A method, device and electronic equipment for calculating the reliability of pile horizontal deformation

Abstract

The present invention discloses a method, device and electronic device for calculating the reliability of horizontal deformation of pile foundations, belonging to the technical field of pile foundation engineering, the method comprising: collecting engineering data; determining factors affecting the horizontal deformation of pile foundations and their simulation calculation value ranges; designing a three-dimensional finite element numerical simulation calculation scheme for the horizontal deformation of pile foundations based on the uniform test design method; obtaining the horizontal deformation value of pile foundations based on the three-dimensional finite element numerical simulation calculation; establishing a training sample data set and a test sample data set; establishing a horizontal deformation proxy model of pile foundations based on a machine learning model; establishing a limit state equation for calculating the reliability of horizontal deformation of pile foundations; calculating the failure probability and reliability index of horizontal deformation of pile foundations. The present invention has the advantages of good fault tolerance, strong adaptability, high fitting and prediction accuracy, and solves the technical problem that it is difficult to construct a high-precision proxy model of pile foundation deformation due to too many input variables.

Description

Technical Field

[0001] The present invention belongs to the technical field of pile foundation engineering, and in particular relates to a method, a device and an electronic device for calculating the reliability of horizontal deformation of pile foundation based on an agent model. Background Art

[0002] Pile foundations in highway, railway and marine engineering fields are generally subjected to large horizontal loads, and the resulting horizontal deformation (specifically the horizontal displacement of the pile top) directly affects the safety and reliability of the pile foundation and its superstructure. Since the factors affecting the horizontal deformation of pile foundations are numerous and complex, and have strong randomness and uncertainty, in order to ensure the safety and reliability of such pile foundations, it is necessary to conduct in-depth research on the reliability evaluation method of the horizontal deformation of the piles in such pile foundations.

[0003] For example, the 7th issue of Rock and Soil Mechanics in 2004 introduced a reliability calculation method for pile top horizontal displacement using the py curve method. This document 1 used the py curve method to calculate the pile top horizontal displacement and constructed a functional function based on it, and used the point estimation method to estimate the failure probability of the pile. However, the py curve model and model parameters in complex terrain and multi-layer foundations are difficult to determine, and the existing py curve model is not suitable for calculating the pile top horizontal displacement of foundation piles in multiple strata or inclined ground.

[0004] For example, the Journal of Railway Science and Engineering, No. 1, 2005, introduced a reliability calculation method for pile top horizontal displacement based on quadratic polynomial response surface method. This document 2 uses the finite bar element method to calculate the pile top horizontal displacement, uses quadratic polynomials to construct the response surface expression of the pile top horizontal displacement (i.e., the pile top horizontal displacement proxy model) and the function function, and uses the verification point method (i.e., JC method) to solve the reliability index. However, the finite bar element method assumes the reaction force of the foundation soil on the pile side as a linear elastic foundation reaction force model, that is, the foundation soil is assumed to be a spring, providing a linear elastic horizontal resistance, while the resistance provided by the soil on the pile side in actual engineering usually presents a nonlinear elastic or elastoplastic form. Therefore, the pile horizontal displacement calculated by the finite bar element method is often inconsistent with the actual situation.

[0005] The 12th issue of Rock and Soil Mechanics in 2007 introduced a reliability calculation method for pile top horizontal displacement based on the fourth-order polynomial response surface method. This document 3 still uses the finite bar element method to calculate the pile top horizontal displacement, but uses the fourth-order polynomial to construct the response surface expression and function of the pile top horizontal displacement, and also uses the verification point method to solve the reliability index. Therefore, like document 1, it also has the problem that the calculated pile horizontal displacement is often inconsistent with the actual situation.

[0006] The dissertation "Reliability Analysis of Pile-column Bridge Foundation Deformation and Its Engineering Application Research" (Cheng Ying, 2014) introduced a reliability calculation method for pile top horizontal displacement based on the m-method. The document4 used the m-method in the linear elastic foundation reaction method to calculate the pile top horizontal displacement and construct a function function based on it, and used the verification point method to solve the reliability index. The m-method actually also uses the linear elastic foundation reaction model, so the horizontal displacement calculated by it is also difficult to conform to reality.

[0007] The 2015 Supplement 1 of the Chinese Journal of Geotechnical Engineering also introduced a reliability calculation method for pile top horizontal displacement based on the m method. Like document 4, this document 5 has the problem that the pile horizontal displacement does not match the actual situation.

[0008] In the existing technology of the proxy model construction method, directly using the traditional random finite element method (i.e. directly calling the three-dimensional finite element calculation model for each sampling) to calculate the deformation failure probability and reliability index of the pile foundation will face a huge computational workload. To this end, a small number of samples can be first used to perform three-dimensional finite element simulation calculations to construct a pile foundation deformation proxy model, and then the deformation failure probability and reliability index calculations are performed based on the constructed deformation proxy model. However, on the one hand, the random sampling method used in the existing three-dimensional finite element simulation calculation sampling leads to uneven distribution of sample points due to too blind sampling, while the orthogonal test method sampling will still result in a large workload of three-dimensional finite element calculations due to too many sampling points; on the other hand, in the existing technology, quadratic or high-order polynomial functions are usually used when constructing the horizontal deformation proxy model of the pile foundation, and the number of variables processed is also less than 10. Therefore, for the reliability analysis of horizontal deformation of piles on multi-layer foundations with a large number of random variables (for example, up to 20 or even more), it is necessary to find a three-dimensional finite element numerical simulation calculation sampling method with good distribution characteristics and suitable for a large number of parameters and levels, as well as a deformation proxy model that can adapt to more variables.

[0009] In the prior art, point estimation method or direct Monte Carlo sampling method is used to calculate failure probability and reliability index. For the reliability analysis of pile deformation with many variables, high nonlinearity and usually very small failure probability, the calculation accuracy is very low. Therefore, it is necessary to seek a method for calculating failure probability and reliability index with higher accuracy. Summary of the invention

[0010] The present invention discloses a method, device and electronic equipment for calculating the reliability of horizontal deformation of pile foundations based on a proxy model. The method adopts a three-dimensional finite element method based on a rock and soil elastoplastic constitutive model to calculate the horizontal deformation of pile foundations. The nonlinear and elastoplastic characteristics of rock and soil bodies, complex foundation conditions such as multi-layers and inclinations, and various load conditions can be considered. The calculation results are more in line with the actual engineering situation, thereby effectively solving at least one technical problem involved in the background technology.

[0011] To achieve the above object, the technical solution of the present invention is:

[0012] A method for calculating the reliability of horizontal deformation of pile foundations comprises the following steps:

[0013] S110, collecting stratigraphic data, pile foundation data, load data, and pile-soil interface data of the project site and adjacent areas;

[0014] S120, determine the factors affecting the horizontal deformation of pile foundation and their value range based on research and analysis;

[0015] S130, a three-dimensional finite element numerical simulation calculation scheme for the horizontal deformation of pile foundations based on the uniform design method;

[0016] S140, establishing a three-dimensional finite element calculation model based on a three-dimensional finite element numerical simulation calculation scheme, and obtaining the horizontal deformation values ​​of the foundation piles under different factor levels by calculation;

[0017] S150, establishing a training sample data set based on the horizontal deformation value of the pile and the corresponding level value of each factor, and establishing a test sample data set by random sampling;

[0018] S160, by comparing multiple machine learning models, a high-precision proxy model of pile horizontal deformation is established;

[0019] S170, establishing a limit state equation for reliability calculation of a proxy model for pile horizontal deformation;

[0020] S180, based on the pile horizontal deformation proxy model and limit state equation, combined with the importance sampling method, calculates the pile horizontal deformation failure probability and reliability index.

[0021] As a preferred improvement of the present invention, the formation data includes the number of formation layers, formation inclination, and the weight, elastic modulus, Poisson's ratio, internal friction angle, and cohesion of the formation rock and soil; the pile foundation data includes the weight, elastic modulus, Poisson's ratio, diameter and length of the pile; the load data includes the vertical load size, horizontal load size and direction, bending moment size and direction; the pile-soil interface data includes the friction coefficient of the pile-soil interface.

[0022] As a preferred improvement of the present invention, step S120 specifically includes: first analyzing and determining the various factors that affect the horizontal deformation of the foundation piles, and then analyzing and determining the lower and upper limits of the values ​​of each influencing factor in the numerical simulation calculation based on the survey report and design data of the analyzed project, and in combination with the engineering survey and design data of adjacent areas in the area and regional empirical data, to determine the calculation value range.

[0023] As a preferred improvement of the present invention, step S130 specifically includes: determining the number of levels of each factor and the corresponding values ​​according to the number of factors and their value ranges determined in step S120; designing or selecting an existing uniform design table according to the number of factors and the number of levels; arranging the values ​​of each factor and its level number into the uniform design table to obtain a three-dimensional finite element numerical simulation calculation scheme for horizontal deformation of pile foundations.

[0024] As a preferred improvement of the present invention, in step S150, several groups of factor values ​​are randomly selected from the value range of each influencing factor to form several calculation schemes, and then the horizontal deformation value of the pile foundation is obtained by three-dimensional finite element calculation and used as the test sample data set.

[0025] As a preferred improvement of the present invention, step S160 includes the following steps:

[0026] S1601, standardizing the training sample data set and the test sample data set by transformation;

[0027] S1602, selecting several appropriate machine learning models, learning the standardized training sample data set, and obtaining corresponding pile horizontal deformation proxy models;

[0028] S1603, selecting a machine learning model with the highest comprehensive accuracy based on the fitting accuracy of the pile horizontal deformation proxy model obtained by training on the training sample data set and the prediction accuracy on the test sample data set;

[0029] S1604, mixing the training sample data set and the test sample data set, retraining the selected machine learning model, and obtaining a high-precision proxy model of pile horizontal deformation.

[0030] As a preferred improvement of the present invention, the expression of the limit state equation is:

[0031] Z = g(X) = [s h ]-s h (X)

[0032] Where Z is the state variable, Z>0 means the structure is in a reliable state, Z=0 means it is in a limit state, and Z<0 means it is in a failure state; X is a random vector composed of the factors affecting the horizontal deformation of the pile, X=[x1,x2,…,x n ]; g(X) is the performance function; [s h ] is the maximum allowable horizontal deformation of the pile; s h (X) is the calculated value of horizontal deformation of pile top.

[0033] As a preferred improvement of the present invention, step S180 specifically includes the following steps:

[0034] S1801, various factors affecting the horizontal deformation of pile foundations are considered as random variables, and the probability distribution, mean value and coefficient of variation of each random variable are determined by statistical analysis of engineering survey, design and regional experience data;

[0035] S1802, based on the limit state equation, the design verification point is obtained by using the verification point method;

[0036] S1803, based on the probability distribution and coefficient of variation of each random variable, taking the design verification point as the center of importance sampling, using the Monte Carlo method to randomly select N groups of random variable values, that is, calculating parameter values;

[0037] S1804, substituting the extracted N groups of calculation parameter values ​​into the pile foundation horizontal deformation proxy model to calculate and obtain N pile top horizontal deformation values, wherein the N groups of calculation parameter values ​​and their corresponding pile top horizontal deformation values ​​constitute N samples;

[0038] S1805, substitute N samples into the limit state equation in batches, and calculate the pile horizontal deformation failure probability P by combining the importance sampling calculation formula f And the reliability index β, the calculation expression is:

[0039]

[0040] β=-Φ -1 (P f )

[0041] In the formula, v i is the importance sampling random vector; g x (·) is the performance function; I(·) is the characteristic function; Z is the limit state variable; f x (·) is the original probability density function; P v (·) is the importance sampling probability density function; Φ -1 (·) is the inverse function of the standard normal distribution function.

[0042] The present invention also provides a device for calculating the reliability of horizontal deformation of pile foundation, comprising:

[0043] Data acquisition module, used to collect stratum data, pile data, load data and pile-soil interface data of the project site and adjacent areas;

[0044] Factor determination module, used to analyze and determine the various factors affecting the horizontal deformation of the pile foundation and their value ranges;

[0045] The calculation scheme design module is used to design the three-dimensional finite element numerical simulation calculation scheme of the horizontal deformation of the foundation pile based on the uniform design method;

[0046] A modeling and calculation module is used to establish a three-dimensional finite element calculation model based on a three-dimensional finite element numerical simulation calculation scheme, and obtain the horizontal deformation value of the foundation pile under different factor levels through calculation;

[0047] A sample set construction module is used to construct a training sample data set and a test sample data set;

[0048] The proxy model building module is used to build a high-precision proxy model of pile horizontal deformation by comparing multiple machine learning models;

[0049] The limit state equation construction module is used to construct the limit state equation for reliability calculation of the pile horizontal deformation proxy model based on the principle that the pile top horizontal deformation does not exceed the specified limit;

[0050] The reliability analysis module is used to calculate the failure probability and reliability index of the horizontal deformation of the pile based on the horizontal deformation proxy model and the limit state equation combined with the importance sampling method; the reliability analysis module includes:

[0051] A calculation parameter determination unit is used to determine various random variables affecting the horizontal deformation of the foundation pile and their probability distribution, mean value and coefficient of variation;

[0052] A sampling center determination unit is used to obtain a design verification point as the center of importance sampling by using a verification point method;

[0053] The calculation parameter sampling unit is used to randomly select N groups of calculation parameter values ​​using the Monte Carlo method based on the probability distribution and coefficient of variation of each random variable and taking the design verification point as the center of importance sampling;

[0054] The pile top horizontal deformation calculation unit is used to substitute the extracted N groups of calculation parameter values ​​into the pile foundation horizontal deformation proxy model, and calculate N pile top horizontal deformation values ​​to form N samples;

[0055] The failure probability and reliability index calculation unit is used to substitute N samples into the limit state equation in batches, and calculate the horizontal deformation failure probability and reliability index of the pile foundation in combination with the importance sampling calculation formula.

[0056] The present invention also provides an electronic device, comprising a memory and a processor, wherein a computer program is stored in the memory, and the computer program is read by the processor and executes the method when it is run.

[0057] The beneficial effects of the present invention are as follows:

[0058] 1. The three-dimensional finite element method based on the rock and soil elastoplastic constitutive model is used to calculate the horizontal deformation of the foundation pile. The nonlinear and elastoplastic characteristics of the rock and soil, complex foundation conditions such as multi-layer and inclined layers, and various load conditions can be considered. The calculation results are more in line with the actual engineering situation.

[0059] 2. Adopt machine learning models, such as multi-layer perceptron MLP, random forest RF, XGBoost and other machine learning models to build a proxy model for pile horizontal deformation, which has technical advantages such as good fault tolerance, strong adaptability, high fitting and prediction accuracy, and solves the technical problem of difficulty in building a high-precision proxy model for pile deformation due to too many input variables;

[0060] 3. The uniform test design method is used to design the three-dimensional finite element numerical simulation calculation scheme for pile deformation. According to the simulation calculation results, a sample database of pile horizontal deformation is constructed to provide training data samples for establishing a proxy model of pile horizontal deformation. Compared with the existing random sampling and orthogonal test method sampling, the uniform design not only has the advantages of uniform distribution of test points and a small number of tests, but is also particularly suitable for multi-factor and multi-level test designs. Therefore, it can not only ensure uniform sampling points, but also reduce the amount of calculation and improve calculation efficiency.

[0061] 4. The value range of each factor in the uniform design is obtained by statistical analysis based on the test data of all exploration points in a certain engineering site, which breaks through the limitation of the existing technology that usually only takes values ​​according to the statistical values ​​of the geotechnical parameters of the exploration points at the analyzed pile foundation according to the 3σ principle, and overcomes the defect that the test results cannot reflect the overall picture of the site due to the single sampling location and small number of samples;

[0062] 5. The direct importance sampling method is used to calculate the failure probability and reliability index of the horizontal deformation of the foundation pile, which solves the problem that the existing point estimation method and direct Monte Carlo sampling method will cause distortion of the calculation results when calculating the reliability index of the foundation pile deformation with a very small failure probability, and improves the calculation accuracy and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work, among which:

[0064] Figure 1 It is a schematic flow chart of a method for calculating the reliability of horizontal deformation of pile foundation according to the present invention;

[0065] Figure 2 is a specific flow chart of step S180;

[0066] Figure 3 A structural frame diagram of a pile foundation horizontal deformation reliability calculation device according to the present invention;

[0067] Figure 4 This is the structural framework diagram of the reliability analysis module;

[0068] Figure 5 A structural framework diagram of an electronic device of the present invention;

[0069] Figure 6 A schematic diagram of a three-dimensional finite element calculation model provided in Example 1;

[0070] Figure 7 This is a schematic diagram of the three-dimensional finite element calculation model provided in Example 2. DETAILED DESCRIPTION

[0071] The technical solutions in the embodiments of the present invention will be described clearly and completely below in combination with the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0073] In addition, in the present invention, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0074] In the present invention, unless otherwise clearly specified and limited, the terms "connection", "fixation", etc. should be understood in a broad sense. For example, "fixation" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0075] In addition, the technical solutions between the various embodiments of the present invention can be combined with each other, but it must be based on the fact that ordinary technicians in the field can implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0076] like Figure 1 and Figure 2 As shown, the present invention provides a method for calculating the reliability of horizontal deformation of pile foundation, comprising the following steps:

[0077] S110, collect stratigraphic data, pile data, load data and pile-soil interface data of the project and site under study, as well as the projects and sites in the surrounding area;

[0078] Among them, the formation data include the number of formation layers, formation inclination, and the weight, elastic modulus, Poisson's ratio, internal friction angle, cohesion and other data of the formation rock and soil; the pile foundation data include the weight, elastic modulus, Poisson's ratio, diameter and length and other data of the pile; the load data include the vertical load size, horizontal load size and direction, bending moment size and direction and other data; the pile-soil interface data include the friction coefficient of the pile-soil interface.

[0079] By extensively collecting information data on various factors that may affect the horizontal deformation of pile foundations, the information is made more comprehensive.

[0080] S120, determine the factors affecting the horizontal deformation of pile foundation and their value range based on research and analysis;

[0081] Specifically, we first analyze and determine the various factors that affect the horizontal deformation of the foundation piles, and then, based on the survey report and design data of the analyzed project, combined with the engineering survey and design data of the adjacent areas in the area and regional experience data, analyze and determine the lower and upper limits of the values ​​of each influencing factor in the numerical simulation calculation, thereby determining the calculation value range.

[0082] The value range of each factor is obtained by statistical analysis based on the test data of all exploration points in a certain engineering site, which breaks through the limitation of the existing technology that usually only takes values ​​according to the statistical values ​​of the geotechnical parameters of the exploration points at the analyzed pile foundation according to the 3σ principle, and overcomes the defect that the test results cannot reflect the overall picture of the site due to the single sampling location and small number of samples; and because the influencing factors are comprehensively considered and the value range is wide, it can be guaranteed that the deformation proxy model established subsequently is more in line with reality and has a larger scope of application.

[0083] S130, a three-dimensional finite element numerical simulation calculation scheme for the horizontal deformation of pile foundations based on the uniform design method;

[0084] First, the number of levels of each factor and the corresponding values ​​are determined according to the number of factors and their value ranges determined in step S120, and it is better to control the number of levels to 5 to 9 levels; then, according to the number of factors and the number of levels, a more suitable uniform design table is designed or selected; finally, each factor and its level value are arranged into the uniform design table, so as to obtain a three-dimensional finite element numerical simulation calculation scheme for the horizontal deformation of the foundation pile.

[0085] In this step, the uniform experimental design method is used to arrange the three-dimensional finite element numerical simulation calculation scheme for the horizontal deformation of pile foundations. Even when the number of factors and the number of levels are large, the distribution of sampling points can be uniform and the number of sampling times can be small, thereby reducing the number of three-dimensional finite element simulation calculations and improving the calculation efficiency, providing a reliable technical guarantee for the subsequent establishment of a proxy model of the horizontal deformation of pile foundations that is more in line with the actual engineering situation.

[0086] S140, establishing a three-dimensional finite element calculation model based on a three-dimensional finite element numerical simulation calculation scheme, wherein the rock and soil body adopts an elastic-plastic constitutive model, and the horizontal deformation values ​​of the foundation piles under different factor levels are obtained by calculation;

[0087] S150, based on the pile horizontal deformation value and its corresponding factor level value, a training sample data set is established, several groups of factor values ​​are randomly selected from the value range of each influencing factor to form several calculation schemes, and then the pile horizontal deformation value is obtained by three-dimensional finite element calculation, which is used as a test sample data set. Preferably, the size of the test sample data set is 10% to 20% of the training sample data set.

[0088] In this step, since the calculation scheme is obtained according to the uniform design, all sample data obtained by simulation and calculation according to the calculation scheme are used as the training sample data set, which ensures the uniformity of the sample points and improves the realism of the deformable proxy model established subsequently.

[0089] S160, by comparing and selecting multiple machine learning models, a high-precision pile horizontal deformation proxy model is established; specifically, the following steps are included:

[0090] S1601, standardizing the training sample data set and the test sample data set by transformation to eliminate the influence of different data dimensions and large size differences on modeling;

[0091] S1602, selecting several appropriate machine learning models, such as multi-layer perceptron (MLP), random forest (RF), BP neural network, gradient boosting decision tree (GBDT), XGBoost and other models, learning the standardized training sample data set, and obtaining the corresponding pile horizontal deformation proxy model;

[0092] S1603, selecting a machine learning model with the highest comprehensive accuracy based on the fitting accuracy of the pile horizontal deformation proxy model obtained by training on the training sample data set and the prediction accuracy on the test sample data set;

[0093] S1604, the training sample data set and the test sample data set are mixed, and the selected machine learning model is retrained to obtain a high-precision pile horizontal deformation proxy model for subsequent pile horizontal deformation reliability calculation.

[0094] In this step, the direct call of the three-dimensional finite element model based on the elastic-plastic constitutive model of rock and soil is replaced by the call of the deformation proxy model to calculate the horizontal deformation of the foundation pile, which greatly improves the calculation efficiency of the foundation pile deformation.

[0095] S170, establish the limit state equation for the reliability calculation of the horizontal deformation proxy model of the pile foundation. According to the engineering requirements, the horizontal deformation of the pile foundation should not exceed the specified limit [s h ]; the expression of the limit state equation is:

[0096] Z = g(X) = [s h ]-s h (X)

[0097] Where Z is the state variable, Z>0 means the structure is in a reliable state, Z=0 means it is in a limit state, and Z<0 means it is in a failure state; X is a random vector composed of the factors affecting the horizontal deformation of the pile, X=[x1,x2,…,x n ]; g(X) is the performance function; [s h ] is the maximum allowable horizontal deformation value of the foundation pile, which is determined according to relevant industry standards; s h (X) is the calculated value of the horizontal deformation of the pile top, which is calculated by a high-precision proxy model of the pile horizontal deformation.

[0098] S180, based on the pile horizontal deformation proxy model and the limit state equation, combined with the importance sampling method, calculates the pile horizontal deformation failure probability and reliability index, specifically including the following steps:

[0099] S1801, various factors affecting the horizontal deformation of pile foundations are considered as random variables, and the probability distribution, mean value and coefficient of variation of each random variable are determined by statistical analysis of engineering survey, design and regional experience data;

[0100] S1802, based on the limit state equation, the design verification point is obtained by using the verification point method (i.e., JC method), and the design verification point is used as the center of importance sampling;

[0101] S1803, based on the probability distribution form and coefficient of variation of each random variable, taking the design verification point as the sampling center, using the Monte Carlo method to randomly select N groups of random variable values, that is, calculating the parameter value;

[0102] S1804, substituting the extracted N groups of calculation parameter values ​​into a high-precision pile foundation horizontal deformation proxy model to calculate and obtain N pile top horizontal deformation values, so that the N groups of calculation parameter values ​​and their corresponding pile top horizontal deformation values ​​constitute N samples;

[0103] S1805, substitute N samples into the limit state equation in batches, and calculate the pile horizontal deformation failure probability P by combining the importance sampling calculation formula f And the reliability index β, the calculation expression is:

[0104]

[0105] β=-Φ -1 (P f )

[0106] In the formula, v i is the importance sampling random vector; g x (·) is the performance function; I(·) is the characteristic function; Z is the limit state variable; f x (·) is the original probability density function; P v (·) is the importance sampling probability density function, which can take the same form as the original probability density function; Φ -1 (·) is the inverse function of the standard normal distribution function.

[0107] In this step, since the direct importance sampling method converts the sampling center to the design verification point on the limit state surface, this reduces the variance of the failure probability estimate and improves the calculation accuracy of the failure probability.

[0108] like Figure 3 As shown, the present invention also provides a device for calculating the reliability of horizontal deformation of pile foundation, comprising:

[0109] The data acquisition module 310 is used to collect the stratum data, pile foundation data, load data and pile-soil interface data of the engineering site and the adjacent area;

[0110] The factor determination module 320 is used to analyze and determine various factors affecting the horizontal deformation of the pile foundation and their value ranges;

[0111] A calculation scheme design module 330 is used to design a three-dimensional finite element numerical simulation calculation scheme for horizontal deformation of pile foundations based on a uniform design method;

[0112] A modeling and calculation module 340 is used to establish a three-dimensional finite element calculation model based on a three-dimensional finite element numerical simulation calculation scheme, and obtain the horizontal deformation values ​​of the foundation piles under different factor levels by calculation;

[0113] A sample set construction module 350 is used to construct a training sample data set and a test sample data set;

[0114] The proxy model building module 360 ​​is used to build a high-precision proxy model of pile horizontal deformation by comparing multiple machine learning models;

[0115] The limit state equation construction module 370 is used to construct the limit state equation for calculating the reliability of the pile horizontal deformation proxy model based on the principle that the pile top horizontal deformation does not exceed the specified limit;

[0116] Reliability analysis module 380 is used to calculate the failure probability and reliability index of pile horizontal deformation based on the pile horizontal deformation proxy model and limit state equation combined with importance sampling method; Figure 4 As shown, the reliability analysis module 380 includes:

[0117] The calculation parameter determination unit 381 is used to determine various random variables affecting the horizontal deformation of the pile foundation and their probability distribution, mean value and coefficient of variation;

[0118] A sampling center determination unit 382 is used to obtain a design verification point as the center of importance sampling by using a verification point method;

[0119] The calculation parameter sampling unit 383 is used to randomly select N groups of calculation parameter values ​​using the Monte Carlo method based on the probability distribution and coefficient of variation of each random variable and taking the design verification point as the sampling center;

[0120] The pile top horizontal deformation calculation unit 384 is used to substitute the extracted N groups of calculation parameter values ​​into the pile foundation horizontal deformation proxy model, and calculate N pile top horizontal deformation values ​​to form N samples;

[0121] The failure probability and reliability index calculation unit 385 is used to substitute N samples into the limit state equation in batches, and calculate the horizontal deformation failure probability and reliability index of the pile foundation in combination with the importance sampling calculation formula.

[0122] like Figure 5 As shown, the present invention further provides an electronic device 100, including a memory 101 and a processor 102, wherein the memory 101 stores a computer program, and the computer program is read by the processor 102 and executes the method described when it is run.

[0123] Embodiment 1

[0124] This embodiment provides a method for calculating the reliability of horizontal deformation of piles in a horizontal layered foundation based on a machine agent model, comprising the following steps:

[0125] S110, for a bridge pile foundation project in a horizontal layered foundation, collect stratum data, pile foundation data, load data and pile-soil interface data; the stratum data include the number of stratum layers, stratum inclination, and the weight, elastic modulus, Poisson's ratio, internal friction angle and cohesion of the stratum rock and soil, wherein the number of stratum layers is 2 and the stratum inclination is 0°; the pile foundation data include the weight, elastic modulus, Poisson's ratio, diameter and length of the pile; the load data include the vertical load size, horizontal load size and direction, bending moment size and direction; the pile-soil interface data include the friction coefficient of the pile-soil interface.

[0126] S120, through analysis, it was determined that there are 21 factors that affect the horizontal deformation of the foundation piles. At the same time, the engineering survey report and design data were comprehensively analyzed, and combined with the engineering survey and design data of the adjacent areas of the region and the regional experience data, the lower and upper limits of the values ​​of each influencing factor in the numerical simulation calculation were analyzed and determined, thereby obtaining the calculation value range shown in Table 1 below.

[0127] Table 1 Influencing factors and value ranges of pile deformation in double-layer foundation

[0128]

[0129]

[0130] S130, a three-dimensional finite element numerical simulation calculation scheme for the horizontal deformation of pile foundations based on the uniform experimental design method;

[0131] In order to obtain more comprehensive calculation data reflecting the relationship between various influencing factors and pile deformation with fewer tests, the uniform design method is combined with the finite element method. Nine levels are taken for the 21 determined influencing factors. The uniform test table U is used to calculate the relationship between the influencing factors and the pile deformation. 108 (9 21 )(where U represents uniform design, 108 is the number of tests, 9 is the number of levels, and 21 is the number of factors) a numerical simulation calculation scheme for the deformation of bridge piles in double-layer foundation was designed, as shown in Table 2 below.

[0132] Table 2 Numerical simulation calculation scheme for horizontal deformation of foundation piles in double-layer foundation

[0133]

[0134]

[0135] S140, establishing a three-dimensional finite element calculation model for the 108 numerical simulation calculation schemes designed in step S130, such as Figure 6 As shown, the rock and soil body adopts an elastic-plastic constitutive model, and the horizontal deformation values ​​of the piles under different factor levels (i.e. different rock and soil parameters and pile design parameter values) are obtained by calculation;

[0136] S150, 108 samples are formed by the horizontal deformation values ​​of the pile foundation and the corresponding level values ​​of each factor as the training sample data set; at the same time, 22 groups of factor values ​​are randomly selected from the value range of each influencing factor to form 22 calculation schemes, and then the horizontal deformation values ​​of the pile foundation are obtained by three-dimensional finite element calculation, which are used as the test sample data set, wherein the size of the test sample data set is approximately 20% of the training sample data set.

[0137] S160, establishing a high-precision pile horizontal deformation proxy model based on a machine learning model, specifically comprising the following steps:

[0138] S1601, perform standardization processing on the training sample data set and the test sample data set according to the following formula:

[0139]

[0140] In the formula, X ij represents the standardized data value of the jth level of the i-th factor; x ij Represents the original data value of the jth level of the i-th factor in the database; μ i represents the mean of the i-th factor; σ i represents the standard deviation of the ith factor;

[0141] S1602, input the standardized training sample data set into the alternative multi-layer perceptron MLP and random forest RF machine learning models for learning. The fitting performance of the trained MLP and RF models on the training set and the prediction performance on the test set are shown in Table 3 below.

[0142] Table 3 Performance evaluation results of pile horizontal deformation proxy model

[0143]

[0144]

[0145] From Table 3, we can see that the comprehensive performance of the multi-layer perceptron MLP is the best, so the multi-layer perceptron MLP is selected to establish the horizontal deformation proxy model of the foundation pile;

[0146] S1603, mixing the 108 training sample data with the 22 test sample data, and retraining the multi-layer perceptron MLP to obtain a high-precision pile horizontal deformation proxy model for subsequent pile horizontal deformation reliability calculation.

[0147] S170, establish the limit state equation for reliability calculation of pile horizontal deformation proxy model, according to which the pile horizontal deformation should not exceed the specified limit [s h ], establish the limit state equation as follows:

[0148] Z = g(X) = [s h ]-s h (X)

[0149] Where Z is the state variable, Z>0 means the structure is in a reliable state, Z=0 means it is in a limit state, and Z<0 means it is in a failure state; X is a random vector composed of the factors affecting the horizontal deformation of the pile, X=[x1,x2,…,x n ]; g(X) is the performance function; [s h ] is the maximum allowable horizontal deformation of the pile foundation, which is determined according to relevant industry standards and can be taken as [s h ]=6mm~10mm;s h (X) is the calculated value of the horizontal deformation of the pile top, which is calculated by the MLP pile horizontal deformation proxy model based on the multi-layer perceptron constructed after comparison.

[0150] S180, calculating the failure probability and reliability index of the horizontal deformation of the pile foundation, specifically including the following steps:

[0151] S1801, the 21 factors that affect the horizontal deformation of pile foundations are regarded as random variables. By statistically analyzing the engineering survey, design and regional experience data, the probability distribution, mean and coefficient of variation of each random variable are determined, as shown in Table 4 below.

[0152] Table 4 Statistical characteristics of calculation parameters

[0153]

[0154]

[0155] S1802, based on the limit state equation, the design verification point is obtained by using the verification point method (i.e., JC method), and the design verification point is used as the center of importance sampling;

[0156] S1803, based on the probability distribution form and coefficient of variation of each random variable, taking the design verification point as the sampling center, using the Monte Carlo method to randomly select N groups of random variable values, that is, calculating the parameter value;

[0157] S1804, substituting the extracted N groups of calculation parameter values ​​into a high-precision pile foundation horizontal deformation proxy model to calculate and obtain N pile top horizontal deformation values, so that the N groups of calculation parameter values ​​and their corresponding pile top horizontal deformation values ​​constitute N samples;

[0158] S1805, substitute N samples into the limit state equation in batches, and combine with the importance sampling calculation formula to calculate the piles at the allowable horizontal deformation value [s h] is 6mm, 7mm, 8mm, 9mm and 10mm when the horizontal deformation failure probability P f And the reliability index β, as shown in Table 5.

[0159] Table 5 Failure probability of pile horizontal deformation P f and reliability index β

[0160]

[0161]

[0162] Embodiment 2

[0163] This embodiment provides a method for calculating the reliability of horizontal deformation of piles in an inclined foundation based on a machine proxy model, comprising the following steps:

[0164] S110, for a bridge pile foundation project in a certain inclined foundation, collect stratum data, pile foundation data, load data and pile-soil interface data; the stratum data include the number of stratum layers, stratum inclination, and the weight, elastic modulus, Poisson's ratio, internal friction angle and cohesion of the stratum rock and soil, where the number of stratum layers is 1; the pile foundation data include the weight, elastic modulus, Poisson's ratio, diameter and length of the pile; the load data include the vertical load size, horizontal load size and direction, bending moment size and direction; the pile-soil interface data include the friction coefficient of the pile-soil interface.

[0165] S120, through analysis, it is determined that there are 15 factors that affect the horizontal deformation of the foundation piles. At the same time, the engineering survey report and design data are comprehensively analyzed, and combined with the engineering survey and design data of the adjacent areas of the area and the regional experience data, the lower and upper limits of the values ​​of each influencing factor are analyzed and determined, thereby obtaining the calculation value range shown in the following Table 6.

[0166] Table 6 Factors affecting pile deformation in inclined foundation and their value ranges

[0167]

[0168]

[0169] S130, a three-dimensional finite element numerical simulation calculation scheme for the horizontal deformation of pile foundations based on the uniform experimental design method;

[0170] In order to obtain more comprehensive calculation data reflecting the relationship between various influencing factors and pile deformation with fewer tests, the uniform design method is combined with the finite element method. Seven levels are taken for the 15 determined influencing factors. The uniform test table U is used to calculate the relationship between the influencing factors and the pile deformation. 70 (7 15)(where U represents the uniform design table, 70 represents the required number of tests, 7 represents the number of levels, and 15 represents the number of factors) a numerical simulation calculation scheme for the deformation of bridge piles in inclined foundation is designed, as shown in Table 7 below.

[0171] Table 7 Numerical simulation calculation scheme for horizontal deformation of piles in inclined foundation

[0172]

[0173]

[0174] S140, establishing a three-dimensional finite element calculation model for the 70 numerical simulation calculation schemes designed in step S130, such as Figure 7 As shown, the rock and soil body adopts an elastic-plastic constitutive model, and the horizontal deformation values ​​of the piles under different factor levels (i.e. different rock and soil parameters and pile design parameter values) are obtained by calculation;

[0175] S150, 70 samples consisting of the horizontal deformation values ​​of the pile foundation and the corresponding level values ​​of each factor are used as the training sample data set; at the same time, 14 groups of factor values ​​are randomly selected from the value range of each influencing factor to form 14 calculation schemes, and then the horizontal deformation values ​​of the pile foundation are obtained by three-dimensional finite element calculation, which are used as the test sample data set, wherein the size of the test sample data set is approximately 20% of the training sample data set.

[0176] S160, establishing a high-precision pile horizontal deformation proxy model based on a machine learning model, specifically comprising the following steps:

[0177] S1601, normalizing the training sample data set and the test sample data set using a minimum-maximum scaling method (Min-Max Scaling);

[0178] S1602, input the normalized training sample data set into the candidate BP neural network model, gradient boosted decision tree (GBDT) model, random forest (RF) model and XGBoost model based on Bayesian optimization; the determination coefficient R of the four machine learning models constructed 2 They are 0.678, 0.370, 0.433 and 0.982 respectively, among which the XGBoost model based on Bayesian optimization has the highest accuracy (R 2 =0.982), so the XGBoost model based on Bayesian optimization was selected to establish the horizontal deformation proxy model of pile foundation;

[0179] S1603, mixing the 70 training sample data with the 14 test sample data, and retraining the XGBoost model based on Bayesian optimization to obtain a high-precision pile horizontal deformation proxy model for subsequent pile horizontal deformation reliability calculation.

[0180] S170, establish the limit state equation for reliability calculation of pile horizontal deformation proxy model, according to which the pile horizontal deformation should not exceed the specified limit [s h ], establish the limit state equation as follows:

[0181] Z = g(X) = [s h ]-s h (X)

[0182] Where Z is the state variable, Z>0 means the structure is in a reliable state, Z=0 means it is in a limit state, and Z<0 means it is in a failure state; X is a random vector composed of the factors affecting the horizontal deformation of the pile, X=[x1,x2,…,x n ]; g(X) is the performance function; [s h ] is the maximum allowable horizontal deformation of the pile foundation, which is determined according to relevant industry standards and can be taken as [s h ]=6mm~10mm;s h (X) is the calculated value of the horizontal deformation of the pile top, which is calculated by the XGBoost horizontal deformation proxy model based on Bayesian optimization constructed after comparison.

[0183] S180, calculating the failure probability and reliability index of the horizontal deformation of the pile foundation, specifically including the following steps:

[0184] S1801, the 15 factors that affect the horizontal deformation of pile foundations are regarded as random variables. Through statistical analysis of engineering survey, design and regional experience data, the probability distribution, mean and coefficient of variation of each random variable are determined, as shown in Table 8 below.

[0185] Table 8 Statistical characteristics of calculation parameters

[0186] S1802, based on the limit state equation, the design verification point is obtained by using the verification point method (i.e., JC method), and the design verification point is used as the center of importance sampling;

[0187] S1803, based on the probability distribution form and coefficient of variation of each random variable, taking the design verification point as the sampling center, using the Monte Carlo method to randomly select N groups of random variable values, that is, calculating the parameter value;

[0188] S1804, substituting the extracted N groups of calculation parameter values ​​into a high-precision pile foundation horizontal deformation proxy model to calculate and obtain N pile top horizontal deformation values, so that the N groups of calculation parameter values ​​and their corresponding pile top horizontal deformation values ​​constitute N samples;

[0189] S1805, substitute N samples into the limit state equation in batches, and combine with the importance sampling calculation formula to calculate the allowable horizontal deformation value of the pile [s h ]=6mm horizontal deformation failure probability P f =4.9×10 -4 and reliability index β = 3.3.

[0190] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A method for calculating the reliability of horizontal deformation of pile foundation, characterized in that: The following steps are involved: S110, collecting stratigraphic data, pile foundation data, load data, and pile-soil interface data of the project site and adjacent areas; S120, determine the factors affecting the horizontal deformation of pile foundation and their value range based on research and analysis; S130, designing a three-dimensional finite element numerical simulation calculation scheme for horizontal deformation of pile foundations based on the uniform design method; step S130 specifically includes: determining the number of levels and corresponding values ​​of each factor according to the number of factors and their value ranges determined in step S120; designing or selecting an existing uniform design table according to the number of factors and the number of levels; arranging the values ​​of each factor and its number of levels into the uniform design table, and obtaining a three-dimensional finite element numerical simulation calculation scheme for horizontal deformation of pile foundations; S140, establishing a three-dimensional finite element calculation model based on a three-dimensional finite element numerical simulation calculation scheme, and obtaining the horizontal deformation values ​​of the foundation piles under different factor levels by calculation; S150, establishing a training sample data set based on the horizontal deformation value of the pile and the corresponding level value of each factor, and establishing a test sample data set by random sampling; S160, by comparing multiple machine learning models, a high-precision proxy model of pile horizontal deformation is established; S170, establishing a limit state equation for reliability calculation of a proxy model for pile horizontal deformation; S180, based on the pile horizontal deformation proxy model and the limit state equation, combined with the importance sampling method, calculate the pile horizontal deformation failure probability and reliability index; step S180 specifically includes the following steps: S1801, various factors affecting the horizontal deformation of pile foundations are considered as random variables, and the probability distribution, mean value and coefficient of variation of each random variable are determined by statistical analysis of engineering survey, design and regional experience data; S1802, based on the limit state equation, the design verification point is obtained by using the verification point method; S1803, based on the probability distribution and coefficient of variation of each random variable, the design verification point is used as the center of importance sampling, and the Monte Carlo method is used to randomly select N Set random variable values, that is, calculate parameter values; S1804, the extracted N Substitute the calculated parameter values ​​into the pile horizontal deformation proxy model and calculate N The horizontal deformation value of the pile top is N The group calculation parameter value and its corresponding pile top horizontal deformation value are composed of N samples; S1805, will N Substitute the samples into the limit state equation in batches and calculate the probability of failure of pile horizontal deformation by combining the importance sampling calculation formula. P f and reliability indicators β , the calculation expression is: ； In the formula, v i Sampling random vectors for importance; is a functional function; is the indicative function; Z is the limit state variable; is the original probability density function; is the importance sampling probability density function; is the inverse function of the standard normal distribution function.

2. A pile foundation horizontal deformation reliability calculation method according to claim 1, characterized in that: The stratum data include the number of stratum layers, stratum inclination, and the weight, elastic modulus, Poisson's ratio, internal friction angle, and cohesion of the stratum rock and soil; the foundation pile data include the weight, elastic modulus, Poisson's ratio, diameter, and length of the pile; the load data include the magnitude of the vertical load, the magnitude and direction of the horizontal load, and the magnitude and direction of the bending moment; the pile-soil interface data include the friction coefficient of the pile-soil interface.

3. A pile foundation horizontal deformation reliability calculation method according to claim 1, characterized in that: Step S120 specifically includes: first analyzing and determining the various factors that affect the horizontal deformation of the foundation piles, and then analyzing and determining the lower and upper limits of the values ​​of each influencing factor in the numerical simulation calculation based on the survey report and design data of the analyzed project, and combining the engineering survey and design data of the adjacent areas of the region and regional empirical data, and determining the calculation value range.

4. A method for calculating the reliability of horizontal deformation of pile foundation according to claim 1, characterized in that: In step S150, several groups of factor values ​​are randomly selected from the value range of each influencing factor to form several calculation schemes, and then the horizontal deformation value of the pile foundation is obtained by three-dimensional finite element calculation and used as the test sample data set.

5. The method for calculating the reliability of pile horizontal deformation according to claim 1 is characterized in that: Step S160 includes the following steps: S1601, standardizing the training sample data set and the test sample data set by transformation; S1602, selecting several appropriate machine learning models, learning the standardized training sample data set, and obtaining corresponding pile horizontal deformation proxy models; S1603, selecting a machine learning model with the highest comprehensive accuracy based on the fitting accuracy of the pile horizontal deformation proxy model obtained by training on the training sample data set and the prediction accuracy on the test sample data set; S1604, mixing the training sample data set and the test sample data set, retraining the selected machine learning model, and obtaining a high-precision proxy model of pile horizontal deformation.

6. A method for calculating the reliability of horizontal deformation of pile foundation according to claim 1, characterized in that: The expression of the limit state equation is: ； Where Z is the state variable, Z >0 means the structure is in a reliable state. Z =0 means it is in the limit state. Z <0 means it is in failure state; X is a random vector composed of the factors affecting the horizontal deformation of the pile foundation, X =[ x 1, x 2, …, x n ]; g ( X ) is a functional function; [ s h ] is the maximum allowable horizontal deformation of the pile; s h ( X ) is the calculated value of the horizontal deformation of the pile top.

7. A device for calculating the reliability of pile horizontal deformation for implementing the method described in any one of claims 1 to 6, characterized in that: include: Data acquisition module, used to collect stratum data, pile data, load data and pile-soil interface data of the project site and adjacent areas; Factor determination module, used to analyze and determine the various factors affecting the horizontal deformation of the pile foundation and their value ranges; The calculation scheme design module is used to design the three-dimensional finite element numerical simulation calculation scheme of the horizontal deformation of the foundation pile based on the uniform design method; A modeling and calculation module is used to establish a three-dimensional finite element calculation model based on a three-dimensional finite element numerical simulation calculation scheme, and obtain the horizontal deformation value of the foundation pile under different factor levels through calculation; A sample set construction module is used to construct a training sample data set and a test sample data set; The proxy model building module is used to build a high-precision proxy model of pile horizontal deformation by comparing multiple machine learning models; The limit state equation construction module is used to construct the limit state equation for reliability calculation of the pile horizontal deformation proxy model based on the principle that the pile top horizontal deformation does not exceed the specified limit; Reliability analysis module, which is used to calculate the failure probability and reliability index of pile horizontal deformation based on the pile horizontal deformation proxy model and limit state equation combined with the importance sampling method; The reliability analysis module includes: A calculation parameter determination unit is used to determine various random variables affecting the horizontal deformation of the foundation pile and their probability distribution, mean value and coefficient of variation; A sampling center determination unit is used to obtain a design verification point as the center of importance sampling by using a verification point method; The calculation parameter sampling unit is used to randomly select the design verification point based on the probability distribution and coefficient of variation of each random variable, using the Monte Carlo method to randomly select the design verification point as the center of importance sampling. N Group calculation parameter values; The pile top horizontal deformation calculation unit is used to extract N Substitute the calculated parameter values ​​into the horizontal deformation proxy model of the pile foundation to obtain N The horizontal deformation value of the pile top is N samples; Failure probability and reliability index calculation unit, used to N The samples are batch-substituted into the limit state equation, and the failure probability of horizontal deformation of the pile and the reliability index are calculated in combination with the importance sampling calculation formula.

8. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the computer program is read by the processor and executes the method according to any one of claims 1 to 6 when executed.

Images