Underground powerhouse multi-working-face excavation organization construction method based on tunnel boring machine (TBM) middle pilot tunnel
By adopting a multi-face excavation construction method based on the pilot tunnel of a TBM, and combining SPSS and Abaqus software with the SSA-BP neural network model, the problem of predicting the deformation of the surrounding rock in the underground powerhouse was solved, thus improving construction efficiency and safety.
Patent Information
- Application Number
- CN202511340600.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-19
- Publication Date
- 2025-12-23
AI Technical Summary
Existing numerical simulation results are insufficient to accurately predict the excavation deformation characteristics of the surrounding rock of underground powerhouses, leading to difficulties in construction guidance.
A multi-face excavation construction method based on the TBM pilot tunnel was adopted. SPSS statistical analysis, Abaqus finite element software and SSA-BP neural network model were combined to optimize the surrounding rock mechanical parameters and predict deformation. The construction organization sequence was reasonably allocated according to the prediction results.
It improved the efficiency of underground powerhouse excavation and support construction, enabled accurate prediction of surrounding rock deformation and reasonable support, and shortened the construction cycle.
Smart Images
Figure CN121189085A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of underground powerhouse construction, in particular to a multi-working face excavation organization construction method of underground powerhouse based on a TBM pilot tunnel. BACKGROUND
[0002] The underground tunnel group of pumped storage power stations has large scale, many projects and large engineering quantity, especially the cross section form of the underground tunnel is various and the size is different, which is more suitable for drill and blast method construction. However, the drill and blast method excavation has low degree of mechanization, large labor input, high safety risk, long construction period and poor working environment, etc. Therefore, the full-face rock tunnel boring machine (TBM) is introduced in some pumped storage power stations, which has been widely used in tunnel excavation of highways, railways and water conservancy projects. The pilot tunnel provides convenience for the construction of the top arch layer of the main powerhouse and shortens the construction period.
[0003] The underground powerhouse is the layout place of the hydraulic generator of the pumped storage power station and is the key construction point, and the simulation and prediction of the excavation displacement of the underground powerhouse has important significance for construction guidance. However, considering the complexity of the construction process of the underground powerhouse, the limitation of the surrounding rock geological survey, and the simplification of the numerical simulation theory and model, the subjectivity of the boundary and parameter condition setting, the numerical simulation result is often difficult to meet the actual excavation deformation characteristics of the surrounding rock, and the surrounding rock excavation deformation prediction cannot be better realized. SUMMARY
[0004] In order to solve the above technical problems, the present application provides a multi-working face excavation organization construction method of underground powerhouse based on a TBM pilot tunnel, which solves the problem that the numerical simulation result of the prior art is often difficult to meet the actual excavation deformation characteristics of the surrounding rock and cannot better realize the surrounding rock excavation deformation prediction.
[0005] To solve the above technical problems, a technical scheme provided by the present application is: a multi-working face excavation organization construction method of underground powerhouse based on a TBM pilot tunnel, which comprises the following steps:
[0006] Step one, based on the statistical analysis software SPSS, the construction process time of each support under different drilling depth conditions is statistically analyzed;
[0007] Step two, a three-dimensional physical model of the underground cavern is established, and the surrounding rock mechanical parameters of the underground powerhouse excavation construction are inversed and optimized by combining the finite element software Abaqus and the SSA-BP neural network model;
[0008] Step 3: Based on the central tunnel formed by the TBM, and according to the construction design requirements of the first floor of the underground powerhouse and real-time data from the construction site, the deformation prediction of the surrounding rock excavation of the underground powerhouse is carried out by combining numerical simulation technology and the inverted and optimized SSA-BP neural network model.
[0009] Step 4: Based on the deformation prediction results and the analysis results from Step 1, rationally allocate the construction organization sequence.
[0010] Furthermore, in step one, the support uses an integrated anchor drilling and grouting trolley to realize the three processes of drilling, grouting, and rod insertion. The time of the three processes of drilling, grouting, and rod insertion at different drilling depths in the past is obtained, and the process that needs to be optimized is determined based on the time of each process.
[0011] Furthermore, in step two, the process of inverting and optimizing the surrounding rock mechanical parameters for the underground powerhouse excavation is as follows:
[0012] The arch of the excavated underground powerhouse was monitored using testing instruments, and testing data was obtained.
[0013] Randomly select monitoring points and form a dataset from the monitoring point data. Then, proportionally allocate the dataset to form a training dataset and an inversion dataset.
[0014] The SSA-BP neural network model was trained using a training dataset and then displayed using the finite element software Abaqus.
[0015] The trained SSA-BP neural network model is processed by inverting the data set, and the results are compared with the actual monitoring data to adjust the relevant parameters.
[0016] Furthermore, the core issue in the inversion of surrounding rock mechanical parameters is to find a reasonable combination of rock mass mechanical parameters within a reasonable range of values, such that the sum of the absolute values of the differences between the numerical simulation calculation results and the field monitoring results of the displacements at each measuring point is minimized. This combination can be considered to be in good agreement with the field rock mass parameters, and can be expressed as follows:
[0017]
[0018] In the formula: m is the number of monitoring points, The calculated deformation and displacement values of the surrounding rock; These are the deformation and displacement values of the surrounding rock monitored on-site. E1 and E2 are the elastic moduli of Class III and IV surrounding rock, respectively; c1 and c2 are the cohesion of the corresponding surrounding rock. and This represents the friction angle corresponding to the surrounding rock.
[0019] Due to the displacement value monitored on site If Y is a fixed value, the key to the above problem lies in establishing a mapping f with nonlinear and highly complex characteristics, and solving for the set of surrounding rock parameters that can minimize Y.
[0020] By utilizing the backpropagation characteristics of BP neural networks, a nonlinear mapping model is directly established based on the input and output. At the same time, the SSA algorithm is used to find the optimal neural network weights and thresholds, thus solving the problem that BP neural network models are prone to getting trapped in local optima.
[0021] Furthermore, the specific parameter settings for the SSA-BP neural network model are as follows:
[0022] Construct a BP neural network with 6 input nodes, 1 output node, and 4 hidden layer nodes, which is the topology of the BP neural network.
[0023] Set the BP network training parameters, with 1000 training iterations and a training target value of 10. -6 The learning rate is set to 0.001;
[0024] SSA algorithm parameter initialization: initial sparrow population size is 20, maximum number of iterations is 50, explorer proportion (PD) is 0.2, scout proportion (SD) is 0.2, and warning value (ST) is set to 0.8.
[0025] Furthermore, the construction design requirements for the first floor of the underground plant are as follows: based on the layout characteristics of the TBM central tunnel, the excavation of the main plant's roof arch layer will proceed simultaneously from the ventilation tunnels and traffic tunnels at both ends of the plant, forming a multi-face excavation scheme.
[0026] Furthermore, the reasonable allocation of the construction organization sequence is as follows: the arch layer of the factory building is excavated in sections, the main concentrated location of deformation in each section is determined according to the predicted data, and the support priority is determined according to the degree of deformation and the time of each construction procedure under different drilling depth conditions. The anchor bolts that need to be supported first and the anchor bolts that can be supported later are identified, and the support operation is carried out according to the priority.
[0027] The factory building's arched roof is divided into three zones: the central tunnel excavation zone, the sidewall expansion excavation zone I, and the sidewall expansion excavation zone II.
[0028] Furthermore, the support operation in the central pilot tunnel excavation area is as follows: The main location of the deformation during the pilot tunnel excavation is concentrated at the top of the central pilot tunnel excavation face. The overall deformation is relatively small and within a controllable range. Therefore, after the TBM central pilot tunnel, priority support is given to the top. The specific parameters are: one 6m mortar anchor rod C28 and one 9m prestressed anchor rod C32, arranged at 1.5m intervals.
[0029] Furthermore, the support operation for the sidewall expansion excavation zone I is as follows: The main location of the excavation deformation in the sidewall expansion excavation zone is concentrated at the top of the main plant roof arch excavation surface. The closer to the sidewall, the smaller the deformation value. Therefore, during the trenching and expansion excavation of the expansion zone I, the top is supported. The specific parameters are: 3 six-meter mortar anchor rods C28 and 3 nine-meter prestressed anchor rods C32, arranged alternately at 1.5 meters intervals, and activated according to the trenching position.
[0030] Furthermore, the support operation for the excavation of the sidewall in Zone II is as follows: During the excavation of the sidewall in Zone II, the specific support parameters are: 7 C28 6m mortar anchors and 7 C32 9m prestressed anchors, staggered at 1.5m intervals.
[0031] The beneficial effects of this invention are as follows:
[0032] This application addresses the construction characteristics of TBMs where the pilot tunnel is constructed first, proposing a multi-face excavation organization strategy. Simultaneously, during the excavation process, based on surrounding rock deformation monitoring data, and combining the SSA-BP neural network model with the finite element software Abaqus, the mechanical parameters of the surrounding rock are inverted and analyzed. The numerical simulation results obtained from the inversion, along with the SSA-BP neural network model, are used to predict the excavation deformation of the underground powerhouse. Based on the deformation prediction results, a zoned support organization strategy is further proposed, significantly improving the construction efficiency of underground powerhouse excavation and support.
[0033] To make the above and other objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached image description:
[0034] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only twelve of the drawings in this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0035] Figure 1 This is a flowchart of the application.
[0036] Figure 2 Flowchart for the inversion of surrounding rock mechanical parameters in the SSA-BP model;
[0037] Figure 3 A comparison chart of monitoring point data and simulation data from the inversion scheme;
[0038] Figure 4 A comparison chart of prediction results for the training set;
[0039] Figure 5 A comparison chart of prediction results for the test set;
[0040] Figure 6 This is a schematic diagram of the excavation of the first floor of the main and auxiliary powerhouses based on the TBM pilot tunnel;
[0041] Figure 7 This is a cross-sectional view of the excavation of the first floor of the main and auxiliary powerhouses based on the TBM pilot tunnel;
[0042] Figure 8 Deformation cloud diagram of the pilot tunnel excavation on the first floor of the main plant (unit: m);
[0043] Figure 9 Schematic diagram of the support structure for the expansion excavation of the pilot tunnel in Zone I of the main plant building;
[0044] Figure 10 Deformation cloud map of the excavation area of the main plant's first floor (unit: m);
[0045] Figure 11 Schematic diagram of the support structure for the excavated section I of the main plant building;
[0046] Figure 12 Schematic diagram of the support structure for the expanded excavation of Zone II of the main plant building (Level I). Detailed implementation method:
[0047] Embodiments of the invention will now be described in detail with reference to the accompanying drawings. While some embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the invention. It should be understood that the drawings and embodiments of the invention are for illustrative purposes only and are not intended to limit the scope of protection of the invention.
[0048] The names of messages or information exchanged between the various devices, systems, equipment, and modules in the embodiments of this invention are for illustrative purposes only and are not intended to limit the scope of these messages or information.
[0049] Example
[0050] like Figures 1-12 As shown, a multi-face excavation and construction method for an underground powerhouse based on a TBM pilot tunnel is described, and its steps are as follows:
[0051] Step S1: Based on the statistical analysis software SPSS, perform statistical analysis on the construction time of each construction procedure of the support under different drilling depth conditions;
[0052] Step S2: Establish a three-dimensional physical model of the underground cavern, and combine the finite element software Abaqus with the SSA-BP neural network model to perform inversion optimization of the surrounding rock mechanical parameters for the excavation of the underground powerhouse.
[0053] Step S3: Based on the central tunnel formed by the TBM, according to the construction design requirements of the first floor of the underground powerhouse and real-time data from the construction site, the deformation prediction of the surrounding rock excavation of the underground powerhouse is carried out by combining numerical simulation technology and the inverted and optimized SSA-BP neural network model.
[0054] Step S4: Based on the deformation prediction results and the analysis results from Step 1, rationally allocate the construction organization sequence.
[0055] In step one, the support uses an integrated anchor drilling and grouting trolley to complete the three processes of drilling, grouting, and rod insertion. The time for the three processes of drilling, grouting, and rod insertion at different drilling depths is obtained, and the processes that need to be optimized are determined based on the time of each process.
[0056] The drilling depths are generally 6 meters and 9 meters. Therefore, this study focuses on these two drilling depths and uses an integrated anchor drilling and grouting trolley to measure the time required for the three processes of drilling, grouting, and rod insertion at 6-meter and 9-meter drilling depths. Specifically, the study collects past operation times and uses a trimmed average method to obtain the specific operation time.
[0057] In step two, to implement the inversion process, an L is designed with elastic modulus, cohesion, and friction angle as variables. 25 (5 3 An orthogonal experimental scheme (see Table 1) was used. An SSA-BP inversion analysis program was written using Matlab software, and numerical simulation calculations were performed using the finite element software Abaqus to realize the inversion study of the surrounding rock mechanical parameters of underground powerhouse excavation based on the SSA-BP neural network.
[0058] Table 1. Inversion parameters of surrounding rock elastic-plastic mechanics from orthogonal test L 25 (5 3 )design
[0059]
[0060] Considering that the parameters to be inverted have a certain range and many possible combinations, in order to ensure generality and minimize the influence of random errors, the inversion results of the SSA-BP neural network were screened, and a set of relatively reasonable parameter combinations was selected (see Table 2) for presentation. After numerical simulation calculation, the results are as follows: Figure 2 As shown, the deformation calculation results under this set of parameters can accurately reflect the deformation characteristics of the surrounding rock excavation, and can be used for subsequent excavation deformation prediction analysis.
[0061] Table 2 Comparison of Surrounding Rock Inversion Parameter Schemes and Monitoring Values
[0062]
[0063] The process of inverting and optimizing the surrounding rock mechanical parameters for underground powerhouse excavation is as follows:Figure 3 As shown,
[0064] The excavated underground powerhouse roof arch was monitored and data was obtained using testing instruments (anchor bolt stress gauge, four-point multi-point displacement gauge, prestressed anchor bolt force gauge, anchor cable stress gauge, etc.).
[0065] Randomly select monitoring points and form a dataset from the monitoring point data. Then, proportionally allocate the dataset to form a training dataset and an inversion dataset.
[0066] The SSA-BP neural network model was trained using a training dataset and then displayed using the finite element software Abaqus.
[0067] The trained SSA-BP neural network model is processed by inverting the data set, and the results are compared with the actual monitoring data to adjust the relevant parameters.
[0068] The core issue in the inversion of surrounding rock mechanical parameters is to find a reasonable combination of rock mass mechanical parameters within a reasonable range of values, such that the sum of the absolute values of the differences between the numerical simulation calculation results and the field monitoring results of the displacements at each measuring point is minimized. This combination is considered to be in good agreement with the field rock mass parameters and can be expressed as follows:
[0069]
[0070] In the formula: m is the number of monitoring points, The calculated deformation and displacement values of the surrounding rock; These are the deformation and displacement values of the surrounding rock monitored on-site. E1 and E2 are the elastic moduli of Class III and IV surrounding rock, respectively; c1 and c2 are the cohesion of the corresponding surrounding rock. and This represents the friction angle corresponding to the surrounding rock.
[0071] Due to the displacement value monitored on site If Y is a fixed value, the key to the above problem lies in establishing a mapping f with nonlinear and highly complex characteristics, and solving for the set of surrounding rock parameters that can minimize Y.
[0072] By utilizing the backpropagation characteristics of BP neural networks, a nonlinear mapping model is directly established based on the input and output. At the same time, the SSA algorithm is used to find the optimal neural network weights and thresholds, thus solving the problem that BP neural network models are prone to getting trapped in local optima.
[0073] The specific parameter settings for the SSA-BP neural network model are as follows:
[0074] Construct a BP neural network with 6 input nodes, 1 output node, and 4 hidden layer nodes, which is the topology of the BP neural network.
[0075] Set the BP network training parameters, with 1000 training iterations and a training target value of 10. -6 The learning rate is set to 0.001;
[0076] SSA algorithm parameter initialization: initial sparrow population size is 20, maximum number of iterations is 50, explorer proportion (PD) is 0.2, scout proportion (SD) is 0.2, and warning value (ST) is set to 0.8.
[0077] The monitoring data was divided into training and testing sets in an 8:2 ratio. The BP neural network and SSA-BP neural network were trained using data from every 7 days to establish a nonlinear mapping relationship between the deformation values of the first 6 days and the deformation values of the 7th day. Figure 4 and Figure 5 The comparison shows the prediction results for the training set and the test set, respectively. Figure 4 As shown, in the training set results, the predicted values of both models fit the measured values well. However, at certain deformation points, especially those with large deformation amplitudes, the BP neural network exhibits a significant error between the predicted and simulated values. This indicates that the BP neural network has limitations in predicting deformation trends. Figure 5 As shown, in the test set results, the SSA-BP neural network is more accurate in judging the upward trend of the data, and the prediction results are closer to the simulated values, thus outperforming the BP neural network.
[0078] To better demonstrate the superiority of the SSA-BP neural network model, the error judgment metrics (MAE, MBE, RMSE) of the BP neural network and the SSA-BP neural network are compared using the test set results as an example. Table 3 shows that the SSA-BP neural network outperforms the BP neural network in all error metrics, proving that the SSA-BP neural network has better accuracy and performance than the BP neural network in predicting underground powerhouse excavation.
[0079] Table 3 Comparison of Evaluation Indicators for Different Neural Network Methods
[0080]
[0081] In step S3, the construction design requirements for the first floor of the underground plant are as follows: based on the layout characteristics of the pilot tunnel in the TBM, the excavation of the main plant's roof arch layer is carried out simultaneously from the ventilation tunnels and traffic tunnels at both ends of the plant, forming a multi-face excavation scheme.
[0082] The multi-face excavation scheme is as follows: Figure 6As shown, the left end wall area of the plant will be excavated: excavation will be carried out upstream and downstream of the opening of the left-end central guide tunnel, requiring the upstream working face to be staggered by 30m from the downstream working face and proceeding to both sides. The right end wall area of the plant will also be excavated: excavation will be carried out upstream and downstream of the opening of the right-end central guide tunnel, requiring the upstream working face to be staggered by 30m from the downstream working face and proceeding to both sides. The excavation of both ends will proceed simultaneously, with the downstream of the left end excavation leading the upstream, and the upstream of the right end excavation leading the downstream, ensuring that the deformation characteristics on both sides are basically consistent, facilitating subsequent support construction. If there is a stronger construction schedule requirement, additional trenching for excavation in the middle of the first floor of the main plant can also be considered.
[0083] In step S4, the construction organization sequence is reasonably allocated as follows: the roof arch layer of the factory building is excavated in sections, the main concentrated location of deformation in each section is determined according to the predicted data, and the support priority is determined according to the degree of deformation and the time of each construction procedure under different drilling depth conditions. The anchor bolts that need to be supported first and the anchor bolts that can be supported later are identified, and the support operation is carried out according to the priority.
[0084] The roof arch layer of the factory building is divided into three zones: the central tunnel excavation zone, the sidewall widening excavation zone I, and the sidewall widening excavation zone II. Figure 7 As shown.
[0085] The support operation in the excavation area of the central pilot tunnel is as follows: Figure 8 and Figure 9 As shown, the main deformation during the excavation of the central pilot tunnel is concentrated at the top of the excavation face. The overall deformation is relatively small and within a controllable range. Therefore, after the TBM central pilot tunnel, priority support is given to the top. The specific parameters are: one 6m mortar anchor C28 and one 9m prestressed anchor C32, arranged at 1.5m intervals.
[0086] The support operation for the sidewall excavation zone I is as follows: Figure 10 and Figure 11 As shown, the main deformation of the sidewall excavation area is concentrated at the top of the arch excavation surface of the main plant roof. The closer to the sidewall, the smaller the deformation value. Therefore, during the trenching and excavation of the expansion area I, the top is supported. The specific parameters are: 3 six-meter mortar anchor rods C28 and 3 nine-meter prestressed anchor rods C32, arranged alternately at 1.5 meters intervals, and activated according to the trenching position.
[0087] The support operation for the expanded excavation zone II of the sidewall is as follows: Figure 10 and Figure 12 As shown, during the trenching and widening process in Zone II, the specific support parameters are: 7 x 6m C28 mortar anchors and 7 x 9m C32 prestressed anchors, staggered at 1.5m intervals.
[0088] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.
Claims
1. A method for organizing the excavation and construction of an underground powerhouse with multiple working faces based on a TBM pilot tunnel, comprising the following steps: Step 1: Based on the statistical analysis software SPSS, perform statistical analysis on the construction time of each construction procedure of the support under different drilling depth conditions; Step 2: Establish a three-dimensional physical model of the underground cavern, and combine the finite element software Abaqus with the SSA-BP neural network model to perform inversion optimization of the surrounding rock mechanical parameters for the excavation of the underground powerhouse. Step 3: Based on the central tunnel formed by the TBM, and according to the construction design requirements of the first floor of the underground powerhouse and real-time data from the construction site, the deformation prediction of the surrounding rock excavation of the underground powerhouse is carried out by combining numerical simulation technology and the inverted and optimized SSA-BP neural network model. Step 4: Based on the deformation prediction results and the analysis results from Step 1, rationally allocate the construction organization sequence.
2. The method for organizing and constructing multi-face excavation of an underground powerhouse based on a TBM pilot tunnel as described in claim 1, characterized in that: In step one, the support uses an integrated anchor drilling and grouting trolley to realize the three processes of drilling, grouting, and rod insertion. The time of the three processes of drilling, grouting, and rod insertion at different drilling depths in the past is obtained, and the process that needs to be optimized is determined based on the time of each process.
3. The method for organizing and constructing multi-face excavation of an underground powerhouse based on a TBM pilot tunnel as described in claim 1, characterized in that: In step two, the process of inverting and optimizing the surrounding rock mechanical parameters for the underground powerhouse excavation is as follows: The arch of the excavated underground powerhouse was monitored using testing instruments, and testing data was obtained. Randomly select monitoring points and form a dataset from the monitoring point data. Then, proportionally allocate the dataset to form a training dataset and an inversion dataset. The SSA-BP neural network model was trained using a training dataset and then displayed using the finite element software Abaqus. The trained SSA-BP neural network model is processed by inverting the data set, and the results are compared with the actual monitoring data to adjust the relevant parameters.
4. The method for organizing and constructing multi-face excavation of an underground powerhouse based on a TBM pilot tunnel as described in claim 3, characterized in that: The core issue in the inversion of surrounding rock mechanical parameters is to find a reasonable combination of rock mass mechanical parameters within a reasonable range of values, such that the sum of the absolute values of the differences between the numerical simulation calculation results and the field monitoring results of the displacements at each measuring point is minimized. This combination can be considered to be in good agreement with the field rock mass parameters, and can be expressed as follows: In the formula: m is the number of monitoring points, The calculated deformation and displacement values of the surrounding rock; These are the deformation and displacement values of the surrounding rock monitored on-site. E1 and E2 are the elastic moduli of Class III and IV surrounding rock, respectively; c1 and c2 are the cohesion of the corresponding surrounding rock. and This represents the friction angle corresponding to the surrounding rock. Due to the displacement value monitored on site If Y is a fixed value, the key to the above problem lies in establishing a mapping f with nonlinear and highly complex characteristics, and solving for the set of surrounding rock parameters that can minimize Y. By utilizing the backpropagation characteristics of BP neural networks, a nonlinear mapping model is directly established based on the input and output. At the same time, the SSA algorithm is used to find the optimal neural network weights and thresholds, thus solving the problem that BP neural network models are prone to getting trapped in local optima.
5. The method for organizing and constructing multi-face excavation of an underground powerhouse based on a TBM pilot tunnel as described in claim 4, characterized in that: The specific parameter settings for the SSA-BP neural network model are as follows: Construct a BP neural network with 6 input nodes, 1 output node, and 4 hidden layer nodes, which is the topology of the BP neural network. Set the BP network training parameters, with 1000 training iterations and a training target value of 10. -6 The learning rate is set to 0.001; SSA algorithm parameter initialization: initial sparrow population size is 20, maximum number of iterations is 50, explorer proportion (PD) is 0.2, scout proportion (SD) is 0.2, and warning value (ST) is set to 0.
8.
6. The method for organizing and constructing multi-face excavation of an underground powerhouse based on a TBM pilot tunnel as described in claim 1, characterized in that: underground The construction design requirements for the first floor of the plant are as follows: Based on the layout characteristics of the TBM central tunnel, the excavation of the main plant's roof arch layer will proceed simultaneously from the ventilation tunnels and traffic tunnels at both ends of the plant, forming a multi-face excavation scheme.
7. The method for organizing and constructing multi-face excavation of an underground powerhouse based on a TBM pilot tunnel according to claim 1, characterized in that: The reasonable allocation of construction organization sequence is as follows: the arch layer of the factory building is excavated in sections, the main concentrated location of deformation in each section is determined according to the predicted data, and the support priority is determined according to the degree of deformation and the time of each construction procedure under different drilling depth conditions. The anchor bolts that need to be supported first and the anchor bolts that can be supported later are identified, and the support operation is carried out according to the priority. The factory building's arched roof is divided into three zones: the central tunnel excavation zone, the sidewall expansion excavation zone I, and the sidewall expansion excavation zone II.
8. The method for organizing and constructing multi-face excavation of an underground powerhouse based on a TBM pilot tunnel as described in claim 7, characterized in that: The support operation in the pilot tunnel excavation area is as follows: The main deformation of the pilot tunnel excavation is concentrated at the top of the central pilot tunnel excavation face. The overall deformation is small and within the controllable range. Therefore, after the central pilot tunnel of the TBM, the top is given priority support. The specific parameters are: one 6m mortar anchor C28 and one 9m prestressed anchor C32, arranged at a distance of 1.5m.
9. The method for organizing and constructing multi-face excavation of an underground powerhouse based on a TBM pilot tunnel as described in claim 7, characterized in that: The support operation for the sidewall expansion excavation zone I is as follows: The main deformation of the sidewall expansion excavation zone is concentrated at the top of the main plant roof arch excavation surface. The closer to the sidewall, the smaller the deformation value. Therefore, during the trenching and expansion excavation of the expansion zone I, the top is supported. The specific parameters are: 3 six-meter mortar anchor rods C28 and 3 nine-meter prestressed anchor rods C32, staggered at 1.5 meters intervals, and activated according to the trenching position.
10. The method for organizing and constructing multi-face excavation of an underground powerhouse based on a TBM pilot tunnel according to claim 7, characterized in that: The support operation for the sidewall excavation area II is as follows: During the trenching and excavation process of the sidewall area II, the specific support parameters are: 7 C28 6m mortar anchors and 7 C32 9m prestressed anchors, arranged alternately at 1.5m intervals.