TBM tunnel support intelligent decision-making method based on multi-source fusion data

Abstract

The invention discloses a TBM tunnel support intelligent decision-making method based on multi-source fusion data, and relates to the technical field of tunnel construction, and the method is characterized in that geological exploration data, rock mass information, TBM tunneling data and other influence factors closely related to the same support condition and the surrounding rock condition are widely selected; and multiple algorithms are deeply fused, so that intelligent decision-making of the TBM tunnel support parameters is realized, and the intelligent level of the TBM tunnel is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel construction, and more specifically, to a TBM tunnel support intelligent decision-making method based on multi-source fusion data. Background Art

[0002] As a key equipment for tunnel construction, the safe and efficient operation of TBM is directly related to the quality and progress of the project. However, the complex geological conditions and changing environment during tunnel construction have brought great challenges to TBM tunnel support. In recent years, domestic and foreign scholars and engineering technicians have carried out a lot of research work on TBM tunnel support and made remarkable progress. However, due to the diversity and complexity of geological conditions, existing research still faces many challenges. For example, in soft strata and broken zones, TBM excavation is very likely to cause stratum deformation and support failure, which seriously affects the construction progress and safety. Therefore, how to make scientific and reasonable support decisions based on multi-source information fusion, and make real-time dynamic adjustments according to the construction situation, so as to achieve efficient, fast and safe support measures, has become a key technical problem that needs to be solved urgently.

[0003] The selection of existing support parameters is mainly based on the engineering analogy method, which is greatly influenced by human factors, the results are rough, and the support parameter decisions are difficult to quantify. In complex geological environments such as rock bursts and large deformations, there are problems of support waste and support failure, and it is difficult to adapt to the complex and changeable geological environment. Summary of the invention

[0004] The purpose of the present invention is to provide a TBM tunnel support intelligent decision-making method based on multi-source fusion data. The method widely selects various influencing factors closely related to the support situation and surrounding rock situation, such as geological survey data, rock mass information, TBM excavation data, etc., and deeply integrates multiple algorithms based on the support parameters actually implemented on site, so as to realize the intelligent decision-making of TBM tunnel support parameters and improve the intelligence level of TBM tunnel.

[0005] The above technical objectives of the present invention are achieved through the following technical solutions: A TBM tunnel support intelligent decision-making method based on multi-source fusion data, comprising the following steps:

[0006] 1) Extensively collect the excavation parameters of TBM tunnels that have been built or are under construction to form an initial sample library of excavation data;

[0007] 2) The collected TBM tunnel excavation parameters are divided into excavation state and non-excavation state according to formula (1), and the data of excavation state are retained;

[0008] f i =F i ·v i ·Ti ·n i (1)

[0009] Where: F i ——Total thrust at the i-th time step, unit: kN; v i ——the excavation speed at the ith time step, unit: mm / min; T i ——The cutter head torque at the i-th time step, unit: MN·m; n i ——The cutter head speed at the i-th time step, unit: rpm; if f i =0, then all data of the i-th time step are discarded;

[0010] 3) Use the 3σ method to process the excavation status data set and calculate the average value of each excavation parameter With standard deviation σ, remove Abnormal data points outside

[0011] 4) The maximum inter-class difference adaptive method is used to extract the excavation cycle and divide it into the cycle rising section and the stable section. The average value of the stable section of each excavation cycle is retained to represent the state of the excavation parameters at this stage, and it is matched with the excavation mileage, and finally a database of excavation parameters with the excavation mileage as the coordinate is formed;

[0012] 5) Collect geological survey data, design support parameters, and change design parameters of TBM tunnels that have been built or are under construction, take the design support parameters as the benchmark, change the design parameters as the modification sample, and use the mileage as the coordinates to establish a field support parameter combination sample, and splice it with the excavation parameter database of step 4) to form a TBM tunnel excavation support sample database;

[0013] 6) Construct a support decision framework. First, arrange the data in a sequence form, and perform a convolution operation on the arranged input vector using formula (4):

[0014]

[0015] Where x∈R n×d is the input vector, W c ∈R k×d is the convolution kernel, k is the size of the convolution kernel, b 1 is the bias, h 1 [t] is the result of the convolution output at the tth time step.

[0016] Then, a global average pooling operation is performed to calculate the output h of the convolutional layer using formula (5): 1 Perform a global average operation:

[0017]

[0018] Then, a nonlinear transformation is performed using the fully connected layer with the ReLU activation function through formula (6):

[0019] h 3 [i]=max(0,W 2 h 2 +b 2 ) (6)

[0020] Then, a nonlinear transformation is performed using the fully connected layer with a sigmoid activation function through formula (7):

[0021]

[0022] Then, BiLSTM is used to transform it through formula (8):

[0023]

[0024] In the formula, is the forward LSTM, and the calculation formula is shown in formula (9). is the backward LSTM, and the calculation formula is shown in formula (10);

[0025]

[0026] Then, a nonlinear transformation is performed using the fully connected layer with the ReLU activation function through formula (11):

[0027] h 6 [i]=max(0,W 5 h 5 +b 5 ) (11)

[0028] Then, the output is converted into a probability distribution using Softmax through equation (12):

[0029]

[0030] 7) Based on the established support decision framework, the TBM tunnel excavation support sample database established in step 5) is used for training, and finally a TBM tunnel support decision algorithm based on multi-source fusion core data is obtained; this algorithm can be applied to TBM tunnels for intelligent support decision-making and improve the intelligence level of TBM tunnels.

[0031] The present invention is further configured as follows: the steps of the maximum inter-class difference adaptive method are as follows:

[0032] (1) Initialize the threshold g between the rising segment and the stable segment of the total propulsion force of the cycle 0 , and the total propulsion force in the cycle is less than g 0 The number of data is recorded as N 0 , the total thrust is greater than g0 The number of data is recorded as N 1 ;

[0033] (2) The proportion of the ascending segment data in the total cycle propulsion force is recorded as w 0 , and calculate the average value μ of the total propulsion force in the ascending phase of the cycle 0 , if the total propulsion force F <g 0 , it is regarded as the ascending stage; the proportion of the stable stage data in the total propulsion force of the cycle to the entire cycle data is recorded as w 1 , and calculate the average value μ of the total propulsion force in the stable cycle 1 , if the total propulsion force F>=g 0 , it is considered as a stable segment;

[0034] (3) According to the proportion of the rising phase and the stable phase data in the total propulsion force of the cycle w 0 、w 1 and its average value μ 0 , μ 1 , the mean μ and inter-class variance S of the total propulsion force of the entire cycle can be calculated:

[0035]

[0036] By further simplifying, we can obtain the equivalent formula:

[0037] S=w 0 w 1 (μ 0 -μ 1 ) 2 (3)

[0038] (4) The mean value μ of the total propulsion force of the entire cycle is taken as the new threshold g between the rising segment and the stable segment of the total propulsion force of the cycle 1 , and then calculate the new mean μˊ and inter-class variance S′ of the total cycle propulsion force according to steps (1) to (3). By comparing the inter-class variances S and S′, if S′>S, the mean μˊ of the total cycle propulsion force is used as the final threshold between the rising segment and the stable segment of the total cycle propulsion force. At this time, the division of the data between the rising segment and the stable segment of the cycle is completed; otherwise, continue to execute steps (1) to (3) until the cycle division is completed.

[0039] In summary, the present invention has the following beneficial effects:

[0040] 1. The current selection of TBM tunnel support parameters is based on human experience, which is slow and has a lot of subjectivity. The TBM tunnel support decision-making method established based on this technical solution can quantify the decision-making indicators, replace the influence of human factors, and achieve rapid decision-making;

[0041] 2. The input indicators such as excavation parameters used in the method provided by the present invention can provide real-time feedback on the current excavation surrounding rock conditions with high accuracy, thereby achieving real-time decision-making on TBM tunnel support parameters;

[0042] 3. In the past, the factors for support decision selection were relatively simple, such as surrounding rock strength, integrity and other parameters. This technical method is based on a machine learning algorithm and can widely adopt multiple excavation parameters and geological survey data that provide real-time feedback on surrounding rock conditions, and the results are more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is the support decision algorithm framework in Example 1 of the present invention;

[0044] Figure 2 are some of the excavation parameters in Example 2 of the present invention;

[0045] Figure 3 It is the division of the excavation cycle in Embodiment 2 of the present invention;

[0046] Figure 4 This is the result of the support decision model in Example 2 of the present invention. DETAILED DESCRIPTION

[0047] The following is combined with Figure 1-4 The present invention is described in further detail.

[0048] Example 1: A TBM tunnel support intelligent decision-making method based on multi-source fusion data, such as Figure 1 As shown, the following steps are included:

[0049] 1) Extensively collect the excavation parameters of TBM tunnels that have been built or are under construction to form an initial sample library of excavation data;

[0050] 2) The collected TBM tunnel excavation parameters are divided into excavation state and non-excavation state according to formula (1), and the data of excavation state are retained;

[0051] f i =F i ·v i ·T i ·n i (1)

[0052] Where: F i ——Total thrust at the i-th time step, unit: kN; v i ——the excavation speed at the ith time step, unit: mm / min; T i ——The cutter head torque at the i-th time step, unit: MN·m; n i ——The cutter head speed at the i-th time step, unit: rpm; if f i=0, then all data of the i-th time step are discarded;

[0053] 3) Use the 3σ method to process the excavation status data set and calculate the average value of each excavation parameter With standard deviation σ, remove Abnormal data points outside

[0054] 4) The maximum inter-class difference adaptive method is used to extract the excavation cycle and divide it into the cycle rising section and the stable section. The average value of the stable section of each excavation cycle is retained to represent the state of the excavation parameters at this stage, and it is matched with the excavation mileage, and finally a database of excavation parameters with the excavation mileage as the coordinate is formed;

[0055] The steps of the maximum inter-class difference adaptive method are as follows:

[0056] (1) Initialize the threshold g between the rising segment and the stable segment of the total propulsion force of the cycle 0 , and the total propulsion force in the cycle is less than g 0 The number of data is recorded as N 0 , the total thrust is greater than g 0 The number of data is recorded as N 1 ;

[0057] (2) The proportion of the ascending segment data in the total cycle propulsion force is recorded as w 0 , and calculate the average value μ of the total propulsion force in the ascending phase of the cycle 0 , if the total propulsion force F <g 0 , it is regarded as the ascending stage; the proportion of the stable stage data in the total propulsion force of the cycle to the entire cycle data is recorded as w 1 , and calculate the average value μ of the total propulsion force in the stable cycle 1 , if the total propulsion force F>=g 0 , it is considered as a stable segment;

[0058] (3) According to the proportion of the rising phase and the stable phase data in the total propulsion force of the cycle w 0 、w 1 and its average value μ 0 , μ 1 , the mean μ and inter-class variance S of the total propulsion force of the entire cycle can be calculated:

[0059]

[0060] By further simplifying, we can obtain the equivalent formula:

[0061] S=w 0 w 1 (μ 0 -μ 1 ) 2 (3)

[0062] (4) The mean value μ of the total propulsion force of the entire cycle is taken as the new threshold g between the rising segment and the stable segment of the total propulsion force of the cycle 1 , and then calculate the new mean μˊ and inter-class variance S′ of the total cycle propulsion force according to steps (1) to (3). By comparing the inter-class variances S and S′, if S′>S, the mean μˊ of the total cycle propulsion force is used as the final threshold between the rising segment and the stable segment of the total cycle propulsion force. At this time, the division of the data between the rising segment and the stable segment of the cycle is completed; otherwise, continue to execute steps (1) to (3) until the cycle division is completed.

[0063] 5) Collect geological survey data, design support parameters, and change design parameters of TBM tunnels that have been built or are under construction, take the design support parameters as the benchmark, change the design parameters as the modification sample, and use the mileage as the coordinates to establish a field support parameter combination sample, and splice it with the excavation parameter database of step 4) to form a TBM tunnel excavation support sample database;

[0064] 6) Construct a support decision framework. First, arrange the data in a sequence form, and perform a convolution operation on the arranged input vector using formula (4):

[0065]

[0066] Where x∈R n×d is the input vector, W c ∈R k×d is the convolution kernel, k is the size of the convolution kernel, b 1 is the bias, h 1 [t] is the result of the convolution output at the tth time step.

[0067] Then, a global average pooling operation is performed to calculate the output h of the convolutional layer using formula (5): 1 Perform a global average operation:

[0068]

[0069] Then, a nonlinear transformation is performed using the fully connected layer with the ReLU activation function through formula (6):

[0070] h 3 [i]=max(0,W 2 h 2 +b 2 ) (6)

[0071] Then, a nonlinear transformation is performed using the fully connected layer with a sigmoid activation function through formula (7):

[0072]

[0073] Then, BiLSTM is used to transform it through formula (8):

[0074]

[0075] In the formula, is the forward LSTM, and the calculation formula is shown in formula (9). is the backward LSTM, and the calculation formula is shown in formula (10);

[0076]

[0077] Then, a nonlinear transformation is performed using the fully connected layer with the ReLU activation function through formula (11):

[0078] h 6 [i]=max(0,W 5 h 5 +b 5 ) (11)

[0079] Then, the output is converted into a probability distribution using Softmax through equation (12):

[0080]

[0081] 7) Based on the established support decision framework, the TBM tunnel excavation support sample database established in step 5) is used for training, and finally a TBM tunnel support decision algorithm based on multi-source fusion core data is obtained; this algorithm can be applied to TBM tunnels for intelligent support decision-making and improve the intelligence level of TBM tunnels.

[0082] Example 2: Taking a railway TBM tunnel as an example to demonstrate the implementation path of the present invention

[0083] (1) Collect TBM excavation parameters including cutter head speed, cutter head torque, cutter head penetration, propulsion speed, total thrust, shield pressure, shield travel displacement, etc., remove non-excavation states and eliminate abnormal data points. Figure 2 The following is a visual display of some parameters;

[0084] (2) Use the maximum inter-class difference adaptive method to divide the excavation cycle, such as Figure 3 As shown;

[0085] (3) Collect geological survey data, design support parameters, and change design parameters. Take the design support parameters as the benchmark, the changed design parameters as the modification sample, and the mileage as the coordinates to establish the on-site support parameter combination sample. Table 1 shows the classification of the actual on-site support combination, which has 8 categories in total.

[0086] Table 1 Statistics of support parameter combinations

[0087]

[0088] (4) Based on the established TBM tunneling support sample database, support decision algorithm training is carried out. Figure 4 The training results are good, with an accuracy rate of up to 99%.

[0089] This specific embodiment is merely an explanation of the present invention and is not a limitation of the present invention. After reading this specification, those skilled in the art may make non-creative modifications to the present embodiment as needed. However, as long as they are within the scope of the claims of the present invention, they are protected by the patent law.

Claims

1. A TBM tunnel support intelligent decision-making method based on multi-source fusion data, characterized by: It includes the following steps: 1) Widely collect the tunneling parameters of existing or under-construction TBM tunnels to form an initial sample database of tunneling data; 2) Divide the collected TBM tunnel tunneling parameters into tunneling states and non-tunneling states according to Equation (1), and retain the data in the tunneling state; f i =F i ·v i ·T i ·n i (1) Where: F i ——Total thrust at the i-th time step, unit: kN; v i ——the excavation speed at the ith time step, unit: mm / min; T i ——The cutter head torque at the i-th time step, unit: MN·m; n i ——The cutter head speed at the i-th time step, unit: rpm; if f i =0, then all data of the i-th time step are discarded; 3) Use the 3σ method to process the excavation status data set and calculate the average value of each excavation parameter With standard deviation σ, remove Abnormal data points outside 4) Use the maximum between-class variance adaptive method to extract the tunneling cycles, divide them into cycle rising segments and stable segments, retain the average value of the stable segment of each tunneling cycle to represent the state of the tunneling parameters in this stage, and match them with the tunneling mileage. Finally, form a tunneling parameter database with the tunneling mileage as the coordinate; 5) Collect the geological exploration data, design support parameters, and design change parameters of existing or under-construction TBM tunnels. Taking the design support parameters as the benchmark, the design change parameters as the modified samples, and the mileage as the coordinate, establish a field support parameter combination sample, and splice it with the tunneling parameter database in step 4) to form a TBM tunnel tunneling support sample database; 6) Construct a support decision framework. First, arrange the data in a sequence form, and perform a convolution operation on the arranged input vector through Equation (4) Subsequently, perform a global average pooling operation, and perform a global average operation on the output h1 of the convolutional layer according to Equation (5): Where x∈R n×d is the input vector, W c ∈R k×d is the convolution kernel, k is the size of the convolution kernel, b1 is the bias, and h1[t] is the result of the convolution output at the tth time step. Subsequently, perform a non-linear transformation through a fully connected layer using the ReLU activation function according to Equation (6): h3[i] = max(0, W2h2 + b2) (6) Subsequently, perform a non-linear transformation through a fully connected layer using the sigmoid activation function according to Equation (7): Subsequently, perform a transformation using BiLSTM according to Equation (8): Subsequently, perform a non-linear transformation through a fully connected layer using the ReLU activation function according to Equation (11): In the formula, is the forward LSTM, and the calculation formula is shown in formula (9). is the backward LSTM, and the calculation formula is shown in formula (10); h6[i] = max(0, W5h5 + b5) (11) Subsequently, convert the output into a probability distribution using Softmax according to Equation (12): 7) Based on the established support decision framework, train with the TBM tunnel tunneling support sample database established in step 5), and finally obtain a TBM tunnel support decision algorithm based on multi-source fusion core data; using this algorithm can be applied to TBM tunnels for intelligent support decision-making to improve the intelligent level of TBM tunnels.

2. A method for intelligent support decision-making of TBM tunnels based on multi-source fusion data according to claim 1, characterized in that: The steps of the maximum between-class variance adaptive method are as follows: (1) Initialize the threshold g0 between the rising segment and the stable segment in the total advancing force of the cycle, and count the number of data with the total advancing force less than g0 in the cycle as N0, and the number of data with the total advancing force greater than g0 as N1; (2) Denote the proportion of the data in the rising segment of the total advancing force of the cycle in the entire cycle data as w0, and calculate the average value μ0 of the total advancing force in the rising segment of the cycle. If the total advancing force F < g0, it is regarded as the rising segment; denote the proportion of the data in the stable segment of the total advancing force of the cycle in the entire cycle data as w1, and calculate the average value μ1 of the total advancing force in the stable segment of the cycle. If the total advancing force F >= g0, it is regarded as the stable segment; (3) According to the proportions w0, w1 of the data in the rising segment and the stable segment in the total advancing force of the cycle and their average values μ0, μ1, the average value μ of the total advancing force of the entire cycle and the between-class variance S can be calculated: ​ By further simplifying, we can obtain the equivalent formula: S=w0w1(μ0-μ1) 2 (3) (4) The mean value μ of the total propulsion force of the entire cycle is used as the new threshold value g1 between the rising segment and the stable segment of the total propulsion force of the cycle, and then the new mean value μˊ and the inter-class variance S′ of the total propulsion force of the cycle are calculated according to steps (1) to (3). By comparing the inter-class variances S and S′, if S′>S, the mean value μˊ of the total propulsion force of the cycle is used as the final threshold value between the rising segment and the stable segment of the total propulsion force of the cycle, and the division of the data of the rising segment and the stable segment of the cycle is completed; otherwise, continue to execute steps (1) to (3) until the cycle division is completed.

Images