Shield machine posture prediction method based on multi-attention mechanism optimization combination model

Abstract

The present application provides a shield posture prediction method based on a multi-attention mechanism optimized combination model, comprising the following steps: obtaining historical feature data and shield posture data of a target shield machine, performing missing value filling, excavation section data extraction, data normalization, and then performing grey correlation analysis on the correlation between the two; extracting input feature attention by selecting a kernel network; establishing a bidirectional long-short-term memory network to predict implicit outputs containing long-term trends; establishing a gated recurrent unit network to predict implicit outputs containing short-term fluctuation trends; splicing two sets of prediction network implicit outputs, and using a multi-head attention mechanism to extract output attention, capturing the implicit features of the combined model output and the dependency on the shield posture, and realizing the mapping of the implicit output to the final predicted shield posture. The present invention improves the prediction accuracy and prediction efficiency of the shield posture prediction model, contributes to the refined control of the shield posture during the construction phase, and improves the shield construction efficiency and construction safety.

Description

Technical Field

[0001] The present invention relates to the technical field of shield construction, and in particular to a shield posture prediction method based on a multi-attention mechanism optimization combination model. Background Art

[0002] Shield tunneling, known for its safety, efficiency, economy, and minimal environmental impact, is widely used in tunnel construction for subways, municipal projects, railways, highways, water conservancy projects, mining projects, and military projects. With the rapid development of urban construction and shield tunneling technology, shield tunneling in my country is gradually moving towards greater depths and larger cross-sections.

[0003] However, as tunnel cross-sections increase, the complexity of geological conditions introduces uncertainty and risk to tunnel excavation. Coupled with factors such as shallow burial depths and high water pressure, shield machine posture control during construction requires even more sophisticated control. Furthermore, due to the increased inertia of the shield system, changes in shield posture become more complex. Traditional methods of manually observing the shield machine's real-time posture feedback to adjust its excavation parameters often result in ineffective posture adjustment, such as delayed posture correction and serpentine trajectories. Therefore, in-depth research on shield machine posture prediction methods to assist in refined shield construction posture control is of great engineering significance for safe and efficient shield construction. Summary of the Invention

[0004] The purpose of the present invention is to provide a shield attitude prediction method based on a multi-attention mechanism optimization combination model to improve the prediction accuracy of the shield attitude and realize scientific adjustment and control optimization of the shield tunneling attitude.

[0005] In order to solve the above technical problems, the present invention provides a shield posture prediction method based on a multi-attention mechanism optimization combination model, comprising the following steps:

[0006] S1. Obtain and preprocess the historical feature data of the tunneling section and the shield posture data of the target shield machine, then use the preprocessed historical feature data as a comparison feature column and the preprocessed shield posture data as a reference column to obtain the correlation between each historical feature data and the shield posture data;

[0007] S2. Arrange the correlation degrees of each historical feature data in descending order, set the feature selection dimension, and select the feature columns with the highest correlation degrees according to the set feature selection dimension as the feature data after feature selection;

[0008] S3. Process the feature data after feature selection to generate a feature time series dataset, which is used as input data for the deep learning model. The feature time series data is divided into a training set and a test set according to a certain ratio.

[0009] S4, by selecting the kernel unit network, perform feature attention calculation on the input time series data, and adaptively form an implicit feature map with attention;

[0010] S5. Construct a bidirectional long-short-term memory network to predict the long-term trend implicit output of the shield posture: the implicit feature map with attention is used as the input data of the bidirectional long-short-term memory network through a fully connected layer. The network parameters are iteratively optimized according to the loss function model of the bidirectional long-short-term memory network to obtain the optimal network parameters;

[0011] S6. Construct a gated recurrent unit network to predict the implicit output of the short-term fluctuation trend of the shield posture: the implicit feature map with attention in step S4 is mapped through the fully connected layer as the input data of the gated recurrent network, and the network parameters are iteratively optimized according to the network loss function model to obtain the optimal network parameters;

[0012] S7. Combine the long-term trend implicit output and the short-term fluctuation trend implicit output, and adaptively capture the dependency between the long-term trend implicit output and the short-term fluctuation trend implicit output and the model output on the shield posture through the multi-head attention mechanism, and finally obtain the shield posture prediction result.

[0013] In the preferred solution, the preprocessing in step S1 includes missing value filling, excavation section data extraction and data normalization;

[0014] The correlation between each historical feature data and shield posture data is obtained through grey correlation analysis;

[0015] The feature data after feature selection in step S3 is subjected to time window processing according to the selected time series length.

[0016] In a preferred solution, the historical feature data in step S1 is:

[0017] X i (k),i∈[1,n],k∈[1,m],

[0018] where X i (k) is the kth point data in the historical feature data after the i-th preprocessing, n and m are the type and time point of the collected historical feature data respectively;

[0019] The posture data after preprocessing in step S1 is:

[0020] Y o (k),o∈[1,4],k∈[1,m],

[0021] Among them, Y o(k) represents the posture data at time point k after preprocessing, where o represents different types of posture data: head vertical posture, head horizontal posture, tail vertical posture and head horizontal posture, and m is the number of time points of historical feature data collected;

[0022] The data of the kth time point of historical feature data is:

[0023] {X1(k),X1(k),…,X n (k)}

[0024] In step 1, the grey correlation analysis method is used to obtain the correlation between each historical feature data and the shield posture data. Specifically:

[0025] Calculate the grey correlation coefficient:

[0026]

[0027] i∈[1,n],k∈[1,m],o∈[1,4];

[0028] Among them, Y o (k) represents the shield posture data of type o after preprocessing at the kth time point, X i (k) represents the historical feature data of the i-th category after preprocessing at the k-th time point, n is the type of historical feature data collected, m is the number of time points for collecting data, ρ is the resolution coefficient, and its value range is (0,∞), ξ o,i (k) represents the grey correlation coefficient between the historical feature data after preprocessing of type i and the shield posture data after preprocessing of type o at the kth time point.

[0029] Calculate the correlation:

[0030]

[0031] where γ i is the correlation between the historical characteristic data of the i-th shield machine and the shield machine posture data, γ i The closer it is to 1, the higher the correlation between this historical feature data and the shield posture data.

[0032] In the preferred solution, in step S3, the historical feature data after time window processing is as follows:

[0033] x(k)={[X1(k-T+1),...,X c (k-T+1)];...;[X1(k),...,X c (k)]};

[0034] Where x(k) is the network model input time window data at time point k, T is the time window time point length, and c is the number of feature data types selected after grey relational ranking.

[0035] In the preferred solution, the attention mechanism calculation for the time series based on the selected kernel network in step S4 is performed as follows:

[0036] S41, splitting process: take the time window data, i.e., multi-dimensional time series features, as input, and pass them through two sets of one-dimensional convolutions of different sizes to obtain two sets of features, U1 and U2;

[0037] S42, fusion process: The two sets of features are added together to obtain the fusion feature U, and then passed through the one-dimensional global average pooling layer F gp Reduce the fusion feature dimension to a global information vector s of length L:

[0038]

[0039] Where s l is the lth element of the global information vector, U l (k) is the k-th time point element of the l-th feature of the fused feature, and H is the time series length of the fused feature;

[0040] Then pass through the fully connected layer Reduce the dimension of the information vector and generate an attention information vector z with a sequence length of L / r, where r is a controllable parameter;

[0041] S43, selection operation: To strengthen important input features and suppress redundant features, softmax regression is used to obtain feature attention weights:

[0042]

[0043] Among them, A L , B L is the parameter matrix of z, exp is the exponential function with the natural constant e as the base, M A 、M B They are the feature attention weights of U1 and U2 respectively. The feature map V extracted by attention can be obtained through attention calculation. L :

[0044] V L =M A U1+M B U2, M A +M B =1.

[0045] In the preferred solution, the bidirectional long short-term memory network constructed in step S5 includes a fully connected layer and three BiLSTM layers, and the features with attention are mapped to the bidirectional long short-term memory network input through the fully connected layer; each BiLSTM layer processes the input feature data from the positive and negative directions respectively through two independent LSTM layers.

[0046] In the preferred solution, the hidden state and cell state of the forward LSTM at each time step are calculated by the following formula:

[0047] I t =σ(W Ix x t +W Ih h t-1 +b I );

[0048] F t =σ(W Fx x t +W Fh h t-1 +b F );

[0049] O t =σ(W Ox x t +W Oh h t-1 +b O );

[0050]

[0051] h t =O t ⊙tanh(C t );

[0052] Among them, I t 、F t , O t and are the input gate, forget gate, output gate and candidate state representing the current information, W and b are the network weight matrix and bias vector, σ and tanh are the Sigmoid and hyperbolic tangent functions, and ⊙ represents the element-by-element product.

[0053] The reverse LSTM processes the sequence in the opposite direction, and the hidden state h′ at each time step t and cell state C′ t Also calculated by the above formula, the final bidirectional long short-term memory network output is composed of the concatenation of the two directional hidden states: y′ t =[h t ; h′ t ].

[0054] In the preferred solution, the gated recurrent unit network constructed in step S6 includes a fully connected layer and three GRU layers; the features with attention are mapped to the gated recurrent unit network input through the fully connected layer; each GRU unit layer has a corresponding update gate and reset gate, which respectively determine the retained information and discarded information from the previous moment to the current moment. Specifically:

[0055] z t =σ(W zx x t +W zh h t-1 +b z );

[0056] r t =σ(W rx x t +W rh h t-1 +b r );

[0057]

[0058] Among them, z t is the update gate, r t is the reset gate, h t is the hidden state passed to the next moment, is the candidate hidden state, x t represents the network input mapped by the fully connected layer at the current moment, W and b are the network weight matrix and bias vector respectively, σ is the Sigmoid activation function, ⊙ represents the element-by-element product, and the final output is: y″ t =[h t ].

[0059] In a preferred embodiment, the combined long-term trend implicit output and short-term fluctuation trend implicit output in step S7 are formed by concatenating the bidirectional long short-term memory network output in step S5 and the gated recurrent unit network output in step S6: t =[y′ t ;y″ t ].

[0060] In the preferred solution, in step S7, the dependency between the two sets of implicit outputs and the model output shield posture is adaptively captured through a multi-head attention mechanism, and the specific operations are as follows:

[0061] The multi-head attention mechanism uses different linear mappings to map the input matrix of dimension m×d into N groups. Each group of independent self-attention mechanisms contains 3 matrices: value V, key K and query Q, where Q and K have the same length d K , V length is d V :

[0062]

[0063] Where x i is the i-th implicit feature output by the BiLSTM and GRU networks, W i V 、W i K 、W i Q are three independent mapping weight matrices that can be iterated during network training. i represents the output matrix of the i-th group of self-attention operations;

[0064] Finally, the output matrix of the multi-head attention mechanism MA is obtained by concatenating the output matrices of the N groups of self-attention operations and linearly mapping them:

[0065] MA=Multihead(Q,K,V)=W O Concat(A1,...,A N );

[0066] Where W O It is the linear mapping weight matrix from the splicing layer to the multi-head attention input matrix. The feature matrix MA output by the multi-head attention mechanism is obtained by weighted connection and remapping of the implicit features of the original two models.

[0067] The beneficial effects of the present invention are as follows: the present invention innovatively uses a combination of multiple attention mechanisms and prediction models to predict the technical solution, which improves the prediction accuracy of the shield posture to a certain extent, reduces the difficulty of model training, contributes to the refined control of the posture, and improves the shield construction efficiency and construction safety. Grey correlation analysis GRA is used to perform preliminary feature type screening. In view of the high dimension of the characteristic data of the shield process, the features that mainly affect the shield posture are extracted as the input of the model, which effectively reduces the complexity of the model and improves the prediction efficiency; the selective kernel network SKNet is used to extract the input feature attention, which can effectively suppress the expression of redundant feature information and solve the problem that it is difficult to extract effective features from the long sequence feature input of the shield; the bidirectional long short-term memory network BiLSTM is combined with the bidirectional long short-term memory network GRU model for shield posture prediction, which can combine the different representations of the two groups of models and enhance the expression and generalization capabilities of the model; finally, the multi-head attention mechanism MA is used as the mapping network from the output of the combined prediction model to the prediction result, which can help the network automatically learn the importance and relevance of the output mapping and improve the model performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 Flow chart of the method of the present invention.

[0069] Figure 2 The kernel network is selected for extracting input feature attention structure diagram for the present invention.

[0070] Figure 3 This is a model ablation comparison diagram of the regression prediction data and the original data obtained by the model of the present invention to predict the shield posture.

[0071] Figure 4 This is a comparison chart of the shield posture predictions made by the model of the present invention and other models in the same data interval. DETAILED DESCRIPTION

[0072] To verify the effectiveness of the proposed model and method, we applied it to shield machine posture prediction using data from a Guangzhou subway tunnel. To demonstrate the model's effectiveness and generalization capabilities, we conducted simulations using data from two time periods: from 3:00 AM on January 4, 2021, to 4:00 PM on February 4, 2021, and from 10:00 PM on March 29, 2021, to 6:00 AM on April 20, 2021. The data was divided into training and test sets in a 7:3 ratio.

[0073] like Figure 1 Shown is a flow chart of the method of the present invention.

[0074] The following combination Figures 1 to 4 , introduces the specific implementation of the present invention, a shield posture prediction method based on a multi-attention mechanism optimization combination model, specifically comprising the following steps:

[0075] Step 1: Obtain the historical feature data and shield posture data of the target shield machine. Then, the historical feature data and shield posture data are processed sequentially through missing value filling, tunneling segment data extraction, and data normalization methods to obtain pre-processed tunneling segment historical feature data and pre-processed shield posture data. Subsequently, the pre-processed historical feature data is used as a comparison feature column, and the pre-processed shield posture data is used as a reference column. Grey correlation analysis is performed to obtain the correlation between each historical feature data and the shield posture data.

[0076] The historical feature data in step 1 is:

[0077] X i (k),i∈[1,n],k∈[1,m],

[0078] where X i (k) is the kth point data in the historical feature data after the i-th preprocessing, n and m are the type and time point of the collected historical feature data respectively;

[0079] The posture data after preprocessing in step 1 is:

[0080] Y o(k),o∈[1,4],k∈[1,m],

[0081] Among them, Y o (k) represents the posture data at time point k after preprocessing, where o represents different types of posture data: head vertical posture, head horizontal posture, tail vertical posture, head horizontal posture, and m is the number of time points of historical feature data collected;

[0082] The data of the k-th time point of the historical feature data is:

[0083] {X1(k),X1(k),...,X n (k)};

[0084] In step 1, the grey correlation analysis method is used to obtain the correlation between each historical feature data and the shield posture data. Specifically:

[0085] Calculate the grey correlation coefficient:

[0086]

[0087] i∈[1,n],k∈[1,m],o∈[1,4],

[0088] Among them, Y o (k) represents the shield posture data of type o after preprocessing at the kth time point. i (k) represents the historical feature data of type i after preprocessing at the kth time point. n is the type of historical feature data collected, and m is the number of time points at which the data was collected. ρ is the resolution coefficient, and its value range is (0,∞). ξ o,i (k) represents the grey correlation coefficient between the historical feature data after preprocessing of type i and the shield posture data after preprocessing of type o at the kth time point.

[0089] Calculate the correlation:

[0090]

[0091] where γ i is the correlation between the historical characteristic data of the i-th shield machine and the shield machine posture data, γ i The closer it is to 1, the higher the correlation between this historical feature data and the shield posture data.

[0092] Step 2: Arrange the correlation of each historical feature data in descending order, set the feature selection dimension, and select the feature column with the highest correlation according to the set feature selection dimension as the feature data after feature selection;

[0093] Step 3: After feature selection, the feature data is time-windowed according to the selected time series length to generate a feature time series dataset, which serves as the input data for the deep learning model. The feature time series data is split into a training set and a test set according to a certain ratio.

[0094] In step 3, the historical feature data after time window processing is as follows:

[0095] x(k)={[X1(k-T+1),...,X c (k-T+1)];...;[X1(k),...,X c (k)]}'

[0096] Where x(k) is the network model input time window data at time point k, T is the time window time point length, and c is the number of feature data types selected after grey relational ranking;

[0097] Step 4: By selecting the kernel unit network, feature attention calculation is performed on the input time series data to adaptively form an implicit feature map with attention;

[0098] Figure 2 The kernel network is selected for extracting input feature attention structure diagram for this invention, and the overall process is reflected in combination with the following case operation process.

[0099] The attention mechanism calculation based on the selected kernel network for the time series described in step 4 is as follows:

[0100] Step 4.1, splitting process: take the time window data, i.e., the multi-dimensional time series features, as input and perform two sets of one-dimensional convolutions of different sizes. Here, the convolution kernel sizes are selected as 3×1 and 5×1 to obtain two sets of features, U1 and U2.

[0101] Step 4.2, fusion process: The two sets of features are added together to obtain the fusion feature U. Then the fusion feature U is obtained through the one-dimensional global average pooling layer. The fusion features are reduced to a global information vector s of length L.

[0102]

[0103] Where s l is the lth element of the global information vector, U l (k) is the k-th time point element of the l-th feature of the fused feature, and H is the time series length of the fused feature.

[0104] Then pass through the fully connected layer Reduce the dimension of the information vector and generate an attention information vector z with a sequence length of L / r, where r is a controllable parameter.

[0105] Step 4.3, select the operation: To strengthen the important input features and suppress the redundant features, use softmax regression to obtain the feature attention weight:

[0106]

[0107]

[0108] Among them, A L , B L is the parameter matrix of z, exp is the exponential function with the natural constant e as the base, M A 、M B They are the feature attention weights of U1 and U2 respectively. The feature map V extracted by attention can be obtained through attention calculation. L .

[0109] V L =M A U1+M B U2, M A +M B =1;

[0110] Step 5: Construct a bidirectional long-term short-term memory network. Based on the bidirectional long-term short-term memory network's strong memory characteristics for long-term trends, the long-term trend implicit output of the shield posture is predicted. The implicit feature map with attention is used as the input data of the bidirectional long-term short-term memory network through a fully connected layer. The network parameters are iteratively optimized according to the loss function model of the bidirectional long-term short-term memory network to obtain the optimal network parameters.

[0111] The bidirectional LSTM network constructed in step 5 consists of a fully connected layer and three BiLSTM layers. The fully connected layer maps the attention-bearing features into the bidirectional LSTM network input. Each BiLSTM layer processes the input feature data from two independent LSTM layers, one in the forward and one in the reverse direction.

[0112] Specifically, the hidden state and cell state of each time step of the forward LSTM are calculated by the following formula:

[0113] I t =σ(W Ix x t +W Ih h t-1 +b I )

[0114] F t =σ(W Fx x t +W Fh h t-1 +b F )

[0115] O t =σ(W Ox x t +W Oh h t-1 +b O )

[0116]

[0117]

[0118] h t =O t ⊙tanh(C t )

[0119] Among them, I t 、F t , O t and They are the input gate, forget gate, output gate and candidate state representing the current information, W and b are the network weight matrix and bias vector respectively, σ and tanh are Sigmoid and hyperbolic tangent functions respectively, and ⊙ represents the element-by-element product.

[0120] The reverse LSTM processes the sequence in the opposite direction, and the hidden state h′ at each time step t and cell state C′ t It can be calculated by a similar formula. The final output of the bidirectional long short-term memory network is the concatenation of the two directional hidden states: y′ t =[h t ; h′ t ].

[0121] Step 6: Construct a gated recurrent unit network. Based on the fact that the gated recurrent unit network is sensitive to short-term fluctuations, the short-term fluctuation trend of the shield posture is predicted. The implicit feature map with attention in step 4 is mapped through the fully connected layer as the input data of the gated recurrent network. The network parameters are iteratively optimized according to the network loss function model to obtain the optimal network parameters.

[0122] The gated recurrent unit network constructed in step 6 consists of a fully connected layer and three GRU layers. The fully connected layer maps the features with attention to the input of the gated recurrent unit network. Each GRU unit layer has a corresponding update gate and reset gate, which respectively determine the information retained and discarded from the previous moment to the current moment. Specifically:

[0123] z t =σ(W zx x t +W zh h t-1 +b z )

[0124] r t =σ(W rx x t +W rh h t-1 +b r )

[0125]

[0126] Among them, z t is the update gate, r t is the reset gate, h t is the hidden state passed to the next moment, is the candidate hidden state, x t represents the network input mapped by the fully connected layer at the current moment, W and b are the network weight matrix and bias vector respectively, σ is the Sigmoid activation function, and ⊙ represents the element-by-element product. The final output is: y″ t =[h t ];

[0127] In step 7, the long-term trend implicit output and the short-term fluctuation trend implicit output are combined, and the dependency between the two sets of implicit outputs and the model output on the shield posture is adaptively captured through the multi-head attention mechanism, and finally the shield posture prediction result is obtained.

[0128] The combined long-term trend implicit output and short-term fluctuation trend implicit output in step 7 are obtained by concatenating the output of the bidirectional long short-term memory network in step 5 and the output of the gated recurrent unit network in step 6: t =[y′ t ;y″ t ].

[0129] As described in step 7, the multi-head attention mechanism is used to adaptively capture the dependency between the two sets of implicit outputs and the model output shield posture. The specific operations are as follows:

[0130] The multi-head attention mechanism uses different linear mappings to map the input matrix of dimension m×d into N groups. Each group of independent self-attention mechanisms contains 3 matrices: value V, key K, query Q, where Q and K have the same length d K , V length is d V .

[0131]

[0132]

[0133] Where x i is the i-th implicit feature output by the BiLSTM and GRU networks. iV 、W i K 、W i Q are three independent mapping weight matrices that can be iterated during network training. i represents the output matrix of the i-th group of self-attention operations.

[0134] Finally, the output matrix of the multi-head attention mechanism MA is obtained by concatenating the output matrices of the N groups of self-attention operations and linearly mapping them:

[0135] MA=Multihead(Q,K,V)=W O Concat(A1,...,A N )

[0136] Where W O is the linear mapping weight matrix from the concatenation layer to the multi-head attention input matrix. The feature matrix MA output by the multi-head attention mechanism is obtained by weighted connection and remapping of the implicit features of the original two models, which is the final shield posture prediction result.

[0137] Figure 3 A comparison chart of the regression prediction data and the original data obtained by applying the shield posture prediction based on the multi-attention mechanism optimization combination model provided in the embodiment of the present invention. The thicker solid line in the figure is the true posture value, the thinner solid line is the prediction curve of this method, and the remaining curves are the prediction curves of the ablation model. The ablation comparison model includes: Bidirectional Long Short-Term Memory Network (BiLSTM), Gated Recurrent Unit (GRU), Bidirectional Long Short-Term Memory Network-Gated Recurrent Unit (BiLSTM-GRU), Selective Kernel Network-Bidirectional Long Short-Term Memory Network-Gated Recurrent Unit (SKNet-BiLSTM-GRU), Selective Kernel Network-Bidirectional Long Short-Term Memory Network-Gated Recurrent Unit-Multi-Head Attention Mechanism (SKNet-BiLSTM-GRU-MA).

[0138] Table 1 Ablation model evaluation indicators

[0139]

[0140] As can be seen from Table 1, the model and method provided by the present invention have high prediction accuracy, with a root mean square error RMSE of 0.89, a mean absolute error MAE of 0.69, and R 2 It also exceeds 0.98, indicating that the model structure and method design of the present invention are reasonable and effective.

[0141] Figure 4A comparison diagram of the regression prediction data and the original data obtained by applying the shield posture prediction based on the multi-attention mechanism optimization combination model provided in the embodiment of the present invention. The thicker solid line in the figure is the true posture value, the fine dotted line is the prediction curve of this method, and the remaining curves are the prediction curves of the comparison model. The comparison model includes: BP neural network-support vector machine (BP-SVM), multi-layer long short-term memory network (MLSTM), gated recurrent unit-random forest (GRU-RF), selected kernel network-bidirectional long short-term memory network-gated recurrent unit-multi-head attention mechanism (SKNet-BiLSTM-GRU-MA).

[0142] Table 2 Comparison of model evaluation indicators

[0143]

[0144] As can be seen from Table 2, the prediction accuracy of the model and method provided by the present invention is still relatively high in this data segment, with a root mean square error (RMSE) of 2.06, a mean absolute error (MAE) of 2.28, and an R 2 It also exceeds 0.98, indicating that the present invention is superior in model structure and method design.

[0145] In summary, the present invention proposes a shield posture prediction method based on the SKNet-BiLSTM-GRU-MA multi-attention mechanism optimization combination model, which realizes the accurate prediction of the shield posture, contributes to the refined control of the posture during shield tunnel construction, and improves the safety and efficiency of the construction process.

[0146] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A shield machine posture prediction method based on a multi-attention mechanism optimization combination model, characterized by: The following steps are involved: S1. Obtain and preprocess the historical feature data of the tunneling section and the shield posture data of the target shield machine, then use the preprocessed historical feature data as a comparison feature column and the preprocessed shield posture data as a reference column to obtain the correlation between each historical feature data and the shield posture data; S2. Arrange the correlation degrees of each historical feature data in descending order, set the feature selection dimension, and select the feature columns with the highest correlation degrees according to the set feature selection dimension as the feature data after feature selection; S3. Process the feature data after feature selection to generate a feature time series dataset, which is used as input data for the deep learning model. The feature time series data is divided into a training set and a test set according to a certain ratio. S4, by selecting the kernel unit network, perform feature attention calculation on the input time series data, and adaptively form an implicit feature map with attention; S5. Construct a bidirectional long-short-term memory network to predict the long-term trend implicit output of the shield posture: the implicit feature map with attention is used as the input data of the bidirectional long-short-term memory network through a fully connected layer. The network parameters are iteratively optimized according to the loss function model of the bidirectional long-short-term memory network to obtain the optimal network parameters; S6. Construct a gated recurrent unit network to predict the implicit output of the short-term fluctuation trend of the shield posture: the implicit feature map with attention in step S4 is mapped through the fully connected layer as the input data of the gated recurrent network, and the network parameters are iteratively optimized according to the network loss function model to obtain the optimal network parameters; S7. Combine the long-term trend implicit output and the short-term fluctuation trend implicit output, and adaptively capture the dependency between the long-term trend implicit output and the short-term fluctuation trend implicit output and the model output on the shield posture through the multi-head attention mechanism, and finally obtain the shield posture prediction result.

2. The shield machine posture prediction method based on the multi-attention mechanism optimization combination model according to claim 1 is characterized by: The preprocessing in step S1 includes missing value filling, tunneling section data extraction and data normalization; The correlation between each historical feature data and shield posture data is obtained through grey correlation analysis; The feature data after feature selection in step S3 is subjected to time window processing according to the selected time series length.

3. The shield posture prediction method based on the multi-attention mechanism optimization combination model according to claim 1 is characterized by: step The historical feature data in S1 is: X i (k),i∈[1,n],k∈[1,m], where X i (k) is the kth point data in the historical feature data after the i-th preprocessing, n and m are the type and time point of the collected historical feature data respectively; The posture data after preprocessing in step S1 is: Y o (k),o∈[1,4],k∈[1,m], Among them, Y o (k) represents the posture data at time point k after preprocessing, where o represents different types of posture data: head vertical posture, head horizontal posture, tail vertical posture and head horizontal posture, and m is the number of time points of historical feature data collected; The data of the kth time point of historical feature data is: {X1(k),X1(k),...,X n (k)} In step S1, the grey correlation analysis method is used to obtain the correlation between each historical feature data and shield posture data. Specifically: Calculate the grey relational coefficient: i∈[1,n],k∈[1,m],o∈[1,4]; Among them, Y o (k) represents the shield posture data of type o after preprocessing at the kth time point, X i (k) represents the historical feature data of the i-th category after preprocessing at the k-th time point, n is the type of historical feature data collected, m is the number of time points for collecting data, ρ is the resolution coefficient, and its value range is (0,∞), ξ o,i (k) represents the grey correlation coefficient between the historical feature data after preprocessing of type i and the shield posture data after preprocessing of type o at the kth time point; Calculate the correlation: where γ i is the correlation between the historical characteristic data of the i-th shield machine and the shield machine posture data, γ i The closer it is to 1, the higher the correlation between this historical feature data and the shield posture data.

4. The shield posture prediction method based on the multi-attention mechanism optimization combination model according to claim 1 is characterized by: In step S3, the historical feature data after time window processing is: x(k)={[X1(k-T+1),…,X c (k-T+1)];...;[X1(k),...,X c (k)]}; Where x(k) is the network model input time window data at time point k, T is the time window time point length, and c is the number of feature data types selected after grey relational ranking.

5. The shield machine posture prediction method based on the multi-attention mechanism optimization combination model according to claim 1 is characterized by: In step S4, the attention mechanism calculation of the time series is performed based on the selected kernel network. The specific process is as follows: S41, splitting process: take the time window data, i.e., multi-dimensional time series features, as input, and pass them through two sets of one-dimensional convolutions of different sizes to obtain two sets of features, U1 and U2; S42, fusion process: The two sets of features are added together to obtain the fusion feature U, and then passed through the one-dimensional global average pooling layer F gp Reduce the fusion feature dimension to a global information vector s of length L: Where s l is the lth element of the global information vector, U l (k) is the k-th time point element of the l-th feature of the fused feature, and H is the time series length of the fused feature; Then pass through the fully connected layer F fc Reduce the dimension of the information vector and generate an attention information vector z with a sequence length of L / r, where r is a controllable parameter; S43, selection operation: To strengthen important input features and suppress redundant features, softmax regression is used to obtain feature attention weights: Among them, A L , B L is the parameter matrix of z, exp is the exponential function with the natural constant e as the base, M A 、M B They are the feature attention weights of U1 and U2 respectively, and the feature map V extracted by attention is obtained through attention calculation L : V L =M A U1+M B U2,M A +M B =1。 6. The shield machine posture prediction method based on the multi-attention mechanism optimization combination model according to claim 1 is characterized by: The bidirectional long short-term memory network constructed in step S5 includes a fully connected layer and three BiLSTM layers. The fully connected layer maps the features with attention to the bidirectional long short-term memory network input; each BiLSTM layer processes the input feature data from the positive and negative directions through two independent LSTM layers.

7. The shield machine posture prediction method based on the multi-attention mechanism optimization combination model according to claim 6 is characterized by: Specifically, the hidden state and cell state of each time step of the forward LSTM are calculated by the following formula: I t =σ(W Ix x t +W Ih h t-1 +b I ); F t =σ(W Fx x t +W Fh h t-1 +b F ); O t =σ(W Ox x t +W Oh h t-1 +b O ); h t =O t ⊙tanh(C t ); Among them, I t 、F t , O t and are the input gate, forget gate, output gate and candidate state representing the current information, W and b are the network weight matrix and bias vector, σ and tanh are the Sigmoid and hyperbolic tangent functions, and ⊙ represents the element-by-element product. The reverse LSTM processes the sequence in the opposite direction, and the hidden state h′ at each time step t and cell state C′ t The calculation formula is the same as the forward LSTM. The final bidirectional long short-term memory network output is composed of the concatenation of the hidden states in two directions: y′ t =[h t ; h′ t ].

8. The shield machine posture prediction method based on the multi-attention mechanism optimization combination model according to claim 1 is characterized by: The gated recurrent unit network constructed in step S6 includes a fully connected layer and three GRU layers. The features with attention are mapped to the input of the gated recurrent unit network through the fully connected layer. There are corresponding update gates and reset gates on each GRU unit layer, which respectively determine the information to be retained and discarded from the previous moment to the current moment. Specifically: z t =σ(W zx x t +W zh h t-1 +b z ); r t =σ(W rx x t +W rh h t-1 +b r ); Among them, z t is the update gate, r t is the reset gate, h t is the hidden state passed to the next moment, is the candidate hidden state, x t represents the network input mapped by the fully connected layer at the current moment, W and b are the network weight matrix and bias vector respectively, σ is the Sigmoid activation function, ⊙ represents the element-by-element product, and the final output is: y″ t =[h t ].

9. The shield machine posture prediction method based on the multi-attention mechanism optimization combination model according to claim 1 is characterized by: The combined long-term trend implicit output and short-term fluctuation trend implicit output in step S7 are formed by concatenating the bidirectional long short-term memory network output in step S5 and the gated recurrent unit network output in step S6: and t =[y′ t ;and" t ]。 10. The shield machine posture prediction method based on the multi-attention mechanism optimization combination model according to claim 1 is characterized by: In step S7, the multi-head attention mechanism is used to adaptively capture the dependency between the two sets of implicit outputs and the model output shield posture. The specific operations are as follows: The multi-head attention mechanism uses different linear mappings to map the input matrix of dimension m×d into N groups. Each group of independent self-attention mechanisms contains 3 matrices: value V, key K and query Q, where Q and K have the same length d K , V length is d V : Where x i is the i-th implicit feature output by the BiLSTM and GRU networks, W i V 、W i K 、W i Q are three independent mapping weight matrices that are iterated during network training. i represents the output matrix of the i-th group of self-attention operations; Finally, the output matrix of the multi-head attention mechanism MA is obtained by concatenating the output matrices of the N groups of self-attention operations and linearly mapping them: MA=Multihead(Q,K,V)=W O Concat(A1,...,A N ); Where W O It is the linear mapping weight matrix from the splicing layer to the multi-head attention input matrix. The feature matrix MA output by the multi-head attention mechanism is obtained by weighted connection and remapping of the implicit features of the original two models.

Images