Analysis method for bearing capacity of long-short pile composite foundation based on finite element analysis
By using Optistruct finite element analysis and autoencoder self-attention network, a prediction model for the ultimate bearing capacity of composite foundations with long and short piles was established. This model addresses the shortcomings of bearing capacity analysis for composite foundations with long and short piles, enables bearing capacity optimization and risk assessment, and improves the scientific and economic efficiency of engineering design.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA POWER CONSTR FOURTEENTH BUREAU URBAN CONSTR INVESTMENT CO LTD
- Filing Date
- 2025-08-29
- Publication Date
- 2026-06-12
AI Technical Summary
The lack of effective methods for analyzing the bearing capacity of composite foundations with long and short piles in existing technologies leads to a lack of theoretical support and a mismatch between the foundation treatment design and actual needs.
The Optistruct finite element analysis software was used to model the composite foundation of long and short piles. Through meshing and settlement result analysis, combined with autoencoders and self-attention networks, a prediction model of the ultimate bearing capacity of the composite foundation of long and short piles was established to realize bearing capacity analysis and optimization design.
It enables automated analysis of the bearing capacity of composite foundations with long and short piles, optimizes the pile layout, improves bearing efficiency and space utilization, provides engineering design basis, assesses risks and supports decision-making, and ensures project quality and economic benefits.
Smart Images
Figure CN121093695B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of composite foundation technology for long and short piles, and more specifically, to a method for analyzing the bearing capacity of composite foundations for long and short piles based on finite element analysis. Background Technology
[0002] When the bearing capacity or deformation of the foundation cannot meet the design requirements, foundation treatment is required. Composite foundation schemes are widely used in foundation treatment, the most typical of which is the composite foundation of long and short piles.
[0003] Composite foundations with long and short piles are a novel foundation treatment technology. They utilize piles of varying lengths arranged according to specific rules to improve the bearing capacity of the foundation and reduce settlement. This technology has demonstrated significant technical and economic benefits in engineering practice, particularly in reducing foundation settlement and harmonizing the relationship between bearing capacity and deformation. The ultimate bearing capacity of the composite foundation, serving as a crucial basis and theoretical support for the design of long and short pile arrangements, is therefore of paramount importance.
[0004] Therefore, there is an urgent need for a bearing capacity analysis method for composite foundations with long and short piles based on finite element analysis to meet the actual needs of the layout design of composite foundations with long and short piles. Summary of the Invention
[0005] This application provides a method for analyzing the bearing capacity of composite foundations with long and short piles based on finite element analysis to solve the above-mentioned technical problems.
[0006] This application provides a method for analyzing the bearing capacity of composite foundations with long and short piles based on finite element analysis, including:
[0007] Using Optistruct finite element analysis software, the piles, soil layers, and cushion layer in the long and short pile composite foundation are modeled to obtain the composite foundation model; wherein, the composite foundation map of the long and short pile composite foundation in the composite foundation model is meshed.
[0008] Based on the first length information of the long piles, the second length information of the short piles in the composite foundation model, and the arrangement of the long piles and the short piles, the long and short pile distance adjacency matrix is determined.
[0009] Vertical stress is simulated on the composite foundation of long and short piles to obtain the settlement results of the composite foundation of long and short piles; wherein, the settlement results of the composite foundation of long and short piles include the settlement information of each grid in the composite foundation map;
[0010] The settlement information of each grid in the composite foundation map is vectorized to obtain the foundation settlement feature matrix;
[0011] The bearing capacity of the composite foundation of long and short piles is analyzed based on the adjacency matrix of the long and short piles and the foundation settlement characteristic matrix.
[0012] Furthermore, after simulating the application of vertical stress on the composite foundation of long and short piles to obtain the settlement results of the composite foundation, the method further includes:
[0013] Obtain the first load information and the first pile top reaction force of the long pile, and obtain the second load information and the second pile top reaction force of the short pile;
[0014] The first load information, the first pile top reaction force, the second load information, and the second pile top reaction force are vectorized to obtain the stress bearing matrix of long and short piles.
[0015] Furthermore, after simulating the application of vertical stress on the composite foundation of long and short piles to obtain the settlement results of the composite foundation, the method further includes:
[0016] Obtain the soil layer thickness, groundwater conditions, and porosity information of the foundation corresponding to each grid in the composite foundation map;
[0017] The soil layer thickness, groundwater conditions, and void ratio information of the foundation area corresponding to each grid in the composite foundation map are vectorized to obtain the composite foundation stress environment parameter matrix.
[0018] Further, the step of analyzing the bearing capacity of the composite foundation of long and short piles based on the long and short pile distance adjacency matrix and the foundation settlement characteristic matrix includes:
[0019] The bearing capacity of the composite foundation of long and short piles is analyzed based on the adjacency matrix of the long and short piles, the foundation settlement characteristic matrix, the stress bearing capacity matrix of the long and short piles, and the stress environment parameter matrix of the composite foundation.
[0020] Furthermore, after analyzing the bearing capacity of the composite foundation of long and short piles based on the adjacency matrix of the long and short piles, the foundation settlement characteristic matrix, the stress bearing capacity matrix of the long and short piles, and the stress environment parameter matrix of the composite foundation, the method further includes:
[0021] Dimensionality reduction of the long and short pile distance adjacency matrix, the foundation settlement feature matrix, the long and short pile stress bearing matrix, and the composite foundation stress environment parameter matrix includes: inputting the long and short pile distance adjacency matrix, the foundation settlement feature matrix, the long and short pile stress bearing matrix, and the composite foundation stress environment parameter matrix into a preset autoencoder; wherein, the autoencoder uses a feature-based distance metric learning method for data conversion, and the autoencoder reduces the dimensionality of the long and short pile distance adjacency matrix, the foundation settlement feature matrix, the long and short pile stress bearing matrix, and the composite foundation stress environment parameter matrix based on the learned distance metric error;
[0022] The reduced long and short pile distance adjacency matrix, the reduced foundation settlement characteristic matrix, and the reduced long and short pile stress bearing matrix are fused to obtain the first fused matrix;
[0023] The reduced composite foundation stress environment parameters and the reduced foundation settlement characteristic matrix are fused to obtain the second fusion matrix;
[0024] Using the first fusion matrix and the second fusion matrix as input sequence information, a prediction model for the ultimate bearing capacity of composite foundations with long and short piles is established.
[0025] The ultimate bearing capacity prediction model for the long-short pile composite foundation is used to predict the ultimate bearing capacity of the long-short pile composite foundation based on the bearing capacity analysis results. The bearing capacity analysis results include the analysis results of the bearing capacity of the long-short pile composite foundation after simulating the application of vertical stress on the long-short pile composite foundation.
[0026] Furthermore, the prediction model for the ultimate bearing capacity of the composite foundation of long and short piles includes a self-attention network, which includes multiple memory units. These memory units are used to determine the correlation between the bearing capacity of the foundation area corresponding to each grid in the composite foundation map and the location corresponding to each grid in the composite foundation map. The step of establishing the prediction model for the ultimate bearing capacity of the composite foundation of long and short piles using the first fusion matrix and the second fusion matrix as input sequence information includes:
[0027] The input sequence information is input into the feature extraction model, and the initial output matrix corresponding to the input sequence information is determined based on a preset sharing strategy; wherein, the preset sharing strategy is used to instruct the neurons of the feature extraction model to share information and update collaboratively.
[0028] Based on the multiple memory units included in the attention network, the initial output matrix is updated to obtain the updated memory output matrix;
[0029]
[0030] Among them, Output update The updated memory output matrix; Output origi The initial output matrix; WEI mem The set of stress bearing ratios between the foundation region corresponding to each grid and each long or short pile in the composite foundation map stored in the multiple memory units; Distance is the distance adjacency matrix of the long and short piles;
[0031]
[0032] Where N is the number of memory units; Stress_GRID i Stress_STAKE represents the sum of the stress bearing values of the foundation regions corresponding to each grid in the composite foundation map stored in the i-th memory unit. i This represents the sum of the stress bearing values of each long or short pile stored in the i-th memory unit;
[0033] The prediction model for the ultimate bearing capacity of the composite foundation of long and short piles is trained based on the updated memory output matrix, the first fusion matrix, and the second fusion matrix until the loss function converges.
[0034] Furthermore, the loss function is:
[0035]
[0036] Where T represents the number of grids in the composite ground map; M represents the sum of the number of long stakes and short stakes; Distance represents the distance adjacency matrix of the long and short stakes; FEA_Sedime k FEA_Stake represents the foundation settlement characteristics of the foundation region corresponding to the k-th grid in the composite foundation map. j The stress bearing characteristics corresponding to the j-th long pile or short pile; Output update The updated memory output matrix; FEA_Mix1 represents the first fusion matrix; FEA_Mix2 represents the second fusion matrix.
[0037] Furthermore, after modeling the piles, soil layers, and cushion layer in the composite foundation using Optistruct finite element analysis software to obtain the composite foundation model, the method further includes:
[0038] The composite foundation model is adjusted by modifying the design parameters of the long and short pile composite foundation using the Optistruct finite element analysis software. The design parameters of the long and short pile composite foundation include the pile length, pile diameter, pile spacing, and area replacement ratio of the long and short piles.
[0039] Furthermore, the long piles include CFG piles or reinforced concrete piles; the short piles include mixing piles, crushed stone piles or lime piles.
[0040] Furthermore, the material properties of the pile, the soil layer, and the cushion layer include unit weight, shear strength index, elastic modulus, and Poisson's ratio; wherein, the shear strength index includes cohesion, friction angle, modulus, and Poisson's ratio.
[0041] Based on the embodiments provided in this application, the piles, soil layers, and cushion layers in a long-short pile composite foundation are modeled using Optistruct finite element analysis software to obtain a composite foundation model. The composite foundation map of the long-short pile composite foundation in the composite foundation model is meshed. Based on the first length information of the long piles, the second length information of the short piles, and the arrangement of the long and short piles in the composite foundation model, the distance adjacency matrix of the long and short piles is determined. Vertical stress is simulated and applied to the long-short pile composite foundation to obtain the settlement results. The settlement results of the long-short pile composite foundation include the settlement information of each grid in the composite foundation map. The settlement information of each grid in the composite foundation map is vectorized to obtain the foundation settlement characteristic matrix. Based on the long-short pile distance adjacency matrix and the foundation settlement characteristic matrix, the bearing capacity of the long-short pile composite foundation is analyzed. This achieves automated analysis of the bearing capacity of the long-short pile composite foundation. Specifically, it offers the following benefits: Detailed model construction: Optistruct accurately models long and short piles, soil layers, and cushion layers, ensuring the model truly reflects the physical properties and mechanical behavior of the composite foundation. Meshable composite foundation map: Mesh processing improves model resolution, allowing for more detailed and accurate stress and settlement analysis. Optimized long and short pile arrangement: By determining the long and short pile distance adjacency matrix, the pile arrangement can be optimized, improving the bearing efficiency and space utilization of the composite foundation. Vertical stress simulation: Simulating the application of vertical stress allows for prediction of the composite foundation's response under actual loads, providing crucial information for engineering design. Settlement result analysis: The acquired settlement results include settlement information from each grid, helping to identify potential uneven settlement areas and providing data support for subsequent design adjustments. Vectorization of settlement information: The foundation settlement characteristic matrix obtained by vectorizing settlement information facilitates mathematical and statistical analysis, improving the efficiency and depth of data processing. In-depth analysis of bearing capacity: Combining the long and short pile distance adjacency matrix and the foundation settlement characteristic matrix... It can provide a more comprehensive analysis of the bearing capacity of composite foundations, assess the feasibility and safety of different design schemes; design optimization and verification: the analysis results can be used to optimize design parameters, such as pile length, diameter, and arrangement density, and verify the effectiveness of design changes through models; risk assessment and management: by predicting bearing capacity and settlement, engineering risks can be assessed, and timely measures can be taken to avoid or reduce potential engineering problems; cost-benefit analysis: accurate bearing capacity and settlement analysis helps to conduct cost-benefit analysis, ensuring that economic benefits are maximized while meeting safety requirements; improve engineering quality and reliability: through finite element analysis, the scientific nature and reliability of engineering design can be improved, the service life of the project can be extended, and the later maintenance costs can be reduced; support decision-making: it provides quantitative analysis results for engineering decision-makers, supports data-based decision-making, and increases the transparency and credibility of decisions. Attached Figure Description
[0042] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:
[0043] Figure 1 This is a flowchart of an optional finite element analysis-based method for analyzing the bearing capacity of a composite foundation of long and short piles according to an embodiment of this application;
[0044] Figure 2 This is a flowchart of another optional method for analyzing the bearing capacity of a composite foundation of long and short piles based on finite element analysis, according to an embodiment of this application.
[0045] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation
[0046] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.
[0047] Optionally, such as Figure 1 As shown, this application provides a method for analyzing the bearing capacity of composite foundations with long and short piles based on finite element analysis, including:
[0048] S101. Using Optistruct finite element analysis software, the piles, soil layers, and cushion layers in the long and short pile composite foundation are modeled to obtain the composite foundation model; among them, the composite foundation map of the long and short pile composite foundation in the composite foundation model is meshed.
[0049] S102, Based on the first length information of the long piles, the second length information of the short piles, and the arrangement of the long and short piles in the composite foundation model, determine the long and short pile distance adjacency matrix.
[0050] S103, simulate the application of vertical stress on the composite foundation of long and short piles to obtain the settlement results of the composite foundation of long and short piles; wherein, the settlement results of the composite foundation of long and short piles include the settlement information of each grid on the composite foundation map;
[0051] S104, the settlement information of each grid in the composite foundation map is vectorized to obtain the foundation settlement feature matrix;
[0052] S105. Based on the adjacency matrix of long and short piles and the foundation settlement characteristic matrix, the bearing capacity of the composite foundation of long and short piles is analyzed.
[0053] Based on the embodiments provided in this application, the piles, soil layers, and cushion layers in a long-short pile composite foundation are modeled using Optistruct finite element analysis software to obtain a composite foundation model. The composite foundation map of the long-short pile composite foundation in the composite foundation model is meshed. Based on the first length information of the long piles, the second length information of the short piles, and the arrangement of the long and short piles in the composite foundation model, the distance adjacency matrix of the long and short piles is determined. Vertical stress is simulated and applied to the long-short pile composite foundation to obtain the settlement results. The settlement results of the long-short pile composite foundation include the settlement information of each grid in the composite foundation map. The settlement information of each grid in the composite foundation map is vectorized to obtain the foundation settlement characteristic matrix. Based on the long-short pile distance adjacency matrix and the foundation settlement characteristic matrix, the bearing capacity of the long-short pile composite foundation is analyzed. This achieves automated analysis of the bearing capacity of the long-short pile composite foundation. Specifically, it offers the following benefits: Detailed model construction: Optistruct accurately models long and short piles, soil layers, and cushion layers, ensuring the model truly reflects the physical properties and mechanical behavior of the composite foundation. Meshable composite foundation map: Mesh processing improves model resolution, allowing for more detailed and accurate stress and settlement analysis. Optimized long and short pile arrangement: By determining the long and short pile distance adjacency matrix, the pile arrangement can be optimized, improving the bearing efficiency and space utilization of the composite foundation. Vertical stress simulation: Simulating the application of vertical stress allows for prediction of the composite foundation's response under actual loads, providing crucial information for engineering design. Settlement result analysis: The acquired settlement results include settlement information from each grid, helping to identify potential uneven settlement areas and providing data support for subsequent design adjustments. Vectorization of settlement information: The foundation settlement characteristic matrix obtained by vectorizing settlement information facilitates mathematical and statistical analysis, improving the efficiency and depth of data processing. In-depth analysis of bearing capacity: Combining the long and short pile distance adjacency matrix and the foundation settlement characteristic matrix... It can provide a more comprehensive analysis of the bearing capacity of composite foundations, assess the feasibility and safety of different design schemes; design optimization and verification: the analysis results can be used to optimize design parameters, such as pile length, diameter, and arrangement density, and verify the effectiveness of design changes through models; risk assessment and management: by predicting bearing capacity and settlement, engineering risks can be assessed, and timely measures can be taken to avoid or reduce potential engineering problems; cost-benefit analysis: accurate bearing capacity and settlement analysis helps to conduct cost-benefit analysis, ensuring that economic benefits are maximized while meeting safety requirements; improve engineering quality and reliability: through finite element analysis, the scientific nature and reliability of engineering design can be improved, the service life of the project can be extended, and the later maintenance costs can be reduced; support decision-making: it provides quantitative analysis results for engineering decision-makers, supports data-based decision-making, and increases the transparency and credibility of decisions.
[0054] like Figure 2 As shown, further, after simulating the application of vertical stress on the long-short pile composite foundation to obtain the settlement results of the long-short pile composite foundation, the method also includes:
[0055] S201, obtain the first load information and the first pile top reaction force of the long pile, and obtain the second load information and the second pile top reaction force of the short pile;
[0056] S202, the first load information, the first pile top reaction force, the second load information, and the second pile top reaction force are vectorized to obtain the stress bearing matrix of long and short piles.
[0057] Furthermore, after simulating the application of vertical stress on the composite foundation of long and short piles to obtain the settlement results of the composite foundation, the method also includes:
[0058] Obtain information on soil layer thickness, groundwater conditions, and void ratio for each grid in the composite foundation map;
[0059] The soil layer thickness, groundwater conditions, and void ratio information of each grid in the composite foundation map are vectorized to obtain the composite foundation stress environment parameter matrix.
[0060] Furthermore, based on the adjacency matrix of long and short piles and the foundation settlement characteristic matrix, the bearing capacity of the composite foundation of long and short piles is analyzed, including:
[0061] The bearing capacity of the composite foundation of long and short piles is analyzed based on the adjacency matrix of long and short piles, the foundation settlement characteristic matrix, the stress bearing capacity matrix of long and short piles, and the stress environment parameter matrix of the composite foundation.
[0062] Furthermore, after analyzing the bearing capacity of the composite foundation of long and short piles based on the adjacency matrix of long and short piles, the foundation settlement characteristic matrix, the stress bearing capacity matrix of long and short piles, and the stress environment parameter matrix of the composite foundation, the method also includes:
[0063] Dimensionality reduction is performed on the long and short pile distance adjacency matrix, foundation settlement feature matrix, long and short pile stress bearing matrix, and composite foundation stress environment parameter matrix. This includes: inputting the long and short pile distance adjacency matrix, foundation settlement feature matrix, long and short pile stress bearing matrix, and composite foundation stress environment parameter matrix into a preset autoencoder; wherein, the autoencoder uses a feature-based distance metric learning method for data transformation, and the autoencoder reduces the dimension of the long and short pile distance adjacency matrix, foundation settlement feature matrix, long and short pile stress bearing matrix, and composite foundation stress environment parameter matrix based on the learned distance metric error;
[0064] The reduced long and short pile distance adjacency matrix, the reduced foundation settlement characteristic matrix, and the reduced long and short pile stress bearing matrix are fused to obtain the first fused matrix;
[0065] The reduced composite foundation stress environment parameters and the reduced foundation settlement characteristic matrix are fused to obtain the second fusion matrix;
[0066] Using the first and second fusion matrices as input sequence information, a prediction model for the ultimate bearing capacity of composite foundations with long and short piles is established.
[0067] The ultimate bearing capacity prediction model for long and short pile composite foundations is used to predict the ultimate bearing capacity of long and short pile composite foundations based on the bearing capacity analysis results. The bearing capacity analysis results include the analysis results of the bearing capacity of long and short pile composite foundations after simulating the application of vertical stress on the foundation.
[0068] Furthermore, the prediction model for the ultimate bearing capacity of the composite foundation of long and short piles includes a self-attention network, which comprises multiple memory units. These memory units are used to determine the correlation between the bearing capacity of the foundation region corresponding to each grid in the composite foundation map and the location corresponding to each grid in the composite foundation map. Using the first fusion matrix and the second fusion matrix as input sequence information, the prediction model for the ultimate bearing capacity of the composite foundation of long and short piles is established, including:
[0069] The input sequence information is fed into the feature extraction model, and the initial output matrix corresponding to the input sequence information is determined based on a preset sharing strategy; wherein, the preset sharing strategy is used to instruct the neurons of the feature extraction model to share information and update collaboratively.
[0070] Based on the multiple memory units included in the attention network, the initial output matrix is updated to obtain the updated memory output matrix;
[0071]
[0072] Among them, Output update The updated memory output matrix; Output origi This is the initial output matrix; WEI mem The set of stress bearing ratios between the foundation region corresponding to each grid and each long or short pile in the composite foundation map stored for multiple memory units; Distance is the long and short pile distance adjacency matrix;
[0073]
[0074] Where N is the number of memory units; Stress_GRID i Stress_STAKE represents the sum of the stress bearing capacity values of the foundation regions corresponding to each grid in the composite foundation map stored in the i-th memory unit. iThis represents the sum of the stress bearing values of each long or short pile stored in the i-th memory unit;
[0075] The prediction model for the ultimate bearing capacity of composite foundations with long and short piles is trained based on the updated memory output matrix, the first fusion matrix, and the second fusion matrix until the loss function converges.
[0076] Furthermore, the loss function is:
[0077]
[0078] Where T represents the number of grids in the composite geodetic map; M represents the sum of the number of long and short stakes; Distance represents the distance adjacency matrix of long and short stakes; FEA_Sedime k FEA_Stake represents the foundation settlement characteristics of the foundation region corresponding to the k-th grid in the composite foundation map. j The stress bearing characteristics corresponding to the j-th long pile or short pile; Output update This is the updated memory output matrix; FEA_Mix1 represents the first fusion matrix; FEA_Mix2 represents the second fusion matrix.
[0079] Furthermore, after using Optistruct finite element analysis software to model the piles, soil layers, and cushion layer in the composite foundation of long and short piles, and obtaining the composite foundation model, the method also includes:
[0080] The design parameters of the long-short pile composite foundation were modified using the Optistruct finite element analysis software to adjust the composite foundation model. The design parameters of the long-short pile composite foundation include the pile length, pile diameter, pile spacing, and area replacement ratio of the long and short piles.
[0081] Furthermore, long piles include CFG piles or reinforced concrete piles; short piles include mixing piles, crushed stone piles, or lime piles.
[0082] Furthermore, the material properties of the pile, soil layer, and cushion layer include unit weight, shear strength index, elastic modulus, and Poisson's ratio; among which, the shear strength index includes cohesion, friction angle, modulus, and Poisson's ratio.
[0083] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.
Claims
1. A method for analyzing the bearing capacity of composite foundations with long and short piles based on finite element analysis, characterized in that, include: Using Optistruct finite element analysis software, the piles, soil layers, and cushion layer in the long and short pile composite foundation are modeled to obtain the composite foundation model; wherein, the composite foundation map of the long and short pile composite foundation in the composite foundation model is meshed. Based on the first length information of the long piles, the second length information of the short piles in the composite foundation model, and the arrangement of the long piles and the short piles, the long and short pile distance adjacency matrix is determined. Vertical stress is simulated on the composite foundation of long and short piles to obtain the settlement results of the composite foundation of long and short piles; wherein, the settlement results of the composite foundation of long and short piles include the settlement information of each grid in the composite foundation map; The settlement information of each grid in the composite foundation map is vectorized to obtain the foundation settlement feature matrix; Based on the adjacency matrix of the long and short piles and the foundation settlement characteristic matrix, the bearing capacity of the long and short pile composite foundation is analyzed; after simulating the application of vertical stress on the long and short pile composite foundation to obtain the settlement results of the long and short pile composite foundation, the method further includes: Obtain the first load information and the first pile top reaction force of the long pile, and obtain the second load information and the second pile top reaction force of the short pile; The method further includes vectorizing the first load information, the first pile top reaction force, the second load information, and the second pile top reaction force to obtain the long and short pile stress bearing matrix; after simulating the application of vertical stress on the long and short pile composite foundation to obtain the settlement results of the long and short pile composite foundation, the method further includes: Obtain the soil layer thickness, groundwater conditions, and porosity information of the foundation corresponding to each grid in the composite foundation map; The soil layer thickness, groundwater conditions, and void ratio information corresponding to each grid in the composite foundation map are vectorized to obtain the composite foundation stress environment parameter matrix; the bearing capacity of the long and short pile composite foundation is analyzed based on the long and short pile distance adjacency matrix and the foundation settlement characteristic matrix, including: The bearing capacity of the composite foundation of long and short piles is analyzed based on the adjacency matrix of long and short piles, the foundation settlement characteristic matrix, the stress bearing capacity matrix of long and short piles, and the stress environment parameter matrix of the composite foundation. After analyzing the bearing capacity of the composite foundation of long and short piles based on the adjacency matrix of long and short piles, the foundation settlement characteristic matrix, the stress bearing capacity matrix of long and short piles, and the stress environment parameter matrix of the composite foundation, the method further includes: Dimensionality reduction of the long and short pile distance adjacency matrix, the foundation settlement feature matrix, the long and short pile stress bearing matrix, and the composite foundation stress environment parameter matrix includes: inputting the long and short pile distance adjacency matrix, the foundation settlement feature matrix, the long and short pile stress bearing matrix, and the composite foundation stress environment parameter matrix into a preset autoencoder; wherein, the autoencoder uses a feature-based distance metric learning method for data conversion, and the autoencoder reduces the dimensionality of the long and short pile distance adjacency matrix, the foundation settlement feature matrix, the long and short pile stress bearing matrix, and the composite foundation stress environment parameter matrix based on the learned distance metric error; The reduced long and short pile distance adjacency matrix, the reduced foundation settlement characteristic matrix, and the reduced long and short pile stress bearing matrix are fused to obtain the first fused matrix; The reduced composite foundation stress environment parameters and the reduced foundation settlement characteristic matrix are fused to obtain the second fusion matrix; Using the first fusion matrix and the second fusion matrix as input sequence information, a prediction model for the ultimate bearing capacity of composite foundations with long and short piles is established. The ultimate bearing capacity prediction model for the long-short pile composite foundation is used to predict the ultimate bearing capacity of the long-short pile composite foundation based on the bearing capacity analysis results. The bearing capacity analysis results include the analysis results of the bearing capacity of the long-short pile composite foundation after simulating the application of vertical stress on the long-short pile composite foundation.
2. The method for analyzing the bearing capacity of composite foundations with long and short piles based on finite element analysis according to claim 1, characterized in that, The prediction model for the ultimate bearing capacity of the composite foundation with long and short piles includes a self-attention network, which includes multiple memory units. The multiple memory units are used to determine the correlation between the bearing capacity of the foundation area corresponding to each grid in the composite foundation map and the location corresponding to each grid in the composite foundation map. The step of establishing a prediction model for the ultimate bearing capacity of a composite foundation with long and short piles, using the first fusion matrix and the second fusion matrix as input sequence information, includes: The input sequence information is input into the feature extraction model, and the initial output matrix corresponding to the input sequence information is determined based on a preset sharing strategy; wherein, the preset sharing strategy is used to instruct the neurons of the feature extraction model to share information and update collaboratively. Based on the multiple memory units included in the attention network, the initial output matrix is updated to obtain the updated memory output matrix; ， in, This is the updated memory output matrix; This is the initial output matrix; The set of stress bearing ratio values of the foundation area corresponding to each grid and each long pile or short pile in the composite foundation map stored in the multiple memory units; This is the adjacency matrix of the long and short piles; ， in, The number of memory units; This represents the sum of the stress bearing capacity values of the foundation regions corresponding to each grid in the composite foundation map stored in the i-th memory unit; This represents the sum of the stress bearing values of each long or short pile stored in the i-th memory unit; The prediction model for the ultimate bearing capacity of the composite foundation of long and short piles is trained based on the updated memory output matrix, the first fusion matrix, and the second fusion matrix until the loss function converges.
3. The method for analyzing the bearing capacity of composite foundations with long and short piles based on finite element analysis according to claim 2, characterized in that, The loss function is: ， in, This indicates the number of grids in the composite geodetic map; This represents the sum of the number of the long piles and the short piles; This represents the foundation settlement characteristics of the foundation region corresponding to the k-th grid in the composite foundation map. This represents the stress bearing characteristics corresponding to the j-th long pile or short pile; This is the updated memory output matrix; This represents the first fusion matrix; This represents the second fusion matrix.
4. The method for analyzing the bearing capacity of composite foundations with long and short piles based on finite element analysis according to claim 1, characterized in that, After using Optistruct finite element analysis software to model the piles, soil layers, and cushion layer in the composite foundation of long and short piles to obtain the composite foundation model, the method further includes: The composite foundation model is adjusted by modifying the design parameters of the long and short pile composite foundation using the Optistruct finite element analysis software. The design parameters of the long and short pile composite foundation include the pile length, pile diameter, pile spacing, and area replacement ratio of the long and short piles.
5. The method for analyzing the bearing capacity of composite foundations with long and short piles based on finite element analysis according to claim 1, characterized in that, The long piles include CFG piles or reinforced concrete piles; the short piles include mixing piles, crushed stone piles or lime piles.
6. The method for analyzing the bearing capacity of composite foundations with long and short piles based on finite element analysis according to claim 1, characterized in that, The material properties of the pile, the soil layer, and the cushion layer include unit weight, shear strength index, elastic modulus, and Poisson's ratio; wherein the shear strength index includes cohesion, friction angle, modulus, and Poisson's ratio.