Big data driven ground steel bar connection optimization method based on neural network

Abstract

The application belongs to the technical field of information technology, and specifically discloses a big data driving foundation steel bar connection optimization method based on a neural network, which comprises the following steps: acquiring mineral composition data and expansion characteristic indexes through soil sample collection, obtaining a soil expansion potential value by adopting X-ray diffraction analysis and processing the mineral composition data and the expansion characteristic indexes; simulating interface gap width changes according to the soil expansion potential value, determining potential seepage path distribution by adopting finite element analysis to construct a seepage channel model; if the distribution risk value of the potential seepage path in the seepage path distribution exceeds a preset threshold value, adjusting pore structure parameters, obtaining an optimized permeability coefficient value by predicting a permeability coefficient range under geological conditions through a neural network; generating a sealing measure configuration scheme for the optimized permeability coefficient value, and judging the effectiveness of the sealing measure by adopting iterative simulation to verify the interface gap closure degree; and the application aims to solve the problems of poor reliability and adaptability of underground engineering waterproof sealing in the prior art.

Description

Technical Field

[0001] This invention relates to the technical field of information technology, specifically to a big data-driven method for optimizing foundation rebar connections based on neural networks. Background Technology

[0002] In the field of foundation engineering, especially in cohesive foundations with high groundwater levels, the waterproofing and sealing of steel reinforcement connections directly affects the long-term stability and durability of building structures. Once seepage channels appear along the surface of the steel reinforcement, continuous groundwater intrusion will occur, seriously threatening foundation safety. This problem is becoming increasingly prominent in urban underground space development and infrastructure construction. Currently, many waterproofing designs rely primarily on empirical formulas and standard specifications, or adjust schemes through field tests. However, these methods often fail to fully consider the complex behavior of cohesive soils under different mineral compositions and expansion characteristics. Especially at the interface between the soil and steel reinforcement, changes in the soil's pore structure and moisture content significantly affect the formation of seepage paths, leading to insufficient adaptability of waterproofing measures and difficulty in maintaining reliable sealing effects under diverse geological conditions.

[0003] There is a close correlation between the permeability coefficient and void ratio of cohesive soil and the sealing performance of the rebar connection interface. Differences in the clay mineral composition directly affect its expansibility, thus altering the width and connectivity of microscopic fissures at the interface. When clay has high expansibility, the interface fissures may temporarily close, only to reopen under drying or loading, forming hidden seepage channels. In silty clay with lower expansibility, the fissures are more likely to remain stable and open, leading to long-term groundwater infiltration along the rebar surface. These interconnected mechanisms introduce greater uncertainty into interface sealing design, as regional differences in soil properties amplify the risk of seepage, making it difficult for traditional fixed waterproofing methods to simultaneously address the seepage prevention needs of different clay types.

[0004] For example, in highly plastic clay foundations, when the rebar anchorage section is long, although bentonite-based materials can fill part of the gaps through expansion, the excessively long seepage path may still bypass the coating and form channels. In low-plasticity silty clay, while shortening the anchorage length reduces the seepage path, the larger gap width necessitates additional waterproofing structures to block penetration. The waterproofing requirements in these two cases are contradictory, and a single design cannot simultaneously meet them. Accurately grasping the dynamic relationship between clay mineral composition, expansibility, and the gap width at the rebar connection interface in cohesive foundations with high groundwater levels, and flexibly adjusting interface sealing measures and node arrangement accordingly, has become a key issue in achieving reliable waterproofing performance. Therefore, optimizing the seepage prevention design of rebar connections based on the intrinsic connection between soil permeability characteristics and interface microchannels has become a core technical challenge that urgently needs to be solved. Summary of the Invention

[0005] This invention provides a big data-driven optimization method for foundation steel reinforcement connections based on neural networks, aiming to solve the problems of poor reliability and adaptability of waterproof sealing in underground engineering in the prior art.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0007] A big data-driven optimization method for foundation rebar connections based on neural networks includes: acquiring mineral composition data and expansion characteristic indicators through soil sample collection; processing the mineral composition data and expansion characteristic indicators using X-ray diffraction analysis to obtain the soil expansion potential value; simulating the interface gap width change based on the soil expansion potential value; constructing a seepage channel model using finite element analysis to determine the potential seepage path distribution; if the distribution risk value of the potential seepage path in the seepage path distribution exceeds a preset threshold, adjusting the pore structure parameters and predicting the permeability coefficient range under geological conditions using a neural network to obtain an optimized permeability coefficient value; generating a sealing measure configuration scheme based on the optimized permeability coefficient value; using iterative simulation to verify the degree of interface gap closure and judge the effectiveness of the sealing measures; acquiring the verified sealing measure effectiveness data; combining the groundwater level simulation scenario and updating the rebar connection node layout through a dynamic adjustment algorithm to obtain a preliminary waterproof sealing design; evaluating the seepage channel blocking effect based on the preliminary waterproof sealing design; if the blocking effect is lower than a preset threshold, incorporating the expansion characteristic feedback loop to recalculate the mineral composition influence to obtain an enhanced sealing measure scheme; integrating geological condition variables through the enhanced sealing measure scheme, and finally determining the optimized waterproof sealing result of rebar connections adapted to diverse geological conditions using finite element analysis.

[0008] In one aspect of the present invention, the step of acquiring mineral composition data and swelling characteristic indicators through soil sample collection, and processing the mineral composition data and swelling characteristic indicators using X-ray diffraction analysis to obtain the soil swelling potential value includes:

[0009] Soil samples and mineral composition data, as well as swelling characteristics, were collected through sample acquisition.

[0010] The diffraction pattern was obtained by processing the mineral composition data using X-ray diffraction analysis.

[0011] Mineral composition is determined by obtaining mineral types and relative contents from diffraction patterns;

[0012] The proportion of swelling-active minerals is determined based on the mineral composition and swelling characteristics. If the proportion of swelling-active minerals is higher than the preset threshold, it is marked as high swelling potential; otherwise, it is marked as low swelling potential.

[0013] The initial expansion potential value is obtained by calculating the weighted expansion index by combining expansion characteristic indicators with the proportion of expansion-active minerals.

[0014] The random forest algorithm is used to process the initial expansion potential value and mineral composition to determine the corrected expansion potential value;

[0015] Based on the modified expansion potential value and expansion characteristic index, the support vector machine algorithm is used to classify the soil expansion level to obtain the expansion potential classification result.

[0016] In one aspect of the present invention, the step of simulating the change in interface gap width based on the soil swelling potential value and using finite element analysis to construct a seepage channel model to determine the potential seepage path distribution includes:

[0017] A finite element mesh model was established based on the soil expansion potential value to simulate the change in interface gap width.

[0018] Obtain gap width distribution data from the change in interface gap width;

[0019] The initial seepage path is determined by constructing a seepage channel model using the seepage equation;

[0020] The seepage intensity distribution is obtained by calculating the seepage velocity field through the initial seepage path;

[0021] High-risk seepage areas are identified based on the distribution of seepage intensity and the distribution of crack width.

[0022] If the proportion of high-risk seepage areas is higher than a preset threshold, then high-risk seepage paths are marked; otherwise, low-risk seepage paths are marked.

[0023] The support vector machine algorithm is used to process the distribution of high seepage risk paths and seepage intensity to classify the potential seepage path distribution and obtain the final seepage risk level.

[0024] In one aspect of the invention, if the distribution risk value of potential seepage paths in the seepage path distribution exceeds a preset threshold, adjusting the pore structure parameters and obtaining an optimized permeability coefficient value by predicting the permeability coefficient range under geological conditions through a neural network includes:

[0025] By analyzing the correlation between seepage path and distribution state, the judgment result of whether the distribution state exceeds the preset threshold is obtained;

[0026] Based on the judgment result of the distribution state, if it exceeds the preset threshold, the structural parameters are adjusted according to the correlation between the pore structure and the structural parameters to obtain the adjusted pore structure data.

[0027] Using adjusted pore structure data and combining the correlation between geological conditions and permeability coefficient, a basic dataset for permeability coefficient prediction is constructed.

[0028] By processing the basic dataset through a neural network model, and considering the correlation between the permeability coefficient and its range, the optimized permeability coefficient range data is determined.

[0029] Based on the optimized permeability coefficient range data, and combined with the correlation between structural parameters and optimized data, an updated adjustment strategy is obtained.

[0030] By updating the scheme, the potential changing trend of the seepage path is determined based on the correlation between the seepage path and geological conditions, and the final path distribution optimization result is obtained.

[0031] By using the final path distribution optimization results and combining the correlation between the distribution status and the adjustment strategy, the long-term stability data of the seepage path are determined.

[0032] In one aspect of the invention, the step of generating a sealing measure configuration scheme based on the optimized permeability coefficient value and using iterative simulation to verify the degree of closure of the interface gaps to determine the effectiveness of the sealing measures includes:

[0033] Based on the optimized permeability coefficient value, an initial sealing measure configuration scheme is generated;

[0034] Data on the closure of interfacial gaps were obtained by iteratively simulating the seepage process.

[0035] Based on the data on the closure of the interface gaps, determine whether the sealing measures configuration scheme meets the closure requirements;

[0036] If the sealing configuration scheme does not meet the closure requirements, the sealing material distribution parameters are adjusted to obtain updated sealing configuration data;

[0037] Using updated sealing measures and configuration data, an interface stress distribution dataset is constructed.

[0038] The interface stress distribution dataset was processed by finite element analysis to determine the interface closure pressure value.

[0039] Based on the interface closure pressure value, a sealing layer thickness adjustment scheme is generated;

[0040] A sealing layer thickness adjustment scheme was adopted to predict seepage barrier stability data.

[0041] In one aspect of the invention, obtaining verified sealing measure effectiveness data and updating the rebar connection node layout through a dynamic adjustment algorithm based on a groundwater level simulation scenario to obtain a preliminary waterproof sealing design includes:

[0042] After obtaining the verified sealing measures data, construct the corresponding scenario dataset for the simulated scenario of groundwater level change, and obtain the initial mapping relationship between the scenario and the sealing measures.

[0043] Based on the initial mapping relationship and combined with groundwater level data, the stress distribution of the steel reinforcement connection nodes is analyzed using a preset threshold range to determine the initial adjustment direction of the node layout.

[0044] By initially adjusting the direction, a dynamic adjustment strategy is implemented based on the stress distribution data of the rebar connection nodes to update the node layout structure and obtain optimized layout configuration information.

[0045] Using the optimized layout configuration information and combined with the design requirements for waterproof sealing, a corresponding sealing layer distribution model is generated. It is then determined whether the sealing layer distribution meets the preset coverage standard. If not, the stress distribution data is reprocessed.

[0046] Based on the sealing layer distribution model, we obtained the seepage barrier data in the groundwater level simulation scenario, analyzed the stability performance of the seepage barrier, and determined the applicability assessment results of the waterproof sealing scheme.

[0047] Based on the applicability assessment results and the stability data of the waterproof sealing scheme, a final preliminary scheme dataset is constructed to obtain waterproof sealing configuration schemes suitable for different simulation scenarios.

[0048] In one aspect of the invention, the evaluation of the seepage channel blocking effect based on the preliminary waterproof sealing design, and if the blocking effect is lower than a preset threshold, incorporates an expansion characteristic feedback loop to obtain an enhanced sealing measure scheme by recalculating the influence of mineral composition, including:

[0049] Obtain seepage channel data corresponding to the preliminary waterproof sealing design;

[0050] Calculate the blocking performance value based on the seepage channel data;

[0051] Determine the relationship between the blocking performance value and the preset threshold. If the blocking performance value is lower than the preset threshold, start the expansion characteristic feedback loop.

[0052] Mineral composition parameters are extracted from the expansion characteristic feedback loop;

[0053] The intensity distribution of the action was recalculated based on the mineral composition parameters;

[0054] By updating the sealing measure configuration through the distribution of action intensity, an enhanced sealing measure scheme is obtained;

[0055] The seepage channel blocking data was regenerated using an enhanced sealing solution.

[0056] In one aspect of the invention, the method of integrating geological condition variables through enhanced sealing measures and using finite element analysis to ultimately determine the optimized results of waterproof sealing for steel reinforcement connections adapted to diverse geological conditions includes:

[0057] Obtain basic data for sealing measures; and extract corresponding environmental parameters from a pre-established geological condition database to obtain preliminary environmental adaptability data, taking into account the characteristics of the steel reinforcement connection structure.

[0058] Based on preliminary environmental adaptability data, the waterproof performance of the steel reinforcement connection was simulated and analyzed using the finite element method to determine the stress distribution of the connection structure under different geological conditions.

[0059] If the stress distribution exceeds the preset threshold range, the geological condition data of the excess portion will be classified and processed to obtain the connection structure parameters that need to be adjusted.

[0060] By adjusting the connection structure parameters and combining the sealing measures, an optimized configuration scheme is regenerated to determine the direction for improving waterproof performance suitable for specific geological conditions.

[0061] The simulation data of the rebar connection was updated using the improved approach to obtain new environmental adaptability assessment results and determine whether they meet the preset performance assessment standards.

[0062] Based on the updated environmental adaptability assessment results, for the substandard rebar connection parts, the corresponding geological condition variables are extracted to obtain the final optimized configuration adjustment scheme;

[0063] Through the final optimized configuration adjustment scheme, the sealing measures were partially updated to determine the waterproof performance configuration of the steel reinforcement connection that is suitable for various geological environments.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] This invention first obtains the expansion potential value through soil sample collection and X-ray diffraction analysis, and then constructs a seepage channel model using finite element analysis to determine the distribution of potential seepage paths. When the path exceeds the limit, the permeability coefficient range is predicted through neural networks for parameter optimization. Subsequently, a sealing measure configuration scheme is generated and its effectiveness is iteratively verified. The arrangement of steel reinforcement connection nodes is dynamically adjusted by combining groundwater level simulation to form a preliminary design. If the blocking effect is insufficient, an expansion characteristic feedback loop is incorporated to recalculate the influence of mineral composition and propose an enhanced sealing measure scheme. Finally, geological condition variables are integrated through finite element analysis to obtain optimized results that adapt to diverse geological conditions. This invention significantly improves the reliability and adaptability of waterproofing and sealing in underground engineering and effectively reduces the risk of seepage. Attached Figure Description

[0066] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0067] Figure 1This is a flowchart of the big data-driven foundation reinforcement connection optimization method based on neural networks according to the present invention. Detailed Implementation

[0068] The present invention will be further described below with reference to embodiments. These embodiments are merely some, not all, of the embodiments of the present invention. Other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are all within the protection scope of the present invention.

[0069] Please see Figure 1 As shown in the figure, this embodiment discloses a big data-driven method for optimizing foundation rebar connections based on neural networks, which may specifically include:

[0070] S101. Obtain mineral composition data and swelling characteristic indicators by collecting soil samples, and use X-ray diffraction analysis to process the mineral composition data and swelling characteristic indicators to obtain the soil swelling potential value.

[0071] Soil samples, mineral composition data, and swelling characteristic indices were collected through sample acquisition. X-ray diffraction analysis was used to process the mineral composition data and obtain diffraction patterns. The mineral types and relative contents were determined from the diffraction patterns to ascertain the mineral composition. The proportion of swelling-active minerals was determined based on the mineral composition and swelling characteristic indices; if the proportion of swelling-active minerals was higher than a preset threshold, it was marked as high swelling potential; otherwise, it was marked as low swelling potential. A weighted swelling index was calculated using the swelling characteristic indices and the proportion of swelling-active minerals to obtain a preliminary swelling potential value. A random forest algorithm was used to process the preliminary swelling potential value and mineral composition to determine a revised swelling potential value. Based on the revised swelling potential value and the swelling characteristic indices, a support vector machine algorithm was used to classify the soil swelling levels, obtaining the swelling potential classification results.

[0072] Specifically, information technology is used to acquire and analyze the mineral composition data and swelling characteristics of soil samples. First, an automated sampling system is used to collect soil samples at a designated location. Assuming a sampling depth of 1.5 meters and a sample size of 500 grams, the initial moisture content of the soil is recorded as 18.5% using a built-in sensor. Subsequently, the samples are transmitted to a laboratory analysis system, where X-ray fluorescence spectrometry (XRF) is used to detect the mineral composition, obtaining the content data of major minerals, such as 35% montmorillonite, 40% quartz, and 25% feldspar. These data are compared with the system database, and considering the high swelling characteristics of montmorillonite, a preliminary assessment is made that the soil has high swelling potential. Next, the system tests the swelling characteristics of the samples. An automatic pressure dilatometer is used to measure the free swelling rate of the soil, assuming a measured value of 12.3%, and the swelling stress under varying moisture content is recorded as 0.8 MPa. These indicators are then input into the swelling potential assessment model. Furthermore, X-ray diffraction (XRD) technology was used to conduct an in-depth analysis of the soil's crystal structure. The system automatically generated a diffraction pattern, identifying the characteristic peak intensity of montmorillonite as 3.2 (relative unit). Using a preset algorithm, the expansion potential value was calculated as: Expansion Potential = Montmorillonite Content (%) × Characteristic Peak Intensity × Free Expansion Rate (%) × 0.1. The calculated expansion potential value was 35 × 3.2 × 12.3 × 0.1 = 137.76, exceeding the threshold of 100, indicating a high risk of soil expansion. Finally, the system integrated all data into an analysis report, automatically generating a risk level of "High," and linked it to a geological engineering database, recommending targeted reinforcement measures, such as adding 5% lime to improve soil expansibility and ensure project safety. This process formed a complete logical chain from data acquisition to analysis and evaluation, achieving fully automated technical processing.

[0073] S102. Based on the soil expansion potential value, simulate the change in interface gap width, and use finite element analysis to construct a seepage channel model to determine the potential seepage path distribution.

[0074] A finite element mesh model was established based on the soil expansion potential to simulate the variation in interface gap width. Gaps width distribution data was obtained from the variation in interface gap width. A seepage channel model was constructed using the seepage equation to determine the initial seepage path. The seepage velocity field was calculated using the initial seepage path to obtain the seepage intensity distribution. High-risk seepage areas were identified based on the seepage intensity distribution and gap width distribution. If the proportion of high-risk seepage areas exceeded a preset threshold, a high-risk seepage path was marked; otherwise, a low-risk seepage path was marked. A support vector machine algorithm was used to process the high-risk seepage paths and the seepage intensity distribution to classify the potential seepage path distribution and obtain the final seepage risk level.

[0075] Specifically, information technology is used to simulate the effect of soil expansion potential value on the change of interface gap width and to construct a seepage channel model. First, the system imports the previously evaluated expansion potential value of 137.76 as input parameters, combined with a soil modulus of 15 MPa, Poisson's ratio of 0.35, and saturated permeability coefficient of 1.2 × 10⁻⁶. -6 A three-dimensional finite element mesh model was established with a flow rate of m / s and approximately 45,000 mesh elements. Boundary conditions were set as a fixed bottom and a free top, with an applied rainfall infiltration intensity of 50 mm / h. The system ran a hydraulic-mechanical coupled analysis module, using the Van Kinney permeability function to describe the flow in the unsaturated region, i.e., permeability coefficient k(θ) = k_s × [θ - θ_r) / (θ_s - θ_r)]³, where residual water content θ_r = 0.08 and saturated water content θ_s = 0.42. The simulation focused on the volumetric shrinkage process as the water content decreased from an initial 22.5% to a dry state under wet-dry cycles, leading to surface tension-induced crack initiation. Furthermore, the phase-field damage algorithm was used to simulate crack propagation, with the damage variable d satisfying the evolution equation. d / t = (G_c / l_0) × (ψ / σ_c²) × ▽²d, where the critical energy release rate G_c = 150 J / m², the characteristic length l_0 = 0.05 m, and the critical stress σ_c = 0.12 MPa. The initial crack width is calculated to be 0.8 mm, which gradually expands to a maximum width of 3.6 mm driven by expansion potential. Next, the system updates the mesh, prioritizing cracks as seepage channels, increasing the permeability coefficient of the crack zone to 150 times that of the matrix, i.e., 1.8 × 10⁻⁶. -4 The system performed transient seepage analysis at a speed of m / s, revealing that potential seepage paths primarily extend downwards along the surface crack network, with a density of 4.2 paths per meter at the slope crest, increasing seepage flow by 28.7%. Three main channels were identified with depths of 2.1m, 3.4m, and 4.8m. Finally, the system integrated the simulation results to generate a seepage risk distribution map, automatically calculating a safety factor of 1.18. Areas below the threshold of 1.3 were marked as high-risk areas. The system also linked to an engineering database to recommend seepage prevention measures, such as laying geomembranes to reduce the infiltration rate to 35% of its original value, thus forming a complete automated analysis chain from expansion simulation to seepage path determination.

[0076] S103. If the risk value of the potential seepage path distribution in the seepage path distribution exceeds the preset threshold, the pore structure parameters are adjusted, and the range of permeability coefficient under geological conditions is predicted by a neural network to obtain the optimized permeability coefficient value.

[0077] By analyzing the correlation between seepage paths and distribution states, a judgment result is obtained as to whether the distribution state exceeds a preset threshold. Based on the judgment result, if the preset threshold is exceeded, the structural parameters are adjusted according to the correlation between pore structure and structural parameters, resulting in adjusted pore structure data. Using the adjusted pore structure data, combined with the correlation between geological conditions and permeability coefficients, a basic dataset for permeability coefficient prediction is constructed. A neural network model is used to process the basic dataset, and the optimized permeability coefficient range is determined based on the correlation between permeability coefficients and coefficient ranges. Based on the optimized permeability coefficient range data, combined with the correlation between structural parameters and optimized data, an updated adjustment strategy is obtained. Using the updated strategy, the potential changing trend of the seepage path is judged based on the correlation between the seepage path and geological conditions, resulting in the final optimized path distribution. Using the final optimized path distribution result, combined with the correlation between distribution state and adjustment strategy, the long-term stability data of the seepage path is determined.

[0078] Specifically, information technology is used to optimize and adjust the distribution of seepage paths and predict permeability coefficients. First, the system reads the path distribution data obtained from previous seepage analysis. If the density of seepage paths exceeds a preset threshold of 3.5 paths per meter, the pore structure parameter adjustment module is automatically triggered. The initial porosity is set to 0.38 and the pore connectivity to 0.62. Combined with soil particle size distribution parameters, a multilayer perceptron neural network model is imported. This model contains 6 nodes in the input layer, 32 nodes in each of the two hidden layers, and 1 node in the output layer. It uses the ReLU activation function and the Adam optimizer with a learning rate of 0.001. The training dataset consists of 200 sets of measured permeability coefficient values ​​obtained from field borehole sampling and indoor permeability tests, covering a range of 1.5 × 10⁻⁶. -7 Up to 8.9×10 -5 m / s. The system uses current geological parameters, including clay content of 42%, dry density of 1.68 g / cm³, and initial saturation of 68%, as input vectors. Through forward propagation calculations, the predicted basic value for the permeability coefficient is 4.7 × 10⁻⁶ m / s. -6 m / s, and then through sensitivity analysis, the porosity was adjusted to decrease by 0.05 to 0.33, and the connectivity was optimized to 0.55, thus converging the predicted permeability coefficient range to 2.1 × 10 m / s. -6 Up to 3.6×10 -6 m / s, average value is 3.0 × 10 -6 m / s. The optimized permeability coefficient value is automatically updated to the matrix region of the seepage model, and the steady-state seepage calculation is rerun. The seepage flow rate decreases by 21.4% compared to the initial value, the path distribution density decreases to 2.8 paths per meter, and the maximum depth of the main channels decreases to 3.7m. The system further generates a comparison cloud map before and after optimization and calculates that the stability safety factor has increased to 1.42, confirming that the risk level has decreased from high to medium. Thus, the system realizes automated optimization of permeability coefficient and seepage path control based on neural network prediction.

[0079] S104. Generate a sealing measure configuration scheme based on the optimized permeability coefficient value, and use iterative simulation to verify the degree of closure of the interface gaps to determine the effectiveness of the sealing measures.

[0080] An initial sealing configuration scheme is generated based on the optimized permeability coefficient value. The seepage process is iteratively simulated to obtain data on the closure status of interfacial gaps. Based on this data, it is determined whether the sealing configuration scheme meets the closure requirements. If the closure requirements are not met, the sealing material distribution parameters are adjusted to obtain updated sealing configuration data. Using this updated configuration data, an interfacial stress distribution dataset is constructed. Finite element analysis is used to process this dataset to determine the interfacial closure pressure value. Based on this pressure value, a sealing layer thickness adjustment scheme is generated. Using this adjusted scheme, seepage barrier stability data is predicted.

[0081] Specifically, information technology is used to configure sealing measures based on optimized permeability coefficients and verify the closure of interface gaps. The system first imports the pre-optimized permeability coefficient value of 3.0 × 10⁻⁶. -6 Using m / s as the matrix zone parameter, an automatic dam foundation sealing scheme is generated. A plastic concrete cutoff wall combined with double-row curtain grouting is selected. The cutoff wall thickness is set at 0.8m, and the depth extends to 5m below the relatively impermeable layer. The elastic modulus of the wall material is controlled at 1500MPa, and the target permeability coefficient is 1.2×10⁻⁶. - 8 The grouting curtain uses cement grout with a water-cement ratio gradually changing from 5:1 to 0.6:1, with a hole spacing of 2.0m and a row spacing of 1.2m. The grouting pressure is gradually increased to 4.5MPa. A three-dimensional finite element seepage model is constructed, with a mesh of approximately 450,000 tetrahedral elements. The boundary conditions are an upstream water head of 45m and a downstream water head of 10m. The initial permeability coefficient of the sealing zone is assigned a value of 1×10⁻⁶. -7 The initial seepage flow rate was calculated to be 0.028 m³ / s using Darcy's law. An iterative simulation was then initiated, adjusting the permeability coefficient of the sealed zone using the Newton-Raphson algorithm. The convergence criterion was that the difference in seepage flow rate between adjacent iterations should be less than 0.5%. After 8 iterations, the permeability coefficient of the sealed zone converged to 8.5 × 10⁻⁶ m³ / s. -9 The hydraulic gradient at the interface joint decreased to 1.8 cm / s, achieving a closure rate of 96.4%. The total seepage flow decreased to 0.0042 m³ / s, a reduction of 85% from the initial value. The system automatically outputs an isohyetal distribution map and seepage vector field, confirming that the interface joint closure meets the seepage prevention standard of less than 10 cm / s. -8 cm / s, thereby enabling automated configuration and effectiveness verification of sealing measures.

[0082] S105. Obtain the effectiveness data of the verified sealing measures, and combine the groundwater level simulation scenario to update the arrangement of the steel reinforcement connection nodes through dynamic adjustment algorithm to obtain the preliminary waterproof sealing design.

[0083] After obtaining verified sealing measure data, a corresponding scenario dataset is constructed for simulated groundwater level changes, obtaining an initial mapping relationship between the scenario and the sealing measures. Based on the initial mapping relationship and groundwater level data, the stress distribution of the rebar connection nodes is analyzed using a preset threshold range to determine the initial adjustment direction of the node layout. Using this initial adjustment direction, a dynamic adjustment strategy is implemented based on the stress distribution data of the rebar connection nodes to update the node layout structure, resulting in optimized layout configuration information. Using the optimized layout configuration information and the design requirements for waterproof sealing, a corresponding sealing layer distribution model is generated to determine whether the sealing layer distribution meets the preset coverage standard. If not, the stress distribution data is reprocessed. Based on the sealing layer distribution model, seepage barrier data in the groundwater level simulation scenario is obtained, and the stability performance of seepage barrier is analyzed to determine the applicability assessment results of the waterproof sealing scheme. Based on the applicability assessment results and the stability data of the waterproof sealing scheme, a final preliminary scheme dataset is constructed to obtain waterproof sealing configuration schemes suitable for different simulation scenarios.

[0084] Specifically, the system utilizes information technology to acquire data on the effectiveness of verified sealing measures and dynamically optimize the arrangement of rebar connection nodes. The system first extracts the verified permeability coefficient of the sealing zone (9.2 × 10⁻⁶) from the preceding seepage simulation module. - 9Using cm / s and a total seepage flow of 0.0038 m³ / s as effective baseline data, and combining multiple scenarios of groundwater level fluctuations, a time-varying boundary sequence with upstream head variations of 35–52 m and downstream head variations of 8–15 m was automatically loaded to construct a transient seepage-stress coupled finite element model. The mesh used approximately 620,000 hexahedral dominant hybrid elements, and the Drucker-Prager criterion was used to describe the nonlinear behavior of reinforced concrete. The system initiated a dynamic simulation of groundwater level, generating 150 sets of random water level fluctuation paths based on the Monte Carlo method. The tensile and shear stress distributions at the rebar connection nodes under each path were calculated. The initial node arrangement used HRB400 grade rebar with a spacing of 1.5 m and a diameter of 28 mm, and the initial maximum tensile stress reached 312 MPa. Next, the dynamic adjustment algorithm is used, employing a genetic algorithm combined with a finite element sub-model for iterative optimization. The population size is set to 80, the crossover rate to 0.75, the mutation rate to 0.12, and the maximum number of iterations to 120. Optimization variables include node spacing, rebar diameter, and anchorage length. The objective function is to minimize the maximum equivalent stress at the nodes while constraining the probability of waterproof seal failure to be less than 0.5%. After 68 iterations and convergence, the optimized layout scheme is obtained: the node spacing is adjusted to 1.2m, the rebar diameter is increased to 32mm, and the anchorage length is extended to 45 times the diameter. After optimization, the maximum tensile stress at the nodes is reduced to 218MPa, the shear stress is reduced to 96MPa, and the overall stress exceedance rate of the waterproof seal area is reduced to 2.3%. The system automatically outputs the node stress cloud map and displacement vector field, generates a preliminary waterproof seal design report, and pushes it to the subsequent construction drawing module, realizing fully automated optimization and data closure throughout the process.

[0085] S106. Evaluate the seepage channel blocking effect based on the preliminary waterproof sealing design. If the blocking effect is lower than the preset threshold, incorporate the expansion characteristic feedback loop to obtain an enhanced sealing measure solution by recalculating the influence of mineral composition.

[0086] Obtain seepage channel data corresponding to the preliminary waterproof sealing design. Calculate the blocking performance value based on the seepage channel data. Determine the relationship between the blocking performance value and a preset threshold. If the blocking performance value is lower than the preset threshold, initiate an expansion characteristic feedback loop. Extract mineral composition parameters from the expansion characteristic feedback loop. Recalculate the action intensity distribution based on the mineral composition parameters. Update the sealing measure configuration based on the action intensity distribution to obtain an enhanced sealing measure scheme. Regenerate seepage channel blocking data using the enhanced sealing measure scheme.

[0087] Specifically, the system uses information technology to evaluate the effectiveness of the initial waterproof sealing design in blocking seepage channels. First, key data is extracted from the seepage analysis module, calculating the average seepage rate of the current sealing area to be 0.0025 m³ / h. This is compared with the preset threshold of 0.0018 m³ / h, revealing that the blocking effect is insufficient. Next, the system automatically initiates an expansion characteristic feedback loop mechanism, calling upon expansion performance parameters of the sealing material from the material database. Initially, bentonite accounts for 22% of the mineral composition. Using a finite element analysis model, the volume growth rate of a material with an expansion coefficient of 0.15 under humidity changes is simulated. The calculation results show that the volume growth rate is only 8.7%, insufficient to effectively fill the seepage channels. Subsequently, the system adjusts the mineral composition ratio using an iterative algorithm, optimizing the bentonite ratio using a gradient descent method. The step size is set to 0.02, with a maximum of 50 iterations, aiming for a volume growth rate of over 12%. After 32 iterations, the bentonite ratio is adjusted to 28%, and the volume growth rate increases to 12.3%. Simultaneously, the system analyzed the impact of mineral composition adjustment on permeability, recalculated the permeability coefficient, and obtained an adjusted value of 7.5 × 10⁻⁶. -10 The expansion rate decreased by approximately 18% from the initial value (cm / s). To ensure the feasibility of the solution, the system further incorporated environmental humidity fluctuation data to simulate the expansion stability within a humidity range of 60% to 85%. The analysis results showed that the material's expansion rate fluctuated by less than 1.2% within this range, meeting the long-term sealing requirements. Finally, the system generated an enhanced sealing measure scheme, integrating the adjusted mineral composition ratio and permeability parameters into the design basis, automatically outputting an optimized material ratio report, and pushing it to the subsequent verification module to form a closed-loop data processing.

[0088] S107. By integrating geological condition variables through enhanced sealing measures, the optimal results of waterproof sealing of steel reinforcement connections that are suitable for diverse geological conditions are finally determined using finite element analysis.

[0089] The process begins by acquiring basic data for sealing measures. Considering the characteristics of the rebar connection structure, corresponding environmental parameters are extracted from a pre-established geological condition database to obtain preliminary environmental adaptability data. Based on this preliminary data, the finite element method is used to simulate and analyze the waterproofing performance of the rebar connection, determining the stress distribution under different geological conditions. If the stress distribution exceeds a preset threshold, the excess geological condition data is categorized to obtain the connection structure parameters that need adjustment. Using the adjusted connection structure parameters, combined with the sealing measures, an optimized configuration scheme is regenerated to determine the direction for improving waterproofing performance suitable for specific geological conditions. The simulation data for the rebar connection is updated using this improvement direction to obtain new environmental adaptability assessment results, determining whether they meet the preset performance evaluation standards. Based on the updated environmental adaptability assessment results, corresponding geological condition variables are extracted for rebar connections that do not meet the standards, resulting in the final optimized configuration adjustment scheme. This final optimized configuration adjustment scheme is then used to locally update the sealing measures, determining a waterproofing performance configuration for rebar connections suitable for various geological environments.

[0090] Specifically, the system integrates enhanced sealing measures using information technology. First, it accesses a geological database to extract geological condition variables for the target area, including parameters such as rock porosity, groundwater pressure, and soil compressibility. Assuming a porosity of 0.18, groundwater pressure of 0.25 MPa, and soil compressibility coefficient of 0.03, the system inputs this geological data into a finite element analysis model to construct a three-dimensional geological environment simulation. With a mesh density of 1000 elements per cubic meter, the system calculates the stress distribution at the steel reinforcement connection points, obtaining a maximum stress value of 1.8 MPa. Combining this with a database of the compressive strength of waterproof sealing materials, it selects material options with a compressive strength of not less than 2.0 MPa. Subsequently, the system conducted an adaptability analysis for diverse geological conditions. Using a Monte Carlo simulation algorithm, random variables were set to a porosity range of 0.15 to 0.22 and groundwater pressure range of 0.20 to 0.30 MPa. 10,000 simulations were run to evaluate the stability of the sealing scheme under different conditions. The analysis results showed that the scheme exhibited uniform stress distribution in over 90% of scenarios, with a local leakage risk below 0.05. Next, based on the analysis results, the system optimized the rebar connection design and adjusted the sealing layer thickness at the connection points. The initial thickness was 5.0 mm, and through iterative calculations, the optimal thickness was determined to be 6.2 mm. Simultaneously, the material mix parameters were updated to ensure a matching degree of over 95% with the geological conditions. Finally, the system automatically generated an optimized waterproofing and sealing result report, integrating all calculated data and simulation results into the design basis and pushing it to relevant modules for subsequent verification, forming a complete data processing chain to ensure the applicability of the scheme in diverse geological environments.

[0091] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A big data-driven optimization method for foundation rebar connections based on neural networks, characterized in that, include: Mineral composition data and swelling characteristic indicators were obtained by collecting soil samples, and the soil swelling potential value was obtained by X-ray diffraction analysis. Based on the soil swelling potential value, the interface gap width change was simulated, and a seepage channel model was constructed using finite element analysis to determine the potential seepage path distribution. If the risk value of the potential seepage path distribution in the seepage path distribution exceeds the preset threshold, the pore structure parameters are adjusted, and the range of permeability coefficient under geological conditions is predicted by a neural network to obtain the optimized permeability coefficient value. A sealing measure configuration scheme is generated based on the optimized permeability coefficient value, and the effectiveness of the sealing measure is judged by iterative simulation to verify the degree of closure of the interface gap. After obtaining the effectiveness data of the verified sealing measures, and combining the groundwater level simulation scenario, the arrangement of the steel connection nodes is updated by dynamically adjusting the algorithm to obtain the preliminary waterproof sealing design. The effectiveness of the seepage channel blocking is evaluated based on the preliminary waterproof sealing design. If the blocking effect is lower than the preset threshold, an enhanced sealing measure solution is obtained by recalculating the impact of mineral composition through an expansion characteristic feedback loop. By integrating geological condition variables through enhanced sealing measures, finite element analysis was used to finally determine the optimized results of waterproof sealing for steel reinforcement connections that are suitable for diverse geological conditions.

2. The method for optimizing foundation reinforcement connection based on neural networks using big data as described in claim 1, characterized in that, The process of acquiring mineral composition data and swelling characteristic indicators through soil sample collection, and then processing the mineral composition data and swelling characteristic indicators using X-ray diffraction analysis to obtain the soil swelling potential value includes: Soil samples and mineral composition data, as well as swelling characteristics, were collected through sample acquisition. The diffraction pattern was obtained by processing the mineral composition data using X-ray diffraction analysis. Mineral composition is determined by obtaining mineral types and relative contents from diffraction patterns; The proportion of swelling-active minerals is determined based on the mineral composition and swelling characteristics. If the proportion of swelling-active minerals is higher than the preset threshold, it is marked as high swelling potential; otherwise, it is marked as low swelling potential. The initial expansion potential value is obtained by calculating the weighted expansion index by combining expansion characteristic indicators with the proportion of expansion-active minerals. The random forest algorithm is used to process the initial expansion potential value and mineral composition to determine the corrected expansion potential value; Based on the modified expansion potential value and expansion characteristic index, the support vector machine algorithm is used to classify the soil expansion level to obtain the expansion potential classification result.

3. The method for optimizing foundation reinforcement connections based on neural networks using big data as described in claim 1, characterized in that, The process of simulating interface gap width changes based on soil swelling potential values ​​and constructing a seepage channel model using finite element analysis to determine the potential seepage path distribution includes: A finite element mesh model was established based on the soil expansion potential value to simulate the change in interface gap width. Obtain gap width distribution data from the change in interface gap width; The initial seepage path is determined by constructing a seepage channel model using the seepage equation; The seepage intensity distribution is obtained by calculating the seepage velocity field through the initial seepage path; High-risk seepage areas are identified based on the distribution of seepage intensity and the distribution of crack width. If the proportion of high-risk seepage areas is higher than a preset threshold, then high-risk seepage paths are marked; otherwise, low-risk seepage paths are marked. The support vector machine algorithm is used to process the distribution of high seepage risk paths and seepage intensity to classify the potential seepage path distribution and obtain the final seepage risk level.

4. The method for optimizing foundation reinforcement connection based on neural networks using big data as described in claim 1, characterized in that, If the risk value of potential seepage paths in the seepage path distribution exceeds a preset threshold, the pore structure parameters are adjusted, and an optimized permeability coefficient value is obtained by predicting the range of permeability coefficients under geological conditions using a neural network, including: By analyzing the correlation between seepage path and distribution state, the judgment result of whether the distribution state exceeds the preset threshold is obtained; Based on the judgment result of the distribution state, if it exceeds the preset threshold, the structural parameters are adjusted according to the correlation between the pore structure and the structural parameters to obtain the adjusted pore structure data. Using adjusted pore structure data and combining the correlation between geological conditions and permeability coefficient, a basic dataset for permeability coefficient prediction is constructed. By processing the basic dataset through a neural network model, and considering the correlation between the permeability coefficient and its range, the optimized permeability coefficient range data is determined. Based on the optimized permeability coefficient range data, and combined with the correlation between structural parameters and optimized data, an updated adjustment strategy is obtained. By updating the scheme, the potential changing trend of the seepage path is determined based on the correlation between the seepage path and geological conditions, and the final path distribution optimization result is obtained. By using the final path distribution optimization results and combining the correlation between the distribution status and the adjustment strategy, the long-term stability data of the seepage path are determined.

5. The method for optimizing foundation reinforcement connections based on neural networks using big data as described in claim 1, characterized in that, The process of generating a sealing measure configuration scheme based on the optimized permeability coefficient value, and using iterative simulation to verify the degree of closure of the interface gaps to determine the effectiveness of the sealing measures, includes: Based on the optimized permeability coefficient value, an initial sealing measure configuration scheme is generated; Data on the closure of interfacial gaps were obtained by iteratively simulating the seepage process. Based on the data on the closure of the interface gaps, determine whether the sealing measures configuration scheme meets the closure requirements; If the sealing configuration scheme does not meet the closure requirements, the sealing material distribution parameters are adjusted to obtain updated sealing configuration data; Using updated sealing measures and configuration data, an interface stress distribution dataset is constructed. The interface stress distribution dataset was processed by finite element analysis to determine the interface closure pressure value. Based on the interface closure pressure value, a sealing layer thickness adjustment scheme is generated; A sealing layer thickness adjustment scheme was adopted to predict seepage barrier stability data.

6. The method for optimizing foundation reinforcement connections based on neural networks using big data as described in claim 1, characterized in that, The process of obtaining and verifying the effectiveness data of the sealing measures, and then updating the arrangement of the rebar connection nodes through a dynamic adjustment algorithm based on the groundwater level simulation scenario, yields a preliminary waterproof sealing design, including: After obtaining the verified sealing measures data, construct the corresponding scenario dataset for the simulated scenario of groundwater level change, and obtain the initial mapping relationship between the scenario and the sealing measures. Based on the initial mapping relationship and combined with groundwater level data, the stress distribution of the steel reinforcement connection nodes is analyzed using a preset threshold range to determine the initial adjustment direction of the node layout. By initially adjusting the direction, a dynamic adjustment strategy is implemented based on the stress distribution data of the rebar connection nodes to update the node layout structure and obtain optimized layout configuration information. Using the optimized layout configuration information and combined with the design requirements for waterproof sealing, a corresponding sealing layer distribution model is generated. It is then determined whether the sealing layer distribution meets the preset coverage standard. If not, the stress distribution data is reprocessed. Based on the sealing layer distribution model, we obtained the seepage barrier data in the groundwater level simulation scenario, analyzed the stability performance of the seepage barrier, and determined the applicability assessment results of the waterproof sealing scheme. Based on the applicability assessment results and the stability data of the waterproof sealing scheme, a final preliminary scheme dataset is constructed to obtain waterproof sealing configuration schemes suitable for different simulation scenarios.

7. The method for optimizing foundation reinforcement connection based on neural networks using big data as described in claim 1, characterized in that, The process involves evaluating the seepage channel blocking effect based on the initial waterproof sealing design. If the blocking effect is lower than a preset threshold, an enhanced sealing measure scheme is obtained by recalculating the impact of mineral composition through an expansion characteristic feedback loop. This scheme includes: Obtain seepage channel data corresponding to the preliminary waterproof sealing design; Calculate the blocking performance value based on the seepage channel data; Determine the relationship between the blocking performance value and the preset threshold. If the blocking performance value is lower than the preset threshold, start the expansion characteristic feedback loop. Mineral composition parameters are extracted from the expansion characteristic feedback loop; The intensity distribution of the action was recalculated based on the mineral composition parameters; By updating the sealing measure configuration through the distribution of action intensity, an enhanced sealing measure scheme is obtained; The seepage channel blocking data was regenerated using an enhanced sealing solution.

8. The method for optimizing foundation reinforcement connections based on neural networks using big data as described in claim 1, characterized in that, The enhanced sealing measures integrate geological condition variables, and the finite element analysis is used to finally determine the optimized results of the waterproof sealing of steel reinforcement connections to adapt to diverse geological conditions, including: Obtain basic data for sealing measures; and extract corresponding environmental parameters from a pre-established geological condition database to obtain preliminary environmental adaptability data, taking into account the characteristics of the steel reinforcement connection structure. Based on preliminary environmental adaptability data, the waterproof performance of the steel reinforcement connection was simulated and analyzed using the finite element method to determine the stress distribution of the connection structure under different geological conditions. If the stress distribution exceeds the preset threshold range, the geological condition data of the excess portion will be classified and processed to obtain the connection structure parameters that need to be adjusted. By adjusting the connection structure parameters and combining the sealing measures, an optimized configuration scheme is regenerated to determine the direction for improving waterproof performance suitable for specific geological conditions. The simulation data of the rebar connection was updated using the improved approach to obtain new environmental adaptability assessment results and determine whether they meet the preset performance assessment standards. Based on the updated environmental adaptability assessment results, for the substandard rebar connection parts, the corresponding geological condition variables are extracted to obtain the final optimized configuration adjustment scheme; Through the final optimized configuration adjustment scheme, the sealing measures were partially updated to determine the waterproof performance configuration of the steel reinforcement connection that is suitable for various geological environments.

Images