Calculation method and system of ultimate bearing capacity of pile based on BP neural network

Abstract

The present invention provides a method for calculating the ultimate bearing capacity of a pile body based on a BP neural network, and belongs to the technical field of geotechnical material data simulation. The method includes: determining the sensitivity parameters that significantly affect the bearing settlement of the foundation pile; collecting the sensitivity parameters and the ultimate bearing capacity of the pile foundation under the corresponding sensitivity parameters, and constructing training samples; constructing and using the sample data set to train the BP neural network model; collecting the sensitivity parameters of the pile foundation to be calculated, and inputting them into the ultimate bearing capacity prediction model to obtain the ultimate bearing capacity of the pile body. The present invention collects the sensitivity parameters and the ultimate bearing capacity of the pile foundation under the corresponding sensitivity parameters to construct training samples, and the formed sample data set is more accurate. The BP neural network model is constructed and trained using the sample data set, so that the BP neural network model is more adapted to the ultimate bearing capacity to obtain the ultimate bearing capacity prediction model. The ultimate bearing capacity prediction model is used to predict the ultimate bearing capacity of the pile foundation to be calculated, providing a reference basis for actual engineering construction.

Description

Technical Field

[0001] The invention relates to the technical field of geotechnical material data simulation, and in particular to a method for calculating the ultimate bearing capacity of a pile based on a BP neural network. Background Art

[0002] Physical model tests are usually used to simplify and abstractly analyze complex projects, which can help capture the main characteristics of engineering phenomena and better understand and explain actual projects. In actual projects, the karst geological conditions often contain multiple caves underground.

[0003] With the rapid development of numerical simulation, numerical analysis software has been increasingly used in the engineering field and plays an important role. Subsequently, various numerical calculation software has been updated and upgraded due to the emergence of the finite element method, and the role of numerical simulation technology in construction engineering has been further highlighted. The basic idea of ​​numerical simulation is to divide the soil and pile into grid units and set them by assigning mechanical factors. The advantages are fast, convenient and not affected by the test site.

[0004] However, the feasibility and effectiveness of the calculation results are largely related to the correct selection of rock mass mechanical factors. Geotechnical materials have typical inhomogeneity and anisotropy, coupled with the subjectivity of human operation and objective interference in the process of model testing. The combined effect of these influences makes it inevitable that there will be a certain deviation between the actual mechanical factors of the model test materials and the expected design values. Before carrying out systematic numerical simulation analysis, it is necessary to have an accurate grasp of the real physical and mechanical factors of the rock mass, which is a key step to ensure the accuracy of the simulation analysis results. Summary of the invention

[0005] The present invention provides a method for calculating the ultimate bearing capacity of a pile based on a BP neural network, which is used to solve the defects of the numerical simulation technology in the prior art in construction engineering and improve the accuracy of simulation analysis results.

[0006] The invention provides a method for calculating the ultimate bearing capacity of a pile body based on a BP neural network, comprising: determining, through an engineering test, a sensitivity parameter that significantly affects the bearing and settlement of a foundation pile; collecting the sensitivity parameter and the ultimate bearing capacity of the pile foundation under the corresponding sensitivity parameter, constructing a training sample, and forming a sample data set; constructing a BP neural network model and using the sample data set for training, saving the training parameters, and obtaining an ultimate bearing capacity prediction model; collecting the sensitivity parameters of a pile foundation to be calculated, inputting them into the ultimate bearing capacity prediction model, and obtaining the ultimate bearing capacity of the pile body; wherein, through finite element simulation of engineering pile data, forming a sample data set comprises: determining the sensitivity parameter method that significantly affects the bearing and settlement of the foundation pile as follows: determining pile foundation parameters; designing an orthogonal test table according to the pile foundation parameters; conducting a pile foundation settlement test or conducting a finite element analysis on the pile foundation settlement according to the orthogonal test table, and recording the pile foundation settlement; analyzing factors affecting the pile foundation settlement, and obtaining the sensitivity parameter.

[0007] According to a method for calculating the ultimate bearing capacity of a pile based on a BP neural network provided by the present invention, a range analysis method is used to determine factors affecting pile foundation settlement and obtain sensitivity parameters.

[0008] According to a method for calculating the ultimate bearing capacity of a pile based on a BP neural network provided by the present invention, the pile foundation parameters include elastic modulus, cohesion, Poisson's ratio, internal friction angle, and expansion angle.

[0009] According to a method for calculating the ultimate bearing capacity of a pile based on a BP neural network provided by the present invention, the method for constructing a training sample is as follows: using engineering pile tests or finite element simulation analysis to obtain the ultimate bearing capacity of the pile foundation under sensitivity parameters; using the sensitivity parameters as input and the ultimate bearing capacity of the pile foundation as a label to construct a training sample.

[0010] According to a method for calculating the ultimate bearing capacity of a pile based on a BP neural network provided by the present invention, a method for constructing a BP neural network model and training it using a sample data set is as follows: constructing a BP neural network model, determining the number of nodes in the input layer n , the number of hidden layer nodes l , the number of nodes in the output layer is m, and the connection weights between the input layer and the hidden layer are , the connection weights between the hidden layer and the output layer , the threshold a of the hidden layer, the threshold b of the output layer are initialized to obtain the initial structure of the BP neural network model; the sample data set is input into the BP neural network model, and the weights are connected and threshold To calculate the input value from the hidden layer to the output layer;

[0011]

[0012] Where: represents the threshold from the input layer to the jth hidden layer node; f is the transfer function; according to the output value H of the hidden layer and the connection weight And threshold b, calculate the predicted output value O of the neural network;

[0013]

[0014] Where: represents the threshold from the hidden layer to the kth output layer node, represents the predicted output value from the hidden layer to the kth output layer node, represents the output value of the jth node in the hidden layer; the prediction error e is calculated by the obtained predicted output value O and the expected output value y;

[0015]

[0016] Represents the error from the hidden layer to the kth output layer node, and in turn updates the network connection weights through the prediction error e , and thresholds a and b; until the prediction error e meets the set threshold, the training is completed.

[0017] According to a method for calculating the ultimate bearing capacity of a pile based on a BP neural network provided by the present invention, the network connection weights are updated in turn through the prediction error e , The calculation formulas for , and thresholds a and b are as follows:

[0018]

[0019]

[0020]

[0021]

[0022] Where: η is the learning rate, represents the updated connection weight between the input layer and the hidden layer, represents the updated connection weight between the hidden layer and the output layer, represents the threshold from the updated input layer to the jth hidden layer node, Represents the threshold of the updated hidden layer to the kth output layer node.

[0023] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for calculating the ultimate bearing capacity of a pile based on a BP neural network as described above is implemented.

[0024] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for calculating the ultimate bearing capacity of a pile based on a BP neural network as described above is implemented.

[0025] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for calculating the ultimate bearing capacity of a pile based on a BP neural network as described above is implemented.

[0026] The present invention provides a method for calculating the ultimate bearing capacity of a pile body based on a BP neural network. After determining the sensitivity parameters that significantly affect the bearing settlement of the pile foundation, the sensitivity parameters and the ultimate bearing capacity of the pile foundation under the corresponding sensitivity parameters are collected to construct training samples, and the formed sample data set is more accurate. A BP neural network model is constructed and trained using the sample data set, so that the BP neural network model is more adapted to the ultimate bearing capacity to obtain an ultimate bearing capacity prediction model. The ultimate bearing capacity prediction model is used to predict the ultimate bearing capacity of the pile foundation to be calculated, providing a reference basis for actual engineering construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0028] Figure 1 It is a flow chart of a method for calculating the ultimate bearing capacity of a pile based on a BP neural network provided by the present invention;

[0029] Figure 2 It is a schematic diagram of a process for determining a sensitivity parameter having a significant influence on pile bearing settlement provided by the present invention;

[0030] Figure 3 It is a schematic diagram of the finite element model structure provided by the present invention;

[0031] Figure 4 It is a schematic diagram of the process of constructing training samples provided by the present invention;

[0032] Figure 5 It is a flow chart of constructing and training a BP neural network model using a sample data set provided by the present invention;

[0033] Figure 6 It is the architecture diagram of the BP neural network provided by the present invention;

[0034] Figure 7 It is a structural schematic diagram of an actual engineering pile model provided by the present invention;

[0035] Figure 8 It is a comparative analysis diagram of the ultimate bearing capacity of the actual engineering pile model provided by the present invention and the ultimate bearing capacity calculated by the ultimate bearing capacity calculation model;

[0036] Fig. 9 It is a structural schematic diagram of the electronic device provided by the present invention.

[0037] Reference numerals:

[0038] 1. Upper layer materials; 2. Middle layer materials; 3. Bottom layer materials; 4. Pile body; 5. Upper layer covering soil. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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.

[0040] The present invention provides a method for calculating the ultimate bearing capacity of a pile based on a BP neural network. Figure 1 , including the following steps:

[0041] Step 1: Determine the sensitivity parameters that have a significant impact on pile bearing settlement through engineering tests.

[0042] See also Figure 2 First, determine the pile foundation parameters; secondly, design an orthogonal test table based on the pile foundation parameters; then, conduct a pile foundation settlement test or perform a finite element analysis on the pile foundation settlement based on the orthogonal test table, and record the pile foundation settlement; finally, analyze the factors affecting the pile foundation settlement and obtain the sensitivity parameters.

[0043] In one embodiment, the pile foundation parameters include elastic modulus, cohesion, Poisson's ratio, internal friction angle, and expansion angle. These parameters have a great influence on the ultimate bearing capacity of the pile foundation.

[0044] Exemplarily, in this embodiment, the elastic modulus E: 3~8 Gpa, Poisson's ratio μ: 0.3~0.4, the internal friction angle φ: 18°~34°, the cohesion c: 0.9~2.1 Mpa, and the expansion angle: 8°~16°.

[0045] In one embodiment, a finite element model for pile foundation settlement simulation is established, as shown in Table 1, and the The orthogonal table divides each pile foundation parameter into five levels. Each parameter combination in the 25 orthogonal tests is substituted into the established finite element model to obtain the corresponding pile settlement under each working condition.

[0046]

[0047] For example, in this embodiment, see Figure 3 and Figure 7 In this embodiment, a finite element model of double cave foundation piles is constructed. The length, width and height of the finite element model are all 80 cm (the depth of the highly weathered rock layer is 29 cm, the medium weathered rock layer is 26 cm, and the depth of the low weathered rock layer is 25 cm), the pile length is 20 cm (the embedded rock layer part is 17 cm), the length, width and height of the caves are 20 cm, 15 cm and 15 cm respectively, and the caves are kept horizontal with a spacing of 4 cm.

[0048] For example, ABAQUS software is used for finite element analysis. When constructing a finite element model, in the component module, there are many types of built-in unit libraries that can be selected, which are generally divided into continuum units, shell units, membrane units, beam units, rod units, rigid units, connection units and infinite units, as well as some special units for special problems. The component type is selected as solid stretching, and the double cave foundation model includes soil and piles.

[0049] Exemplarily, the corresponding material factors are assigned to the soil and pile respectively, wherein the pile adopts a linear elastic model and the soil adopts an elastoplastic model, which obeys the Mohr-Coulomb yield criterion. Compared with the Tresca yield criterion and the Mises yield criterion, Mohr-Coulomb can better describe the strength characteristics and failure characteristics between materials. Under the action of external loads, the rock layer forms a continuous sliding surface, causing the foundation to be damaged. The Mohr-Coulomb yield criterion can better show the properties of the test model material. In addition, the criterion factor is easy to determine and the concept is clear, which is more in line with actual engineering. Therefore, the Mohr-Coulomb constitutive model is selected as a condition for use.

[0050] For example, boundary conditions are set for the model in the Load module to control the horizontal movement of the model side, and the bottom of the model is set to be fully hinged. The meshing type is selected as C3D8R, and the horizontal and vertical ranges of the location of the cave are encrypted.

[0051] For example, in numerical simulation work, the automatic geostress balance method is used to balance the initial geostress of the soil. Due to the elastic-plastic properties of rock and soil materials, multiple iterations are required to achieve the balance of geostress using the keyword-defined initial stress method, which can easily lead to non-convergence of model calculations, resulting in unsatisfactory balance results. At the same time, if there are too many working conditions, the calculation of the initial stress ODB file result is too cumbersome. In comparison, the automatic geostress balance method is the most preferred.

[0052] Exemplarily, a method of applying load to the pile body is adopted to simulate the stress process of the double cave foundation piles.

[0053] Exemplarily, the interface between the pile body and the rock mass is used as the contact surface, wherein the pile body is used as the main control surface and the rock mass is used as the auxiliary surface, and the contact tracking between the pile foundation and the rock and soil is performed in a limited sliding manner.

[0054] In one embodiment, the range analysis method is used to determine the factors affecting the pile foundation settlement and obtain the sensitivity parameters. By calculating the range value R of the pile settlement under different factors in Table 1, the size of the range value R reflects the significance of the studied factor in affecting the pile bearing settlement. When the R value is larger, the sensitivity of the factor is also greater, which further indicates that the degree of influence of the factor on the settlement of the pile under the action of external load is also greater; the smaller the R value, the smaller the sensitivity of the factor, which indicates that the influence of the factor on the settlement of the pile is smaller.

[0055] Exemplarily, as shown in Table 2, in this embodiment, the calculation results are R1>R2>R4>R5>R3, R1 is the elastic modulus, R2 is the cohesion, R3 is the Poisson's ratio, R4 is the internal friction angle, and R5 is the expansion angle, that is, the sensitivity parameters are the elastic modulus E, cohesion c and internal friction angle. These three indicators have a high significance in affecting the bearing settlement of the pile.

[0056]

[0057] Step 2: See Figure 4 , collect sensitivity parameters and the ultimate bearing capacity of pile foundation under the corresponding sensitivity parameters, construct training samples, and form a sample data set. The method of constructing training samples is as follows: use engineering pile tests or finite element simulation analysis to obtain the ultimate bearing capacity of pile foundation under sensitivity parameters; use sensitivity parameters as input and pile foundation ultimate bearing capacity as labels to construct training samples.

[0058] In one embodiment, finite element simulation analysis is used to obtain the ultimate bearing capacity of the pile foundation under sensitivity parameters, a finite element model is constructed, and load is applied to obtain the ultimate bearing capacity of the pile foundation under different sensitivity parameters. The samples are shown in Table 3.

[0059]

[0060]

[0061]

[0062] Step 3: Construct a BP neural network model and train it using the sample data set, save the training parameters, and obtain the ultimate bearing capacity prediction model.

[0063] See also Figure 5-Figure 6 , build a BP neural network model and determine the number of nodes in the input layer n , the number of hidden layer nodes l , the number of nodes in the output layer is m, and the connection weights between the input layer and the hidden layer are , the connection weights between the hidden layer and the output layer , the threshold a of the hidden layer, the threshold b of the output layer are initialized to obtain the initial structure of the BP neural network model; the sample data set is input into the BP neural network model, and the weights are connected and threshold To calculate the input value from the hidden layer to the output layer;

[0064]

[0065] Where: represents the threshold from the input layer to the jth hidden layer node; f() is the transfer function, represents the weight from the input layer to the jth hidden layer node to the kth output layer node, Represents the input value of the i-th node in the input layer;

[0066] According to the output value H of the hidden layer and the connection weight And threshold b, calculate the predicted output value O of the neural network;

[0067]

[0068] Where: represents the threshold from the hidden layer to the kth output layer node, represents the predicted output value from the hidden layer to the kth output layer node, represents the output value of the jth node in the hidden layer; the prediction error e is calculated by the obtained predicted output value O and the expected output value y;

[0069]

[0070] Represents the error from the hidden layer to the kth output layer node, and in turn updates the network connection weights through the prediction error e , and thresholds a and b; until the prediction error e meets the set threshold, the training is completed.

[0071] For example, the number of input nodes of the BP neural network structure is 3, the number of hidden layer nodes is 7, and the number of output nodes is 1. Number of hidden layer nodes l The general formula To select, where t is a constant in [0,10], in this embodiment t=5.

[0072] Exemplarily, f is a transfer function, and the sigmoid function is selected as its excitation function.

[0073] In one embodiment, the network connection weights are updated in turn by the prediction error e , The calculation formulas for , and thresholds a and b are as follows:

[0074]

[0075]

[0076]

[0077]

[0078] Where: η is the learning rate, represents the updated connection weight between the i-th node in the input layer and the j-th node in the hidden layer, represents the updated connection weight between the jth node in the hidden layer and the kth node in the output layer, represents the threshold from the updated input layer to the jth hidden layer node, Represents the threshold of the updated hidden layer to the kth output layer node.

[0079] Step 4: Collect the sensitivity parameters of the pile foundation to be calculated, input them into the ultimate bearing capacity prediction model, and obtain the ultimate bearing capacity of the pile.

[0080] First, an actual engineering pile model is established, and the actual engineering pile data is input into the ultimate bearing capacity calculation model to obtain the ultimate bearing capacity of the engineering pile. The ultimate bearing capacity prediction value of the actual engineering pile data is 7kN.

[0081] In one embodiment, see Figure 7According to the distribution of rock layers in the actual engineering survey data, the model size part has the upper material 1 29cm thick, the middle material 2 26cm thick, and the bottom material 3 25cm thick. The pile body 4 is located at the center of the horizontal section of the model box and is composed of a PMMA cylinder with a height of 20cm and a width of 4cm. The model pile body 4 is embedded in the material 1 for 17cm, of which 3cm is exposed on the upper outer side of the model box. The dimensions of the rectangular cavity are 20cm, 15cm and 15cm respectively, buried at 37cm thick of the middle layer material 2, and the distance between the bottom of the pile end and the top of the cave is 20cm, which is equal to 5 times the pile diameter. Among them, the rectangular cavity b is located directly below the model pile, and the buried depth of the pile side cave model a and the pile end cave model b is consistent, and the minimum boundary distance between a and b is 4cm, which is equal to 1 times the pile diameter.

[0082] In one embodiment, the ultimate bearing capacity of the pile body in the actual engineering pile model can be detected by using a distributed optical fiber strain sensing system. When the actual engineering pile model is established, optical fiber is added for monitoring.

[0083] For example, see Figure 8 The ultimate bearing capacity of the actual engineering pile model monitored by optical fiber was compared with the ultimate bearing capacity calculated by the ultimate bearing capacity calculation model. The load-displacement curve in the result was consistent with the model test results, and the error of the ultimate bearing capacity was small, which proved that the method had certain feasibility and effectiveness.

[0084] The following is a description of the pile ultimate bearing capacity calculation system based on the BP neural network provided by the present invention. The pile ultimate bearing capacity calculation system based on the BP neural network described below and the pile ultimate bearing capacity calculation method based on the BP neural network described above can refer to each other.

[0085] Fig. 9 An example of a physical structure diagram of an electronic device is provided, and the electronic device may include: a processor 910, a communication interface 920, a memory 930, and a communication bus 940, wherein the processor 910, the communication interface 920, and the memory 930 communicate with each other through the communication bus 940. The processor 910 may call the logic instructions in the memory 930 to execute a method for calculating the ultimate bearing capacity of a pile body based on a BP neural network, and the method includes: forming a sample data set by finite element simulation of engineering pile data; constructing a BP neural network structure according to the variables and results corresponding to the sample data set; training the BP neural network structure using the sample data set to obtain an ultimate bearing capacity calculation model; collecting actual engineering pile data and inputting the ultimate bearing capacity calculation model to obtain the ultimate bearing capacity of the actual engineering pile.

[0086] In addition, the logic instructions in the above-mentioned memory 930 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0087] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute a method for calculating the ultimate bearing capacity of a pile based on a BP neural network provided by the above-mentioned methods. The method includes: forming a sample data set by finite element simulation of engineering pile data; constructing a BP neural network structure according to the variables and results corresponding to the sample data set; training the BP neural network structure using the sample data set to obtain an ultimate bearing capacity calculation model; collecting actual engineering pile data and inputting it into the ultimate bearing capacity calculation model to obtain the ultimate bearing capacity of the actual engineering pile.

[0088] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it is implemented to execute a method for calculating the ultimate bearing capacity of a pile based on a BP neural network provided by the above-mentioned methods. The method includes: forming a sample data set by finite element simulation of engineering pile data; constructing a BP neural network structure according to the variables and results corresponding to the sample data set; training the BP neural network structure using the sample data set to obtain an ultimate bearing capacity calculation model; collecting actual engineering pile data and inputting it into the ultimate bearing capacity calculation model to obtain the ultimate bearing capacity of the actual engineering pile.

[0089] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0090] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for calculating the ultimate bearing capacity of a pile based on a BP neural network, characterized in that: The following steps are involved: Through engineering tests, determine the sensitivity parameters that have a significant impact on pile bearing and settlement; Collect sensitivity parameters and ultimate bearing capacity of pile foundation under corresponding sensitivity parameters, construct training samples, and form sample data sets; Construct a BP neural network model and train it using a sample data set, save the training parameters, and obtain an ultimate bearing capacity prediction model; Collect the sensitivity parameters of the pile foundation to be calculated, input them into the ultimate bearing capacity prediction model, and obtain the ultimate bearing capacity of the pile body; Among them, the method for determining the sensitivity parameters that significantly affect the bearing settlement of pile foundations is as follows: Determine pile foundation parameters; Design orthogonal test table according to pile foundation parameters; Conduct pile foundation settlement test according to orthogonal test table or perform finite element analysis on pile foundation settlement and record the pile foundation settlement; The factors affecting pile foundation settlement are analyzed, and the range analysis method is used to determine the factors affecting pile foundation settlement and obtain sensitivity parameters; The pile foundation parameters include elastic modulus, cohesion, Poisson's ratio, internal friction angle, and expansion angle; The method for constructing training samples is as follows: Use engineering pile tests or finite element simulation analysis to obtain the ultimate bearing capacity of pile foundation under sensitive parameters; Using sensitivity parameters as input and ultimate bearing capacity of pile foundation as label, training samples are constructed. The method of constructing a BP neural network model and training it using a sample data set is as follows: Construct a BP neural network model and determine the number of nodes in the input layer n , the number of hidden layer nodes l The number of hidden layer nodes l is generally calculated according to the formula To select, where t is a constant in [0,10], the number of nodes in the output layer is m, and the connection weight between the input layer and the hidden layer is , the connection weights between the hidden layer and the output layer , the threshold a of the hidden layer, the threshold b of the output layer are initialized to obtain the initial structure of the BP neural network model; the sample data set is input into the BP neural network model, and the weights are connected and threshold To calculate the input value from the hidden layer to the output layer; Where: represents the threshold from the input layer to the jth hidden layer node; f is the transfer function; According to the output value H of the hidden layer and the connection weight And threshold b, calculate the predicted output value of the neural network O ; Where: represents the threshold from the hidden layer to the kth output layer node, represents the predicted output value from the hidden layer to the kth output layer node, Represents the output value of the jth node in the hidden layer; The predicted output value obtained by O Calculate the prediction error e with the expected output value y; Represents the error from the hidden layer to the kth output layer node, and in turn updates the network connection weights through the prediction error e , and thresholds a, b; The training is completed until the prediction error e meets the set threshold.

2. A method for calculating the ultimate bearing capacity of a pile based on a BP neural network according to claim 1, characterized in that: When determining the sensitivity parameter method that affects the significant bearing settlement of pile foundation, a finite element model of double-cavity foundation piles was constructed.

3. A method for calculating the ultimate bearing capacity of a pile based on a BP neural network according to claim 1, characterized in that: The network connection weights are updated in turn through the prediction error e , The calculation formulas for , and thresholds a and b are as follows: Where: η is the learning rate, represents the updated connection weight between the input layer and the hidden layer, represents the updated connection weight between the hidden layer and the output layer, represents the threshold from the updated input layer to the jth hidden layer node, Represents the threshold of the updated hidden layer to the kth output layer node.

4. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method for calculating the ultimate bearing capacity of a pile based on a BP neural network as described in any one of claims 1 to 3 is implemented.

5. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for calculating the ultimate bearing capacity of a pile based on a BP neural network as claimed in any one of claims 1 to 3 is implemented.

6. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method for calculating the ultimate bearing capacity of a pile based on a BP neural network as described in any one of claims 1 to 3 is implemented.

Images