A shallow foundation bearing capacity calculation method fusing machine learning and non-local algorithm

Abstract

This invention discloses a method for calculating the bearing capacity of shallow foundations by integrating machine learning and nonlocal algorithms, belonging to geotechnical engineering. 1. The type of foundation soil is determined through indoor triaxial tests; strain-hardening foundation soils are analyzed; 2. Strain-softening foundation soils are analyzed; 3. A finite element model is established for refined simulation, followed by 7. 3. A constitutive model combined with a nonlocal algorithm is used for refined simulation, then proceeds to 4. 4. The influence range (DL) of the shallow foundation analysis is calibrated. f DL is determined by the softening rate control parameter β. f If suitable, proceed to step 7; otherwise, proceed to step 6. Step 6 involves building and training a machine learning agent model to ultimately obtain a suitable deep learning (DL) algorithm. f Shallow foundation analysis was performed with β1 and 7 to obtain the predicted value of the foundation bearing capacity. The present invention uses the above method to achieve efficient inversion of nonlocal parameters, thereby obtaining a reasonable predicted value of the shallow foundation bearing capacity.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geotechnical engineering, and particularly to a shallow foundation bearing capacity calculation method combining machine learning and non-local algorithm. BACKGROUND

[0002] The bearing capacity calculation of the foundation during the loading process of the shallow foundation is very important for the geotechnical engineering design. In the shallow foundation bearing capacity analysis, the traditional specification method is based on the elastic half-space theory model, which assumes that the stress and strain of the soil body are in linear elastic relationship, which is significantly deviated from the actual nonlinear, elastoplastic and layered characteristics of the soil body. At the same time, the determination of the key parameters (such as the base coefficient) in the specification mainly depends on the empirical formula or the simplified assumption, which is easy to cause the calculation results to deviate from the engineering practice.

[0003] The bearing capacity calculation of the shallow foundation is an important part of the geotechnical engineering. When the finite element method is used to analyze the shallow foundation built on the foundation with softening characteristics, the non-local algorithm needs to be used to eliminate the grid dependency problem of the calculation results in the softening stage. In addition, the stress and strain characteristics of the soil body are complex, and a reasonable constitutive model needs to be used for description. However, when the constitutive model combined with the non-local algorithm is used to analyze the shallow foundation bearing capacity problem, the non-local parameter should only change the softening rate of the load displacement curve to eliminate the grid dependency problem, but the value of the non-local parameter has a great influence on the residual state of the load displacement curve, which leads to the failure to obtain a reasonable prediction value of the foundation bearing capacity.

[0004] Based on this, the present application provides a shallow foundation bearing capacity calculation method combining machine learning and non-local algorithm. SUMMARY

[0005] The purpose of the present application is to provide a shallow foundation bearing capacity calculation method combining machine learning and non-local algorithm. Based on the simulation of the indoor triaxial test, the mutual relationship of the influence range, the softening rate and the problem size is calibrated for the shallow foundation analysis, the influence range and the softening rate control parameters are calibrated, and the machine learning technology is used to increase the efficiency, so that the non-local parameter is efficiently inverted, and a reasonable prediction value of the shallow foundation bearing capacity is obtained.

[0006] To achieve the above purpose, the present application provides a shallow foundation bearing capacity calculation method combining machine learning and non-local algorithm, comprising the following steps:

[0007] S1, judging the type of the foundation soil through the results of the indoor triaxial test, if the type of the foundation soil is the strain hardening type of the foundation soil, then S2 is performed, if the type of the foundation soil is the strain softening type of the foundation soil, then S3 is performed;

[0008] S2. Based on the triaxial test conditions, establish a finite element model, and based on the selected constitutive model, perform a detailed simulation of the entire triaxial test process, calibrate the constitutive model parameters, and then proceed to S7.

[0009] S3. Based on the triaxial test conditions, establish a finite element model and use a constitutive model combined with a nonlocal algorithm to perform a detailed simulation of the entire triaxial test process, and then proceed to S4.

[0010] S4. Calibrate constitutive model parameters and determine the influence range of triaxial analysis (DL) t And the softening rate control parameter β, and then the influence range DL of the triaxial analysis. t Convert to shallow foundation analysis scope DL f ;

[0011] S5. Determine the influence range of the shallow foundation analysis based on the mesh size used in the shallow foundation analysis. f If it is suitable, proceed to S7; if not, proceed to S6.

[0012] S6. Construct and train a machine learning surrogate model, and use the machine learning surrogate model to calibrate the constitutive model parameters, the influence range of the new triaxial experimental simulation, and the new softening rate control parameters, ultimately obtaining a suitable DL. f ;

[0013] S7. Using the calibrated constitutive model parameters and appropriate shallow foundation analysis influence range and softening rate control parameters, perform shallow foundation analysis to obtain the predicted value of foundation bearing capacity.

[0014] Preferably, the process of determining the foundation soil type in S1 based on the results of indoor triaxial tests is as follows:

[0015] S11. Take samples of the selected foundation soil to obtain sample foundation soil, and perform water injection, counterpressure saturation, K0 consolidation and pressure shearing operations on the sample foundation soil according to the "Standard for Geotechnical Testing Methods GB / T50123-2019".

[0016] S12. During the pressure shearing process, record the state of the foundation soil sample every 1-2 seconds to obtain the stress-strain curve. Observe the stress-strain curve. If the curve shape shows that the stress continuously increases with the increase of strain and eventually reaches a peak and no longer changes, then the foundation soil is a strain-hardening type of foundation soil. Proceed to S2. If the curve shape shows that the stress first reaches a peak and then begins to decrease with the increase of strain and eventually no longer changes, then the foundation soil is a strain-softening type of foundation soil. Proceed to S3.

[0017] Preferably, the process of S2 is as follows:

[0018] S21. Based on the actual dimensions of the foundation soil of the sample tested in S1, establish a three-dimensional finite element model that is consistent with the sample dimensions and considers fluid-structure interaction.

[0019] S22. Add contact surface elements to both ends of the established three-dimensional finite element model to simulate the end constraint effect of the test instrument on both ends of the specimen.

[0020] S23. Apply a total stress increment considering gravity gradient to the cylindrical surface of the foundation soil of the sample, and apply a condition that the pore pressure increment is zero at both ends of the sample. Then apply a total stress increment to the cylindrical surface and top of the foundation soil of the sample, and apply a pore pressure increment depending on the in-situ pore water pressure of the foundation soil of the sample to both ends of the sample.

[0021] S24. Apply different total stress increments to the cylindrical surface and both ends of the sample foundation soil to restore the K0 consolidation state of the sample foundation soil, and set the pore pressure increment at both ends of the sample to zero. Then, bind all pore water pressure degrees of freedom at the top and bottom of the sample respectively, apply the condition of zero total stress increment to the cylindrical surface, and apply a constant displacement increment to the top for loading.

[0022] S25. Complete the refined simulation to obtain the stress-strain curve, and then compare it with the measured stress-strain test data in the triaxial test to obtain the calibrated constitutive model parameters, and then proceed to S7.

[0023] Preferably, S3 uses a constitutive model combined with a nonlocal algorithm to perform a refined simulation of the entire triaxial test process as follows:

[0024] S31. Based on the actual dimensions of the foundation soil of the sample tested in S1, establish a three-dimensional finite element model that is consistent with the sample dimensions and considers fluid-structure interaction.

[0025] S32. Add contact surface elements to both ends of the established three-dimensional finite element model to simulate the end constraint effect of the test instrument on both ends of the specimen.

[0026] S33. Use the same method as S23-S24 to perform a refined simulation to obtain the stress-strain curve, and then proceed to S4.

[0027] Preferably, in S4, the constitutive model parameters are calibrated and the influence range of the shallow foundation analysis is DL. f The process of controlling the softening rate parameter β is as follows:

[0028] S41. Select a suitable triaxial experiment influence range (DL) for triaxial experiment simulation. t DL t The value is 1-2 times the mesh size used in the triaxial test simulation;

[0029] S42. Compare the stress-strain curves obtained in S33 with the measured stress-strain test data in the triaxial test, and calibrate the constitutive model parameters and the softening rate control parameter β.

[0030] S43. Using the problem size, calculate the influence range DL of the triaxial test calibrated in the triaxial test simulation. t Scaled to the shallow foundation analysis impact range (DL) for analyzing shallow foundation problems. f The method is as follows:

[0031] DL f / B f =DL t / B t ;

[0032] Where B represents the problem size, the subscript f indicates shallow foundation analysis, and the subscript t indicates triaxial experiment.

[0033] Preferably, in S5, the influence range of shallow foundation analysis (DL) is determined. f Whether it is appropriate depends on having enough Gaussian points to calculate nonlocal strain, thereby discretizing the excessively concentrated strain on the shear band to surrounding elements, reducing stress / strain concentration on the shear band, and thus eliminating mesh dependency in finite element calculations. Therefore, if DL... f If the value of is greater than the size of the largest grid on the shear band, then β and DL will be... f If the appropriate softening rate control parameters and the influence range of the shallow foundation analysis are required, proceed to S7; otherwise, proceed to S6.

[0034] Preferably, the process of constructing and training the machine learning agent model in S6, and obtaining a suitable shallow foundation analysis to define the scope of influence is as follows:

[0035] S61. Using nonlinear multi-scale linear compression technology, data characteristics of the high-dimensional output space are proposed, and the output result set of the finite element method is... From the original space D Y Mapped to reduced space D Z ;

[0036] S62, Based on reduced space D Z By using polynomial chaotic expansion technology, a polynomial approximation function is used to replace the finite element model used in the above triaxial test simulation to construct a proxy model;

[0037] S63. Calculate the triaxial test simulation results using finite element method, extract the data features of the output results using nonlinear multi-scale linear compression technology, and train the surrogate model.

[0038] S64. Select a new triaxial experimental influence range DL' tThe new softening rate control parameter β1 was used, and the constitutive model parameters were adjusted. Then, the triaxial test simulation results were output using the surrogate model.

[0039] S65. If the simulation results obtained in S64 are inconsistent with the experimental results, return to S64 to readjust the constitutive model parameters and softening rate control parameters until the triaxial test simulation results match the experimental data. At this point, the calibrated constitutive model parameters and the new triaxial test influence range DL' are obtained. t And the new softening rate control parameter β1;

[0040] S66. Obtain DL' from S65 t The new shallow foundation analysis influence range DL' is obtained by scaling using the method in S43. f Then determine DL' f If it is suitable, return to S64 if it is not suitable; otherwise, check if β1 and DL' are suitable. f To determine the appropriate softening rate control parameters and the influence range of the shallow foundation analysis, proceed to S7.

[0041] Preferably, the process of performing shallow foundation analysis in S7 to obtain the predicted value of the foundation bearing capacity is as follows:

[0042] S71. If the foundation soil is strain hardening type, then the constitutive model parameters from S2 are received; if the foundation soil is strain softening type, then the constitutive model parameters from S3-S6, along with appropriate softening rate control parameters and the influence range of shallow foundation analysis are received.

[0043] S72. Displacement-controlled loading of shallow foundations is adopted, and the reaction force of the soil on the shallow foundation is obtained by finite element calculation.

[0044] S73. Repeat the process in S72 to obtain the load-displacement curve of the shallow foundation. When the load-displacement curve tends to be stable, apply vertical displacement to the shallow foundation. Stop the finite element analysis when the reaction force of the foundation soil on the shallow foundation remains basically unchanged. The reaction force value obtained at this time is the predicted value of the foundation bearing capacity.

[0045] Therefore, the shallow foundation bearing capacity calculation method of the present invention, which integrates machine learning and nonlocal algorithms using the above-mentioned structure, has the following advantages:

[0046] 1. This solves the problem that when using constitutive models combined with nonlocal algorithms to analyze shallow foundations built on softening soil, the nonlocal parameters were not calibrated using a reasonable method to obtain reasonable predicted values ​​of the foundation bearing capacity.

[0047] 2. The method proposed in this invention is based on the analysis of the influence law of nonlocal parameters. It does not affect the previous triaxial test process and does not require any additional indoor or outdoor tests related to shallow foundations.

[0048] 3. The method proposed in this invention uses numerical simulation and machine learning techniques to calibrate parameters, which can be completed automatically by a high-performance server without the need for consumables. Therefore, it has low labor and economic costs and high repeatability.

[0049] 4. By using shallow foundation analysis to calibrate the relationship between the influence range, softening rate and problem size, constitutive model parameters and nonlocal parameters are calibrated for shallow foundation analysis involving softening. Machine learning technology is used to enhance efficiency, achieving efficient inversion of nonlocal parameters, thereby obtaining reasonable predicted values ​​of shallow foundation bearing capacity, while also taking into account the predicted values ​​of shallow foundation bearing capacity for hardened foundation soil.

[0050] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0051] Figure 1 This is a flowchart of a shallow foundation bearing capacity calculation method that integrates machine learning and nonlocal algorithms according to the present invention;

[0052] Figure 2 This is a flowchart of the machine learning proxy model for a shallow foundation bearing capacity calculation method that integrates machine learning and nonlocal algorithms, as described in this invention. Detailed Implementation

[0053] Example

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0055] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0056] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0057] In the description of this invention, it should be noted that the terms "upper," "lower," "inner," "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the product of this invention is usually placed when in use. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0058] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," and "connect" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0059] The following detailed description of some embodiments of the present invention is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0060] like Figure 1 and Figure 2 As shown, the present invention provides a method for calculating the bearing capacity of shallow foundations that integrates machine learning and nonlocal algorithms, comprising the following steps:

[0061] S1. Determine the type of foundation soil based on the results of indoor triaxial tests. If the foundation soil is strain hardening type, proceed to S2; if the foundation soil is strain softening type, proceed to S3.

[0062] S11. Take samples of the selected foundation soil to obtain sample foundation soil, and perform water injection, counterpressure saturation, K0 consolidation and pressure shearing operations on the sample foundation soil according to the "Standard for Geotechnical Testing Methods GB / T50123-2019".

[0063] S12. During the pressure shearing process, record the state of the foundation soil sample every 1-2 seconds to obtain the stress-strain curve. Observe the stress-strain curve. If the curve shape shows that the stress continuously increases with the increase of strain and eventually reaches a peak and no longer changes, then the foundation soil is a strain-hardening type of foundation soil. Proceed to S2. If the curve shape shows that the stress first reaches a peak and then begins to decrease with the increase of strain and eventually no longer changes, then the foundation soil is a strain-softening type of foundation soil. Proceed to S3.

[0064] S2. Based on the triaxial test conditions, establish a finite element model, and based on the selected constitutive model, perform a detailed simulation of the entire triaxial test process, calibrate the constitutive model parameters, and then proceed to S7.

[0065] S21. Based on the actual dimensions of the foundation soil of the sample tested in S1, establish a three-dimensional finite element model that is consistent with the sample dimensions and considers fluid-structure interaction.

[0066] S22. Add contact surface elements to both ends of the established three-dimensional finite element model to simulate the end constraint effect of the test instrument on both ends of the specimen.

[0067] S23. Apply a total stress increment considering gravity gradient to the cylindrical surface of the sample foundation soil, and apply a condition that the pore pressure increment is zero at both ends of the sample. Then apply a total stress increment to the cylindrical surface and top of the sample foundation soil, and apply a pore pressure increment depending on the in-situ pore water pressure of the sample foundation soil to both ends of the sample. This step needs to be long enough to meet the drainage conditions.

[0068] S24. Apply different total stress increments to the cylindrical surface and both ends of the sample foundation soil to restore the K0 consolidation state of the sample foundation soil, and set the pore pressure increment at both ends of the sample to zero. Then, bind all pore water pressure degrees of freedom at the top and bottom of the sample respectively, apply the condition of zero total stress increment to the cylindrical surface, and apply a constant displacement increment to the top for loading. The loading rate needs to be determined according to the triaxial test conditions.

[0069] S25. Complete the refined simulation to obtain the stress-strain curve, and then compare it with the measured stress-strain test data in the triaxial test to obtain the calibrated constitutive model parameters, and then proceed to S7.

[0070] S3. Based on the triaxial test conditions, establish a finite element model and use a constitutive model combined with a nonlocal algorithm to perform a detailed simulation of the entire triaxial test process, and then proceed to S4.

[0071] S31. Based on the actual dimensions of the foundation soil of the sample tested in S1, establish a three-dimensional finite element model that is consistent with the sample dimensions and considers fluid-structure interaction.

[0072] S32. Add contact surface elements to both ends of the established three-dimensional finite element model to simulate the end constraint effect of the test instrument on both ends of the specimen.

[0073] S33. Use the same method as S23-S24 to perform a refined simulation to obtain the stress-strain curve, and then proceed to S4.

[0074] S4. Calibrate constitutive model parameters and determine the influence range of triaxial analysis (DL) t And the softening rate control parameter β, and then the influence range DL of the triaxial analysis. t Convert to shallow foundation analysis scope DL f ;

[0075] S41. Select a suitable triaxial experiment influence range (DL) for triaxial experiment simulation. t DL t The value is 1-2 times the mesh size used in the triaxial test simulation;

[0076] S42. Compare the stress-strain curves obtained in S33 with the measured stress-strain test data in the triaxial test, and calibrate the constitutive model parameters and the softening rate control parameter β.

[0077] S43. Using the problem size, calculate the influence range DL of the triaxial test calibrated in the triaxial test simulation. t Scaled to the shallow foundation analysis impact range (DL) for analyzing shallow foundation problems. f The method is as follows:

[0078] DL f / B f =DL t / B t ;

[0079] Where B represents the problem size, the subscript f indicates shallow foundation analysis, and the subscript t indicates triaxial experiment.

[0080] S5. Determine the influence range of the shallow foundation analysis based on the mesh size used in the shallow foundation analysis. f If it is suitable, proceed to S7; otherwise, proceed to S6. DL is a parameter in the nonlocal algorithm that controls the range of influence. That is, all Gaussian points within a radius of DL around the current Gaussian point will participate in the calculation of nonlocal strain.

[0081] Determine the scope of influence of shallow foundation analysis (DL) f Whether it is appropriate depends on having enough Gaussian points to calculate nonlocal strain, thereby discretizing the excessively concentrated strain on the shear band to surrounding elements, reducing stress / strain concentration on the shear band, and thus eliminating mesh dependency in finite element calculations. Therefore, if DL... f If the value of is greater than the size of the largest grid on the shear band, then β and DL will be... f If the appropriate softening rate control parameters and the influence range of the shallow foundation analysis are required, proceed to S7; otherwise, proceed to S6.

[0082] S6. Construct and train a machine learning surrogate model, and use the machine learning surrogate model to calibrate the constitutive model parameters, the influence range of the new triaxial experimental simulation, and the new softening rate control parameters, ultimately obtaining a suitable DL. f ;

[0083] S61. Using nonlinear multiscale linear compression (MDS) technology, data features of the high-dimensional output space are proposed, and the output result set of the finite element method is... From the original space DY Mapped to reduced space D Z ;

[0084] S62, Based on reduced space D Z By using polynomial chaotic expansion (PCE) technology, a polynomial approximation function is used to replace the finite element model used in the above triaxial test simulation to construct a proxy model;

[0085] S63. Calculate the triaxial test simulation results using finite element method, extract the data features of the output results using nonlinear multi-scale linear compression technology, and train the surrogate model.

[0086] S64. Select a new triaxial experimental influence range DL' t The new softening rate control parameter β1 was used, and the constitutive model parameters were adjusted. Then, the triaxial test simulation results were output using the surrogate model.

[0087] S65. If the simulation results obtained in S64 are inconsistent with the experimental results, return to S64 to readjust the constitutive model parameters and softening rate control parameters until the triaxial test simulation results match the experimental data. At this point, the calibrated constitutive model parameters and the new triaxial test influence range DL' are obtained. t And the new softening rate control parameter β1;

[0088] S66. Obtain DL' from S65 t The new shallow foundation analysis influence range DL' is obtained by scaling using the method in S43. f Then determine DL' f If it is suitable, return to S64 if it is not suitable; otherwise, check if β1 and DL' are suitable. f To determine the appropriate softening rate control parameters and the influence range of the shallow foundation analysis, proceed to S7.

[0089] S7. Using the calibrated constitutive model parameters and appropriate shallow foundation analysis influence range and softening rate control parameters, perform shallow foundation analysis to obtain the predicted value of foundation bearing capacity.

[0090] S71. If the foundation soil is strain hardening type, then the constitutive model parameters from S2 are received; if the foundation soil is strain softening type, then the constitutive model parameters from S3-S6, along with appropriate softening rate control parameters and the influence range of shallow foundation analysis are received.

[0091] S72. Displacement-controlled loading of shallow foundations is adopted, and the reaction force of the soil on the shallow foundation is obtained by finite element calculation.

[0092] S73. Repeat the process in S72 to obtain the load-displacement curve of the shallow foundation. When the load-displacement curve tends to be stable, apply vertical displacement to the shallow foundation. Stop the finite element analysis when the reaction force of the foundation soil on the shallow foundation remains basically unchanged. The reaction force value obtained at this time is the predicted value of the foundation bearing capacity.

[0093] Therefore, this invention employs a shallow foundation bearing capacity calculation method that integrates machine learning and nonlocal algorithms. This method solves the problem of failing to calibrate nonlocal parameters and obtain reasonable foundation bearing capacity predictions when using constitutive models combined with nonlocal algorithms to analyze shallow foundations built on softening soil. Furthermore, based on the analysis of the influence of nonlocal parameters, it does not affect the initial triaxial experimental process and eliminates the need for any additional indoor or outdoor experiments related to the shallow foundation. Numerical simulation and machine learning techniques are used for parameter calibration, which can be automatically completed by a high-performance server without the need for consumables. Therefore, it has low labor and economic costs and high repeatability. Simultaneously, by utilizing the relationship between the influence range, softening rate, and problem size in shallow foundation analysis, constitutive model parameters and nonlocal parameters are calibrated for shallow foundations involving softening. Machine learning techniques are used to enhance efficiency, achieving efficient inversion of nonlocal parameters, thereby obtaining reasonable shallow foundation bearing capacity predictions, while also considering the shallow foundation bearing capacity predictions for hardened soil types.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method of shallow foundation bearing capacity calculation fusing machine learning with non-local algorithms, characterized in that, Comprising the following steps: S1, judging the type of foundation soil through the results of indoor triaxial test, if the type of foundation soil is strain-hardening type of foundation soil, then S2 is performed; if the type of foundation soil is strain-softening type of foundation soil, then S3 is performed; S2, establishing a finite element model according to the triaxial test conditions, and based on the selected constitutive model, performing fine simulation on the whole process of triaxial test, calibrating the parameters of the constitutive model, and then S7 is performed; S3, establishing a finite element model according to the triaxial test conditions, and using the constitutive model combined with the non-local algorithm to perform fine simulation on the whole process of triaxial test, and then S4 is entered; S4, calibrate constitutive model parameters and triaxial analysis influence range and softening rate control parameters , then convert triaxial analysis influence range to shallow foundation analysis influence range ; S5. Determine the shallow foundation analysis influence range according to the grid size used in the shallow foundation analysis S7. If appropriate, go to S7, if not appropriate, go to S6; S6, construct and train the machine learning agent model, and calibrate the constitutive model parameters and the influence range of the new triaxial test simulation and the new softening rate control parameters using the machine learning agent model, and finally obtain suitable ; S61, using nonlinear multi-scale linear compression technology, the data characteristics of high-dimensional output space are proposed, and the output result set of finite elements is mapped from the original space to the reduced space ;​ S62、based on the reduced space An agent model is constructed by replacing the finite element model used in the above triaxial test simulation with a polynomial approximation function using a polynomial chaos expansion technique. S63, using the triaxial test simulation results calculated by the finite element, using the nonlinear multi-scale linear compression technology to extract the data characteristics of the output results, and training the surrogate model; S64, select new triaxial experiment influence range and new softening rate control parameters and adjust the constitutive model parameters, and then output the triaxial test simulation results using the proxy model; S65, if the simulation result obtained in S64 is inconsistent with the experimental result, return to S64 to adjust the constitutive model parameters and the softening rate control parameters again, until the triaxial test simulation result is consistent with the test data, at this time the calibrated constitutive model parameters and the new and the new softening rate control parameters ; S66, the result obtained in S65 The new shallow foundation analysis influence range is obtained by scaling using the method in S43. Then judge If it is suitable, return to S64 if it is not suitable; otherwise, proceed to the next step. and To determine the appropriate softening rate control parameters and the influence range of the shallow foundation analysis, proceed to S7; S7, using the calibrated constitutive model parameters and appropriate shallow foundation analysis influence range and softening rate control parameters to perform shallow foundation analysis, and obtaining the predicted value of foundation bearing capacity.

2. The method of claim 1, wherein the method is characterized by, The process of judging the type of foundation soil in S1 is as follows: S11, sampling the selected foundation soil to obtain a sample foundation soil, and performing water injection, back pressure saturation, K0 consolidation and pressure shear operations on the sample foundation soil according to the Standard for Soil Test Methods GB / T 50123-2019; S12, during the pressure shear process, the state of the sample foundation soil is recorded once every 1-2 seconds to obtain a stress-strain curve, and the stress-strain curve is observed, if the curve shape increases with strain, the stress continuously increases and finally reaches a peak value that does not change, then the foundation soil is a strain-hardening type of foundation soil, S2 is performed, if the curve shape increases with strain, the stress first reaches a peak value and then starts to decrease and finally does not change, then the foundation soil is a strain-softening type of foundation soil, S3 is performed.

3. The method of claim 2, wherein the method is characterized by, The process of S2 is as follows: S21, according to the true size of the sample foundation soil subjected to triaxial test in S1, a three-dimensional finite element model consistent with the size of the sample and considering fluid-structure coupling is established; S22, contact surface elements for simulating the end constraint effect of the test instrument on the two ends of the sample are added to the established three-dimensional finite element model; S23, total stress increment considering gravity gradient is applied to the cylindrical surface of the sample foundation soil, and the condition of zero pore pressure increment is applied to the two ends of the sample, then total stress increment is applied to the cylindrical surface and the top end of the sample foundation soil, and pore pressure increment dependent on the in-situ pore water pressure of the sample foundation soil is applied to the two ends of the sample; S24, different total stress increments are applied to the cylindrical surface and the two ends of the sample foundation soil to restore the K0 consolidation state of the sample foundation soil, and the pore pressure increment of the two ends of the sample is set to zero, then all the pore water pressure degrees of freedom of the top and bottom ends of the sample are bound, total stress increment of zero is applied to the cylindrical surface, and constant displacement increment is applied to the top end for loading; S25, fine simulation is completed to obtain a stress-strain curve, then the calibrated constitutive model parameters are obtained by comparing the stress-strain test data measured in the triaxial test, and then S7 is entered.

4. The method of claim 3, wherein the method is characterized by, The process of S3 using the constitutive model combined with the non-local algorithm to perform fine simulation on the whole process of triaxial test is as follows: S31, according to the real size of the sample foundation soil in S1, a three-dimensional finite element model is established which is consistent with the size of the sample and considers fluid-structure coupling; S32, contact surface elements are added to both ends of the established three-dimensional finite element model to simulate the end constraint effect of the test instrument on both ends of the sample; S33, the same method as S23-S24 is used for refined simulation to obtain the stress-strain curve, and then go to S4.

5. The method of claim 4, wherein the method is characterized by, S4 calibrates the constitutive model parameters and analyzes the influence range of the shallow foundation with the softening rate control parameter The process is as follows: S41、For triaxial test simulation to select a suitable triaxial experiment influence range , The value of the grid size used for triaxial test simulation is 1-2 times. S42, compare the stress-strain curve obtained in S33 with the measured stress-strain test data in the triaxial test, and calibrate the constitutive model parameters and the softening rate control parameters ; S43, using the problem size to scale the influence range of triaxial test calibration in triaxial test simulation scaling to the shallow foundation analysis influence range for analyzing shallow foundation problems The method is as follows: ； Where B represents the problem size, subscript f represents shallow foundation analysis, and subscript t represents triaxial test.

6. The method of claim 5, wherein the method is characterized by: S5: judging the shallow foundation analysis influence range Whether it is appropriate, need to have enough Gauss points for the calculation of non-local strain, so that the shear band on too concentrated strain to the surrounding elements dispersed, weaken the shear band on the stress / strain concentration, so as to eliminate the grid dependence in the finite element calculation, so if The value of the maximum grid size on the shear band is greater than the value of the maximum grid size on the shear band. At this time, And The softening rate control parameter and the shallow foundation analysis influence range are appropriate, turn to S7, otherwise turn to S6.

7. The method of claim 6, wherein the method further comprises: The process of shallow foundation analysis in S7 to obtain the predicted value of foundation bearing capacity is as follows: S71, if the foundation soil is a strain-hardening type of foundation soil, the constitutive model parameters are received from S2, if the foundation soil is a strain-softening type of foundation soil, the constitutive model parameters and appropriate softening rate control parameters and shallow foundation analysis influence range are received from S3-S6; S72, shallow foundation is loaded by displacement control, and the reaction force value of the foundation to the shallow foundation is obtained by finite element calculation; S73, the process of S72 is repeated to obtain the load displacement curve of the shallow foundation, when the load displacement curve tends to be stable, vertical displacement is applied to the shallow foundation, and when the reaction force of the foundation soil to the shallow foundation is basically unchanged, the finite element analysis is stopped, and the reaction force value obtained at this time is the predicted value of the foundation bearing capacity.

Images