-norm of the vector. In the backtracking layer processing, the residual backtracking matrix W e It is used to describe the relationship between the residual and the modified hidden state sequence, so as to realize the backtracking processing of the residual.

[0084] Among them, solving W e Available through:

[0085] Among them, ┳ represents the transpose operation.

[0086] It can be understood that the relationship between the residual and the modified hidden state sequence can be accurately fitted through least squares backtracking, thereby achieving accurate backtracking of the residual. This helps to improve the model's adaptability to complex changes in the excavation process and avoid falling into a local optimal solution.

[0087] (3) Using the obtained residual backtracking matrix W e And the modified hidden state sequence H′ t−1 , generate the intermediate vector C: ; This vector contains information about the residual backtracking and can be used as the input of the decoding layer to help the decoding layer better understand and handle the complex changes in the tunneling process. The generation of the intermediate vector C provides the decoding layer with rich residual backtracking information. This helps the decoding layer to fully consider the impact of historical residuals when generating predicted values, thereby improving the accuracy and confidence of the predicted values.

[0088] (4) Take the intermediate vector C as the input of the decoding layer: sum the intermediate vector C with the encoding layer output H at the current time step t Combined as the input of the decoding layer. Here, the combination method can be concatenation method: H′′ t =f(H t,C); Where f represents the concatenation operation function: H′′t=[H t ;C]; It can be understood that by combining the intermediate vector C with the encoding layer output H t Combined with the input of the decoding layer, the historical residual backtracking information and the output information of the encoding layer at the current time step can be fully utilized to improve the accuracy and confidence of the prediction value generated by the decoding layer. At the same time, this combination also supports multi-independent variable data driving with unequal lengths of input and output sequences, allowing the model to handle complex changes in the excavation process more flexibly.

[0089] Specifically, Figure 7 As shown, regarding step S203, decoding layer processing: decoding layer processing is a key step following the encoding layer and the backtracking layer. It is responsible for generating a prediction sequence of the dependent variable Y (such as excavation speed, excavation force, etc.) based on the hidden state information provided by the encoding layer and the residual backtracking information generated by the backtracking layer.

[0090] (1) The calculation of the forward and backward GRU units in the decoding layer is similar to that of the encoding layer, but the difference is that the decoding layer is the process of generating predicted values, while the encoding layer is the process of extracting features from the input sequence. The forward GRU unit processes information from left to right in chronological order, while the backward GRU unit processes information from right to left. The combination of the two can capture the bidirectional dependency in the time series: ; ; ; in, represents the candidate hidden state computed by the forward GRU unit at time step t. → indicates that this is a forward (i.e., from left to right, or in chronological order) computation. h′ is the representation of the hidden state. f indicates that it is the output of the forward GRU. t is the index of the time step. GRU f Represents the data-driven operation of the forward GRU unit (see step S200 for details). represents the candidate hidden state calculated by the forward GRU unit at time step t−1. It is the hidden state of the previous time step of the current time step t and is used to transfer information in the forward GRU. represents the candidate hidden state calculated by the backward GRU unit at time step t. b Represents the data-driven operation of the backward GRU unit (see step S200 for details). Represents the candidate hidden state calculated by the backward GRU unit at time step t+1.

[0091] It can be understood that through the combination of forward and backward GRU units, the decoding layer is able to capture the bidirectional dependencies in the time series, thereby generating more accurate predictions. This bidirectional processing mechanism helps improve the model's ability to predict complex changes in the tunneling process.

[0092] (2) Use hidden state to represent h' t And the intermediate vector C generates the prediction sequence S' of the dependent variable Y. A fully connected layer NN can be used: ; S' is a sequence containing the predicted values ​​of multiple dependent variables Y: S' = [Y'1, Y'2, ..., Y' N ]; Preferably, the fully connected layer is usually represented as a combination of a linear transformation plus a nonlinear activation function. In the decoding layer, the fully connected layer receives the hidden state representation h′ t and the intermediate vector C as input, and outputs the predicted sequence S': S'=σ(W[h' t ;C]+b); Where: W is a weight matrix, which is determined by the dimensions of the input features (i.e., the concatenated hidden state representation and the intermediate vector) and the dimensions of the output prediction sequence. [h′ t ;C] represents the hidden state h′ t and the intermediate vector C are concatenated into a long vector. b is the bias vector, which has the same dimension as the output prediction sequence. σ(⋅) is a nonlinear activation function, such as ReLU, sigmoid, or softmax, which is used to increase the nonlinear expression ability of the model.

[0093] It can be understood that the weight matrix W and the bias vector b jointly perform a linear transformation on the input features. The purpose of this step is to map the input features to a new feature space in order to better capture the linear relationship between input and output. The role of the nonlinear activation function σ(⋅) is to introduce nonlinear expression capabilities into the model. By concatenating the hidden state representation h′ t The fully connected layer can fuse features from different sources by combining the intermediate vector C. This fusion helps the model capture the complex dependencies in the input sequence, thereby improving the accuracy of prediction.

[0094] It can be understood that by combining the hidden state representation and the intermediate vector, the decoding layer can handle the impact of multiple independent variables (such as excavation parameters, geological conditions, etc.) on the excavation effect and realize multi-independent variable data-driven. The intermediate vector C contains the information of residual backtracking, which helps the decoding layer to consider the impact of historical residuals when generating predicted values, thereby improving the confidence of the predicted values. Since the decoding layer is processed based on time steps, it can handle situations where the input and output sequences are of unequal length, which enables the model to adapt more flexibly to complex changes in the excavation process. By combining forward and backward GRU units and residual backtracking information, the decoding layer can more comprehensively capture the dependencies in the time series, thereby avoiding falling into local optimal solutions.

[0095] Further, such as Figure 8 As shown, the DS evidence theory algorithm is used to calculate the confidence of different dependent variables Y in the prediction sequence S': that is, for each dependent variable Y' i , calculate its confidence Bel(Y' i ): (1) Identification framework Θ and its propositions: First, we need to define an identification framework Θ, which includes all possible excavation speed ranges. The excavation speed is divided into several intervals, each of which represents a proposition Y. i , indicating the possibility that the excavation speed falls within this interval: Let Θ = {Y1, Y2, ..., Y n}, where Y i The proposition that the excavation speed falls within the i-th interval.

[0096] (2) Execute the basic probability assignment mechanism: a way to assign confidence to each proposition based on the available evidence. Here, we can assign basic probabilities based on the information about tunneling speed provided by the TBM sensor data, that is, calculate each proposition Y i The basic probability distribution m(Y i ). This basic probability is expressed as the likelihood or confidence that the excavation speed falls within this interval: Based on the TBM’s sensor data or other available evidence, we assign a basic probability m(Y i ), indicating the confidence that the excavation speed falls within this interval. Considering the subset structure, we need to calculate the i The sum of the basic probability distributions of all subsets of to obtain the confidence Bel(Y i ). When there are multiple sources of evidence (such as TBM sensor data and external detection data), we use Dempster's combination rule to combine the basic probability assignments of these evidence sources to obtain a comprehensive basic probability assignment.

[0097] Considering the existence of subset structure, and our essential idea is to calculate the confidence more accurately, we need to consider all the sets containing Y i The basic probability distribution of the subset of ′. Let P(Y i ) represents the proposition Y i The set of all subsets of (including Y i itself), then the confidence Bel(Yi ′ ) is calculated as: ; Furthermore, based on this idea, it can be further expanded within the framework of the DS evidence theory algorithm: not limited to the information about the excavation speed provided by the TBM sensor data, but also based on the external detection of the TBM excavation speed (such as underground radar) to obtain two sources of evidence, then: Dempster combination rule is needed to allow us to combine the basic probability distributions from different sources of evidence to obtain a new, comprehensive basic probability distribution. This comprehensive basic probability distribution can then be used to calculate the confidence level: ; Among them, m(Y i ) is the basic probability distribution after combination. m1(A) and m2(B) are the basic probability distributions of the two evidence sources (the sensor data of the TBM and the excavation speed of the external detection TBM), respectively. A and B are propositions in the identification framework Θ, which come from the two evidence sources, respectively. K is a normalization constant used to ensure that the sum of the basic probability distributions after combination is 1: ; Among them, Ø represents the empty set; represents the sum of the products of the basic probability assignments of all pairs of propositions that have no intersection.

[0098] (3) Once the basic probability distribution m(Y i ), we can use it to calculate the confidence Bel(Y i ′). If the predicted value Y i ′ directly corresponds to a proposition Y i (depends on the number of propositions, that is, if conditions permit, the more propositions the better), then the confidence Bel(Y i ′) can be simply represented by m(Y i ) gives: Bel(Y' i ) = m(Y i ); If the predicted value Y i ' does not correspond to a proposition Y i , then it must also be in Y i-1 and Y i+1In the interval, we can determine whether it belongs to Y according to its trend. i-1 or Y i+1 .

[0099] Furthermore, by combining information from different sources, we can reduce the errors or uncertainties that may exist in a single source of evidence, thereby improving the accuracy of the advance rate prediction. In complex or uncertain environments, a single source of evidence may not provide sufficiently reliable information. By combining multiple sources of evidence, we can make the system more robust and better able to cope with various uncertainties and changes. Accurate advance rate prediction is essential for the operation and maintenance of TBMs. By combining information from multiple sources of evidence, we can provide decision makers with more comprehensive and accurate data support to help them make more informed decisions.

[0100] (4) Based on the confidence level, extract the most credible final dependent variable Y', that is, the dependent variable with the highest confidence level: ; It can be understood that by selecting the predicted value with the highest confidence as the final dependent variable, we can ensure that the control decision is based on the most reliable information, thereby improving the control effect of the TBM. In addition, this method can also automatically adapt to different excavation conditions and environmental changes, enhancing the adaptability and stability of the system. In summary, using the DS evidence theory algorithm to calculate the confidence of different dependent variables Y in the prediction sequence S' is an important part of the intelligent control method for cantilever tunneling. By combining information from multiple evidence sources, calculating confidence and selecting the most credible predicted value, we can more accurately evaluate the uncertainty of the excavation speed, provide decision makers with more comprehensive and accurate data support, and further improve the control effect of the TBM.

[0101] In this embodiment, regarding step S3, control is executed: this step is based on the difference and direction between the final dependent variable Y' (i.e., the target excavation speed) and the current excavation speed PR', and a PID (proportional-integral-differential) controller system is used to perform precise calculations to obtain an electrical signal for controlling the excavation speed of the TBM (tunnel boring machine), and the TBM is controlled in real time accordingly.

[0102] Based on the difference and direction between the final dependent variable Y' and the current advance speed PR', a PID controller system is used to calculate to obtain an electrical signal for controlling the advance speed of the TBM and control the TBM.

[0103] Error e(t) = Y' set -PR'; Integral term: I(t) = I(t-1) + e(t)·△t; where I(t-1) is the integral term value at the previous sampling moment, e(t) is the current error, and △t is the time interval (time gain).

[0104] Differential term: ; The differential term can respond to possible changes in the system in advance and improve the response speed and stability of the control.

[0105] Proportional term: P(t) = K p ·e(t); The proportional term can respond quickly to error changes, but it may cause system oscillation, so it needs to be used in combination with other control terms.

[0106] Control electrical signal u(t) = P(t) +K i I(t) + K d ·D(t); Among them, K p is the proportional gain; K i is the integral gain; K d is the differential gain; t is the time gain. The set value Y' is the target excavation speed; I(t) is the integral term value at the current moment; e(t-1) is the error at the previous sampling moment; P(t) and D(t) are the differential term value and proportional term value at the current moment.

[0107] It is understandable that the PID controller can more accurately control the TBM's excavation speed, reduce errors, and improve control accuracy by comprehensively considering the current error, the accumulation of past errors, and the changing trend of future errors. The integral term can eliminate the static error of the system and improve the stability of the system; the differential term can predict and suppress the oscillation of the system, further enhancing the stability of the system.

[0108] Embodiment 2: This embodiment will further provide a training method for the bidirectional Seq-Bi-GRU-DS fuzzy neural network as described in Embodiment 1, and the training parameters include: (1) Reset the gate weight matrix W r and U r , and the bias vector b r ; (2) Update the gate weight matrix W z and U z , and the bias vector b z ; (3) Weight matrix W of candidate hidden states h and U h , and the bias vector b h ; (4) Linear transformation weight matrix W i and the bias vector b i ; (5) Weight matrix W and bias vector b of the fully connected layer.

[0109] Training methods include: P1. Create a data set D = [(X1, Y1), (X2, Y2), ..., (X n ,Y n )]; Among them, (X i ,Y i ) is the i-th training sample, n is the number of samples, where X i is the historical value of the ith independent variable as described in Example 1, Y i The corresponding X i The value of the dependent variable.

[0110] The dataset D is then divided into a training set and a test set, and the training set is used to perform subsequent training steps.

[0111] P2. Execute steps S200 to S203 as described in the first embodiment.

[0112] P3. Calculate the loss: Use a suitable loss function (such as mean square error, cross entropy, etc.) to calculate the error between the predicted sequence S′ and the target sequence Y to obtain the loss value L. Based on the loss value L, the gradient of the loss function relative to each parameter is calculated using the chain rule.

[0113] For the GRU units in the encoding layer and the decoding layer, the gradients of the update gate, reset gate, candidate hidden state, and actual hidden state with respect to the weight matrix and bias vector are calculated.

[0114] For the blurring layer, the gradients of the weight matrix and bias vector of the linear transformation are calculated.

[0115] For the regression layer, the gradient of the residual regression matrix is ​​calculated.

[0116] For fully connected layers, the gradients of the weight matrix and bias vector are calculated.

[0117] P4. Update parameters: Use the gradient descent algorithm (SGD) to update the values ​​of all weight matrices and bias vectors. For each parameter, update it according to the learning rate and the calculated gradient.

[0118] P5, iterative training: Repeat the steps of forward propagation, loss calculation, back propagation and parameter update, and adjust the batch size according to the performance on the test set to optimize the model performance until the preset number of training rounds is reached or the loss value converges to a satisfactory level.

[0119] Furthermore, the Python execution program of the bidirectional Seq-Bi-GRU-DS fuzzy neural network is as follows: import torch import torch.nn as nn import torch.optim as optim class BiGRUCell(nn.Module): def __init__(self, input_size, hidden_size): super(BiGRUCell, self).__init__() self.hidden_size = hidden_size self.reset_gate = nn.Linear(input_size + hidden_size, hidden_size) self.update_gate = nn.Linear(input_size + hidden_size,hidden_size) self.candidate_state = nn.Linear(input_size + hidden_size,hidden_size) self.sigmoid = nn.Sigmoid() self.tanh = nn.Tanh() def forward(self, x, hidden): r_t = self.sigmoid(self.reset_gate(torch.cat([x, hidden], dim=1))) z_t = self.sigmoid(self.update_gate(torch.cat([x, hidden],dim=1))) h_hat_t = self.tanh(self.candidate_state(torch.cat([x, r_t *hidden], dim=1))) h_t = (1 - z_t) * hidden + z_t * h_hat_t return h_t class FuzzyLayer(nn.Module): def __init__(self, hidden_size, num_rules): super(FuzzyLayer, self).__init__() self.hidden_size = hidden_size self.num_rules = num_rules self.linear_transforms = nn.ModuleList([nn.Linear(hidden_size, hidden_size) for _ in range(num_rules)]) self.means = nn.Parameter(torch.randn(num_rules, hidden_size)) self.stds = nn.Parameter(torch.rand(num_rules, hidden_size)) def forward(self, H): T, hidden_size = H.size() H_prime = torch.zeros_like(H) for t in range(T): memberships = torch.exp(-torch.sum((H[t] - self.means) **2 / (2 * self.stds ** 2), dim=1)) weighted_sum = torch.zeros(hidden_size) for i in range(self.num_rules): weighted_sum += memberships[i] * self.linear_transforms[i](H[t]) H_prime[t] = weighted_sum / torch.sum(memberships) return H_prime class Seq2SeqBiGRUDS(nn.Module): def __init__(self, input_size, hidden_size, output_size, num_rules): super(Seq2SeqBiGRUDS, self).__init__() self.encoder_forward = BiGRUCell(input_size, hidden_size) self.encoder_backward = BiGRUCell(input_size, hidden_size) self.fuzzy_layer = FuzzyLayer(hidden_size * 2, num_rules) self.decoder_forward = BiGRUCell(hidden_size * 2, hidden_size) self.decoder_backward = BiGRUCell(hidden_size * 2, hidden_size) self.fc = nn.Linear(hidden_size * 2, output_size) def forward(self, S): T = S.size(1) h_forward = torch.zeros(S.size(0), self.encoder_forward.hidden_size).to(S.device) h_backward = torch.zeros(S.size(0), self.encoder_backward.hidden_size).to(S.device) H_forward = [] H_backward = [] for t in range(T): h_forward = self.encoder_forward(S[:, t], h_forward) H_forward.append(h_forward) for t in reversed(range(T)): h_backward = self.encoder_backward(S[:, t], h_backward) H_backward.append(h_backward) H = torch.stack([torch.cat([H_forward[t], H_backward[T-1-t]],dim=1) for t in range(T)], dim=1) H_prime = self.fuzzy_layer(H) C = torch.zeros_like(H_prime[:, 0]) decoder_outputs = [] for t in range(T): h_decoder_forward = self.decoder_forward(torch.cat([H_prime[:, t], C], dim=1), h_forward) h_decoder_backward = self.decoder_backward(torch.cat([H_prime[:, t], C], dim=1), h_backward) h_decoder = torch.cat([h_decoder_forward, h_decoder_backward], dim=1) decoder_outputs.append(self.fc(h_decoder)) return torch.stack(decoder_outputs, dim=1) The training module is as follows: def train_model(model, train_loader, criterion, optimizer, num_epochs): for epoch in range(num_epochs): model.train() total_loss = 0 for S, Y in train_loader: S, Y = S.to(device), Y.to(device) optimizer.zero_grad() outputs = model(S) loss = criterion(outputs, Y) loss.backward() optimizer.step() total_loss += loss.item() print(f'Epoch [{epoch+1} / {num_epochs}], Loss: {total_loss / len(train_loader):.4f}') # Parameter settings input_size = 10 hidden_size = 50 output_size = 1 num_rules = 5 num_epochs = 20 learning_rate = 0.001 # Initialize the model, loss function, and optimizer model = Seq2SeqBiGRUDS(input_size, hidden_size, output_size, num_rules).to(device) criterion = nn.MSELoss() optimizer = optim.Adam(model.parameters(), lr=learning_rate) # Train the model train_model(model, train_loader, criterion, optimizer, num_epochs) In the above program, the forward and backward Bi-GRU units are used to process the input sequence S respectively, generating the forward hidden state sequence H_forward and the backward hidden state sequence H_backward. The forward and backward hidden states are concatenated to form the final hidden state sequence H.

[0120] Fuzzy processing layer: Apply fuzzy rules to the hidden state sequence H and calculate the membership of the hidden state h(t) at each time step t to each fuzzy set. Use the membership to perform weighted averaging on the linear transformation results to generate a modified hidden state sequence H'.

[0121] Backtracking layer: Calculate the prediction error of the previous time step and use the least squares method to fit the residual to obtain the residual backtracking matrix W e . Using W e And the modified hidden state sequence H' generates the intermediate vector C.

[0122] Decoding layer: Use forward and backward Bi-GRU units to process the modified hidden state sequence H' and the intermediate vector C to generate the hidden state of the decoding layer. The hidden state of the decoding layer is converted into the predicted sequence S' of the dependent variable Y through the fully connected layer.

[0123] The training process is to use the mean square error (MSE) as the loss function and update the model parameters through the Adam optimizer. In each epoch, the training data is forward propagated, the loss is calculated, and then back-propagated and the parameters are updated.

[0124] Embodiment 3: In embodiment 1, when there is a longer input sequence in step S101, it may contain more historical information, which helps the model to more accurately capture the temporal dependency and long-term memory effect of the data; but it will also increase the complexity and computational complexity of the model. Therefore, it is necessary to make a trade-off between the accuracy and computational efficiency of the model. To this end, this embodiment provides an optimization solution for step S101, that is, introducing a dynamic adjustment mechanism to dynamically adjust the length of the input sequence S (i.e., the time step T): P1. Calculate the energy function e ij : ; Among them, q j k represents the query vector of the jth element (i.e., independent variable) in the input sequence S; i Represents the key vector of the i-th element in the input sequence S; d k is the key vector k i ; ⋅ represents the vector dot product operation.

[0125] The dot product attention is calculated by computing the query vector q j and key vector ki The larger the dot product result, the more similar the two vectors are.

[0126] P2. Calculate the weight factor p ij : ; Among them, the weight factor p ij represents the attention weight of the i-th element in the input sequence S to the j-th element in the output sequence; T is the total length of the input sequence. k is the index.

[0127] P3, according to the weight factor p ij The size of the input elements that have little impact on the output is selectively ignored, thereby shortening the actual processing length of the input sequence. If a threshold is set, all weight factors p less than the threshold are ignored. ij The element in the input sequence S corresponding to (i.e., the independent variable) will be removed.

[0128] It is understandable that by ignoring those input elements that have little impact on the output, the amount of data that the model needs to process can be reduced, thereby reducing computational complexity and improving computational efficiency. Dynamic adjustment ensures that the model can still capture key information when processing shortened input sequences, thereby maintaining the accuracy of the model. This is because the dynamic adjustment mechanism can highlight and help us further eliminate those input elements that have a greater impact on the output by dynamically adjusting the attention weights. Since the dynamic adjustment mechanism can dynamically focus on different parts of the input sequence, it can make the model more adaptable to input sequences of different lengths. This allows the model to maintain good performance when processing input data with different time steps T.

[0129] Embodiment 4: In step S201 of embodiment 1, A i is a fuzzy set; this embodiment further defines a geological condition fuzzy set for the specific working environment of TBM, including the following three types: (1) Soft rock fuzzy set A soft (h(x)): Soft rock fuzzy set is used to describe the situation where the geological conditions encountered during tunneling are relatively soft and the rock strength is low. It is defined as a Gaussian fuzzy set, where the mean μ soft Represents the typical soft rock strength value, standard deviation σ soft Reflects the fluctuation range of rock strength: ; Among them, the mean μ soft Based on historical excavation data or geological exploration results, a typical soft rock strength value is determined as the mean. This value represents the central trend of soft rock strength. Standard deviation σ softIt reflects the fluctuation range of soft rock strength, that is, the degree of dispersion of rock strength around the mean. The larger the standard deviation, the wider the fluctuation range of rock strength. The Gaussian function is used as the membership function of the fuzzy set because it is smooth, continuous and differentiable, and can well describe the fuzziness of rock strength within the soft rock range.

[0130] (2) Hard rock fuzzy set A hard (h(x)): The hard rock fuzzy set is used to describe the hard geological conditions encountered during tunneling, where the rock strength is high. It is also defined as a Gaussian fuzzy set, but the mean and standard deviation are adjusted accordingly to reflect the characteristics of hard rock: ; Among them, the mean μ hard Based on historical excavation data or geological exploration results, a typical hard rock strength value is determined as the mean. This value represents the central trend of hard rock strength and is usually larger than the mean of soft rock. Standard deviation σ hard Reflects the fluctuation range of hard rock strength. Since the strength of hard rock is usually relatively stable, its standard deviation may be relatively small. The Gaussian function is also used as the membership function of the fuzzy set to describe the fuzziness of rock strength within the hard rock range.

[0131] (3) Mixed rock fuzzy set A mixed (h(x)): used to describe the complex geological conditions and the mixture of soft and hard rocks during the excavation process. Since the geological conditions of mixed rocks are relatively complex, a single Gaussian fuzzy set may not be able to accurately describe its characteristics. Therefore, this embodiment considers combining multiple Gaussian fuzzy sets or using more complex fuzzy set shapes (such as trapezoidal fuzzy sets) to define: ; Among them, A i (x) represents the i-th Gaussian fuzzy set (which can be a soft rock fuzzy set or a hard rock fuzzy set), w i represents the weight of the i-th fuzzy set, and .

[0132] Or, using the trapezoidal fuzzy set definition: ; Among them, a, b, c, and d are the four parameters of the trapezoidal fuzzy set, representing the starting value of the mixed rock strength range, the transition value between soft rock and hard rock, the typical value of hard rock, and the end value of the mixed rock strength range. By combining multiple Gaussian fuzzy sets, the geological conditions of mixed rocks can be described more flexibly. Each Gaussian fuzzy set can represent rocks of different strength ranges in the mixed rock, and the weight w i This reflects the relative importance of these different intensity ranges in migmatites.

[0133] Trapezoidal fuzzy sets are a simpler fuzzy set shape that defines a trapezoidal region by using four parameters. Among mixed rock fuzzy sets, trapezoidal fuzzy sets can well describe the transition process of rock strength from soft rock to hard rock, as well as the starting and ending values ​​of the mixed rock strength range.

[0134] Through the above definitions, geological condition fuzzy sets can be used to more accurately describe and handle different geological conditions during excavation, thereby improving the intelligent control effect of cantilever tunnel excavation.

[0135] Embodiment 5: In step S203 of embodiment 1, each proposition Y is calculated i The basic probability distribution m(Yi) reflects the probability distribution of proposition Y i The method is: P1. Define the constraints of m(Yi): (1) The basic probability assignment of the empty set is 0: m(∅) = 0. This means that there is no trust in the empty set (i.e., impossible events).

[0136] (2) The sum of the basic probability distributions of all propositions is 1: ∑ m(Yi) = 1, where Y i Iterate over all propositions in the identification frame Θ. This means that the sum of the confidences of all possible answers is 1.

[0137] P2. Specific allocation: For each proposition Y i ,m(Y i ) is assigned a value based on the available evidence. The sensor data is analyzed and processed to assess the likelihood that the advancement rate falls within the interval represented by the proposition. This involves collecting sensor data from the TBM and preprocessing it to extract information related to the advancement rate. Based on the selected features, an assignment mechanism is established to calculate the probability of each proposition Y using statistical methods. i The basic probability distribution m(Y i ): (1) For each proposition Y i , calculate the frequency of the tunneling speed in the sensor data falling within this interval. This can be achieved by counting the number of data points that fall within this interval and dividing it by the total number of data points.

[0138] (2) Basic probability distribution: m(Yi) = frequency(Yi) Among them, frequency (Yi) represents the frequency of the excavation speed falling within the interval represented by proposition Yi.

[0139] (3) Normalization: Since frequency estimation may result in the sum of the basic probability distributions of all propositions not being 1, normalization is required. This is achieved by dividing the basic probability distribution of each proposition by the sum of the basic probability distributions of all propositions: m'(Y i ) = m(Y i ) / ∑ m(Yj) ; Among them, m'(Y i ) represents the normalized basic probability distribution, ∑m(Y j ) represents the sum of the basic probability distributions of all propositions.

[0140] For example, the following sensor data of the excavation speed are provided (unit: m / min): 12, 18, 25, 10, 20, 30, 15, 22, 10, 28 The excavation speed is divided into the following three intervals and the corresponding propositions are defined: Y1: [0, 15]; Y2: (15, 25]; Y3: (25, 35]; Now calculate the basic probability distribution of each proposition according to the above algorithm: The data points that fall within the Y1 interval are: 12, 10, 10, a total of 3, with a frequency of 3 / 10; The data points falling in the Y2 interval are: 18, 20, 15, 22, a total of 4, with a frequency of 4 / 10; The data points that fall within the Y3 interval are: 25, 30, 28, a total of 3, with a frequency of 3 / 10; Basic probability distribution: m(Y1) = 3 / 10; m(Y2) = 4 / 10; m(Y3) = 3 / 10; Normalization: m'(Y1) = (3 / 10) / (3 / 10 + 4 / 10 + 3 / 10) = 3 / 10; m'(Y2) = (4 / 10) / (3 / 10 + 4 / 10 + 3 / 10) = 4 / 10; m'(Y3) = (3 / 10) / (3 / 10 + 4 / 10 + 3 / 10) = 3 / 10.

[0141] Embodiment 6: This embodiment further discloses an intelligent control system for cantilever tunneling as the hardware carrier of all the above embodiments, such as Fig. 9As shown, the system includes a processor and a memory connected to the processor, wherein program instructions are stored in the memory. When the program instructions are executed by the processor, the processor executes the intelligent control method as described above; after the control electrical signal u(t) is obtained, it is input into the TBM control system, and the TBM control system controls the actuator of the TBM to perform corresponding operations.

[0142] It should be pointed out that the TBM control system drives the actuators of the tunnel boring machine (such as the propulsion cylinder, the cutterhead rotation motor, the steering system, etc.) to perform corresponding actions according to the control signals received, so as to achieve precise tunneling operations. The intelligent control system also includes a feedback mechanism, which continuously monitors the actual working status of the tunnel boring machine through sensors and feeds this information back to the processor to form a closed-loop control, further improving the control accuracy and stability. Modern intelligent control systems often also support remote monitoring functions, allowing operators to remotely monitor the tunneling process through the network in a safe area, adjust the control strategy in a timely manner, and even remotely intervene when necessary.

[0143] Test example: 1. Experimental purpose: The purpose of this experiment is to verify the superiority of the bidirectional Seq-Bi-GRU-DS fuzzy neural network in intelligent control of tunnel excavation by comparing the prediction accuracy and soil stress effect of the bidirectional Seq-Bi-GRU-DS fuzzy neural network (experimental group) and the traditional BPNN neural network (control group) when cooperating with the TBM control system for tunnel excavation.

[0144] (II) Experimental and control group settings: 2.1 Experimental Group Control method: the bidirectional Seq-Bi-GRU-DS fuzzy neural network described in Examples 1 to 5; Its characteristics include supporting unequal lengths of input and output sequences, being able to accurately describe the entire confidence interval distribution, and optimizing the evaluation of the PR value of the excavation speed.

[0145] 2.2 Control group: Control method: traditional BPNN neural network; The characteristic is that it does not support unequal lengths of input and output sequences, and its application effect is limited in complex and changeable construction environments.

[0146] 3. Experimental methods Simulation environment: Establish shield tunneling scenarios in ANSYS finite element simulation software such as Fig.10As shown in the figure, the shield (diameter 8600mm) passes through the full-section sandy gravel muddy stratum (soil pressure range: 180~200KP), the shield 440 ring cutter head torque is set to 2200~2800knm; the soil cabin pressure is 1.2~1.5bar, and the propulsion speed is 50~70mm / min. The shield thrust is 10000KN~30000KN.

[0147] Both the experimental group and the control group performed predictions according to the above parameters, output the predicted mileage, and compared it with the simulated mileage to determine the accuracy of the prediction.

[0148] After the comparison procedure of predicting the advancement mileage is completed, the control parameters of the experimental group and the control group are configured in the simulation scenario to control the TBM respectively. The stress effects of the soil layers of the simulated experimental group and the control group are analyzed, including the stress analysis of the upper and lower soil layers and the excavation stress analysis.

[0149] (IV) Experimental results: 4.1 Prediction Accuracy: like Fig.10 As shown in the figure, the prediction accuracy of the experimental group is significantly better than that of the control group. The experimental group can more accurately predict the advancement mileage of the shield machine, which is more consistent with the simulation results.

[0150] 4.2 Stress Effect: Stress analysis of upper and lower soil layers ( Fig.12 ): The stress generated by the experimental group (Part B) was significantly smaller than that of the control group (Part A), indicating that the experimental group performed better in controlling soil stress.

[0151] Tunneling stress analysis ( Fig.13 ): Similarly, the stress generated by the experimental group (Part B) is also significantly smaller than that of the control group (Part A), which further verifies the stress control effect of the experimental group during the excavation process.

[0152] (V) Results Analysis: The reason why the experimental group performed better than the control group in terms of prediction accuracy and stress control is mainly due to the following advantages of the bidirectional Seq-Bi-GRU-DS fuzzy neural network: The bidirectional Seq-Bi-GRU-DS fuzzy neural network can accurately describe the entire confidence interval distribution in advance, so that the model can more accurately evaluate the PR value of the excavation speed when performing statistics. Based on the accurately evaluated PR value, the model can indirectly optimize the stress of the soil layer and reduce the disturbance of the soil layer during the excavation process.

[0153] Different from the traditional BPNN neural network, the bidirectional Seq-Bi-GRU-DS fuzzy neural network supports unequal lengths of input and output sequences, which enables the model to more effectively handle differences in input data dimensions and quantity caused by changes in geological conditions, adjustments to tunnel boring machine operating strategies, etc.

[0154] All the above embodiments only express the implementation methods of the relevant practical applications of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be based on the attached claims.

[0155] For those skilled in the art, it can be further appreciated that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0156] At the same time, those skilled in the art can understand that all or part of the processes in all the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Claims

1. An intelligent control method for cantilever tunneling, characterized in that: Including at the current time step t, perform the following steps: S1, Data Engineering: Collect independent variables X(t) of TBM excavation parameters, including total thrust G t , Cutter power C p , Cutter torque C t , Cutter penetration C h , Cutter speed C s , field penetration FPI, torque penetration index TPI and current advancement speed PR'; let the future advancement speed PR be the dependent variable Y; the independent variable X constitutes the input sequence S; S2, bidirectional Seq-Bi-GRU fuzzy neural network operation: the encoding layer receives the input sequence S, calculates the candidate hidden states of the forward and backward directions through multiple groups of forward and backward GRU units, and forms a hidden state sequence H, which is input into the fuzzy processing layer for fuzzy operation processing, and generates a modified hidden state sequence H' to the decoding layer; the decoding layer calculates the candidate hidden states of the forward and backward directions through multiple groups of forward and backward GRU units based on the modified hidden state sequence H', and forms a prediction sequence S' of the dependent variable Y; the DS evidence theory algorithm is used to calculate the confidence of different dependent variables Y in the prediction sequence S', and the most credible final dependent variable Y' is extracted; S3, execute control: based on the difference and direction between the final dependent variable Y' and the current advance speed PR', the PID controller calculates the control electrical signal u(t) for controlling the advance speed of the TBM at the current time step t and controls the TBM.

2. The intelligent control method according to claim 1, characterized in that: In S1, the independent variable X(t)=[G t ,C p ,C t ,C h ,C s ,FPI,TPI,PR′]; The input sequence S=[X(1),X(2),…,X(t),…,X(T)]; Where T represents the total time steps of the input sequence.

3. The intelligent control method according to claim 1, characterized in that: In S2, the data-driven process of the bidirectional Seq-Bi-GRU fuzzy neural network includes: Encoding layer processing: For each time step t, the forward GRU unit receives the input sequence S and the hidden state of the previous time step , calculate the forward candidate hidden state of the current time step and the forward hidden state ; For each time step t, the backward GRU unit receives the input sequence S and the hidden state of the next time step , calculate the backward candidate hidden state of the current time step and the backward hidden state ; For each time step t, the forward hidden state and the backward hidden state Merge to form the final hidden state h(t); Fuzzy processing layer processing: at each time step t and each fuzzy rule R i , calculate h(t) belongs to the fuzzy set A i The membership degree α i (t); using membership degree α i (t) for the linear transformation F i (h(t)) is weighted averaged to obtain a modified hidden state and a modified hidden state sequence H' = [h'(1), h'(2), ..., h'(T)]; where T is the length of the input sequence S; Decoding layer processing: including the calculation of the forward GRU unit, the calculation of the backward GRU, and the decoding operation composed of the splicing operation, using the hidden state to represent h' t The prediction sequence S' of the dependent variable Y is generated by the intermediate vector C; the confidence of different dependent variables Y in the prediction sequence S' is calculated using the DS evidence theory algorithm, and the most credible final dependent variable Y' is extracted.

4. The intelligent control method according to claim 3, characterized in that: A traceback layer is also included between the encoding layer and the decoding layer, and the traceback layer is used to modify the hidden state sequence H' of the previous time step. t-1 , perform residual backtracking through the least squares method, use historical prediction errors for compensation, and generate an intermediate vector C; use the intermediate vector C as the input of the decoding layer; The decoding layer calculates the candidate hidden states of the forward and backward directions based on the intermediate vector C and the modified hidden state sequence H' through multiple groups of forward and backward GRU units to form a predicted sequence S' of a set of dependent variables Y.

5. The intelligent control method according to claim 3 or 4, characterized in that: In the encoding layer, the forward hidden state sequence : ; The backward hidden state : ; The hidden state sequence H: ; in, , [·;·] represents vector concatenation operation.

6. The intelligent control method according to claim 3 or 4, characterized in that: In the fuzzy processing layer: based on N fuzzy sets A i and the linear transformation function F i Fuzzy rules R i , for each time step t and each of the fuzzy rules R i , calculate the final hidden state h(t) belongs to the fuzzy set A i The membership degree α i (t): ; Then, using the membership degree α i (t) for the linear transformation F i (h(t)) is weighted averaged to obtain the corrected hidden state .

7. The intelligent control method according to claim 4, characterized in that: The data-driven process of the traceback layer includes: Assume E t−1 is the prediction error of the previous time step, that is, the actual value With the predicted value The difference between: ; Use the least squares method to fit the residuals and get the residual backtracking matrix W e , so that: ； Where W is the backtracking matrix, ∥⋅∥ 2 is the square of the 2-norm of the vector; Using the obtained residual backtracking matrix W e And the modified hidden state sequence H′ t−1 , generate the intermediate vector C: ； Combine the intermediate vector C with the encoding layer output H of the current time step t Combined as input to the decoding layer: H′′ t =f(H t ,C); Where f represents the concatenation operation function: H′′t=[H t ;C].

8. The intelligent control method according to claim 4, characterized in that: In the decoding layer, the data-driven process of the DS evidence theory algorithm includes for each dependent variable Y' i , calculate its confidence Bel(Y' i ): The identification framework Θ divides the tunneling speed into several intervals, each of which represents a proposition: Let Θ = {Y1, Y2, ..., Y n }, where Y i The proposition that the excavation speed falls within the i-th interval is the dependent variable Y' i ; Based on the information about the tunneling speed provided by the TBM sensor data, the basic probability distribution m(Y i ); let P(Y i ) represents the proposition Y i The set of all subsets of ′ ) is calculated as: ; Select the confidence level Bel(Yi ′ ) corresponds to the basic probability distribution m(Y i ), find the corresponding proposition Y i ; Then extract the most credible final dependent variable Y', that is, the one with the highest confidence: .

9. The intelligent control method according to claim 3 or 4, characterized in that: In S3, the data driving method of the PID controller is: Error e(t) = Y' set -PR'; Integral term: I(t) = I(t-1) + e(t)·△t; Differential term: ; Proportional term: P(t) = K p ·e(t); Control electrical signal u(t) = P(t) +K i I(t) + K d ·D(t); Among them, K p is the proportional gain; K i is the integral gain; K d is the differential gain; t is the time gain; The set value Y' is the target excavation speed; I(t) is the integral term value at the current moment; I(t-1) is the integral term value at the previous sampling moment; e(t-1) is the error at the previous sampling moment; P(t) and D(t) are the differential term value and proportional term value at the current moment.

10. An intelligent control system for cantilever tunneling, characterized in that: The system comprises a processor and a memory connected to the processor, wherein the memory stores program instructions, and when the program instructions are executed by the processor, the processor executes the intelligent control method according to any one of claims 1 to 9; After the control electrical signal u(t) is obtained, it is input into the TBM control system, and the TBM control system controls the actuator of the TBM to perform corresponding operations.

Images