Method and system for constructing constitutive model of rock-soil material, electronic device and storage medium

Abstract

The application discloses a geotechnical material constitutive model construction method and system, an electronic device and a storage medium, and relates to the technical field of geotechnical engineering. The method comprises the following steps: constructing a full-strain parameter quantitative correlation network; calculating an initial set of constitutive model parameters capable of fully characterizing the full-strain mechanics; comparing the displacement prediction value with the measured displacement value based on the initial set of constitutive model parameters; when the deviation exceeds a predetermined threshold, starting a traceability adaptive algorithm; and finally determining the set of constitutive model parameters of the target geotechnical material. The application effectively solves the defects of the existing constitutive model parameters lacking physical and mathematical correlations in the full-strain range, and overcomes the disadvantages of model static solidification by dynamically adjusting the constitutive model parameter values, thereby achieving better engineering prediction accuracy at a lower cost.

Description

Technical Field

[0001] This invention relates to the field of geotechnical engineering technology, and in particular to methods, systems, electronic devices and storage media for constructing constitutive models of geotechnical materials. Background Technology

[0002] One of the core bottlenecks in the digitalization and intelligentization of geotechnical engineering lies in the lack of an accurate material model capable of autonomously adapting to the dynamic processes of engineering. Currently, advanced constitutive models, such as boundary surface theory and small strain theory, are insufficient for describing soil changes from minimal strain (…). From large strain () When dealing with complex mechanical behaviors across the full strain range, the following shortcomings exist:

[0003] First, the parameter system lacks a physical and mathematical connection. To achieve full strain description, existing models introduce a reference shear strain to control the stiffness decay at small strains. Boundary surface shape parameters that determine yielding and hardening behavior and interpolation parameters A series of parameters. However, these parameters differ from "fundamental physical parameters" (such as porosity) that can be directly measured through conventional tests (consolidation, triaxial, etc.). Compression index Rebound Index There is a lack of quantifiable mathematical correlation between the internal friction angle (φ′) and the internal friction angle (φ′). This leads to parameter calibration relying entirely on the researcher's experience and repeated calculations, making it a time-consuming, highly subjective, and unreproducible "black box" process, hindering the engineering popularization of advanced models.

[0004] Secondly, there is a disconnect between static models and dynamic engineering. The models obtained through the aforementioned empirical calibration are static and fixed. They cannot reflect the continuous and sequential disturbances to the soil state caused by actual engineering construction (such as layered excavation of foundation pits and step-by-step tunneling of shield tunnels). The structure, stress history, and mechanical response characteristics of the soil across the entire strain range evolve dynamically as the project progresses, and the initial parameters may also change due to disturbances. The static model is severely disconnected from the reality of dynamic engineering, and its prediction errors will inevitably accumulate and amplify as construction progresses, failing to meet the requirements of intelligent control based on a "monitoring-prediction" closed loop.

[0005] Existing technological improvements are limited to two directions: one is to rely on expensive and complex special tests (such as resonant columns and bending elements) to directly measure certain advanced parameters, which is extremely costly; the other is to use pure data-driven machine learning methods to simulate global parameters or build surrogate models. Although this may improve the goodness of fit, it is completely detached from the physical mechanism, and the result often has physically unreasonable parameter combinations, as well as poor generalization ability and engineering credibility.

[0006] Therefore, the industry urgently needs a new method for constructing constitutive models that can integrate the physical meaning of parameters, achieve rapid and objective calibration, and evolve dynamically with engineering projects. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method, system, electronic device and storage medium for constructing constitutive models of geotechnical materials. It effectively solves the problems of lack of physical and mathematical correlation of parameter system in constitutive models and inability to dynamically and adaptively predict engineering results. It constructs a full strain parameter quantification correlation network and improves the accuracy of parameters through a source-tracing adaptive algorithm, thereby further improving the accuracy of engineering prediction in a dynamic and adaptive manner.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A method for constructing constitutive models of geotechnical materials includes the following steps:

[0010] S100, for the target geotechnical material, constructs a full strain parameter quantification correlation network including an intrinsic physical layer, a response feature layer, and a constitutive control layer, which is used to establish logical connections between the parameters of the target constitutive model and the parameters of multiple specified intrinsic physical layers.

[0011] S200: Collect the user-inputted intrinsic physical layer parameter values ​​and perform calculations based on the aforementioned full-strain parameter quantization correlation network to obtain an initial constitutive model parameter set Π0 that can fully describe the full-strain mechanical characteristics of the target soil and rock material. The initial constitutive model parameter set Π0 includes the input intrinsic physical layer parameter values, the calculated target constitutive model parameter values, and other constitutive model parameter values ​​required to determine displacement. These other constitutive model parameters include the initial shear modulus located in the response feature layer. and the slope of the critical state line And the plastic modulus interpolation parameters located in the constitutive control layer The calculation includes using a preset parameter quantization correlation formula to calculate the target constitutive model parameter values ​​of the response feature layer and constitutive control layer layer by layer based on the intrinsic physical layer parameter values; the parameter quantization correlation formula includes a reference shear strain for calculating the response feature layer parameters. The first formula and the boundary surface model shape parameters used to calculate the constitutive control layer parameters The second formula;

[0012] The first formula is ,in, It is the compression index; Porosity; It has an over-consolidation ratio; and The fitting coefficients are obtained based on experimental fitting calibration; the second formula is... ,in, The rebound index; The initial boundary surface shape parameters are reference values ​​obtained based on the modified Cambridge model yield surface shape. These are the fitting coefficients, obtained based on experimental fitting calibration.

[0013] S300, calculate the predicted displacement value and the deviation between the predicted displacement value and the measured displacement value based on the aforementioned initial set of constitutive model parameters Π0;

[0014] S400, when the deviation between the predicted displacement value and the measured displacement value exceeds a predetermined threshold, the source-tracing adaptive algorithm is initiated; the source-tracing adaptive algorithm is configured to: identify one or more intrinsic physical layer parameters that cause the aforementioned deviation, adjust one or more intrinsic physical layer parameters and iteratively calculate the displacement value until the aforementioned deviation does not exceed the predetermined threshold, and determine the constitutive model parameter set Π of the target geotechnical material based on the adjusted intrinsic physical layer parameter values. e .

[0015] Furthermore, in step S300, after calculating the deviation between the predicted displacement value and the measured displacement value, if the deviation does not exceed a predetermined threshold, the constitutive model parameter set Π of the target soil and rock material is determined using the parameter values ​​in the aforementioned initial constitutive model parameter set Π0. e ;

[0016] The specified multiple intrinsic physical layer parameters include porosity. Compression index Rebound Index Internal friction angle φ′ and overconsolidation ratio .

[0017] Furthermore, the source tracing adaptive algorithm includes the following steps:

[0018] S410, Construct the mapping relationship between the original physical layer parameters and the displacement prediction values;

[0019] S420, using the reverse tracing method to trace back to one or more questionable parameters in the original physical layer; the questionable parameters are those parameters in the original physical layer parameter values ​​of the target soil and rock material input by the user, which, due to numerical deviation, cause the aforementioned calculated displacement prediction value and the measured displacement value to deviate from a predetermined threshold.

[0020] S430, under the constraint of the physical feasible domain Ω of geotechnical materials, the value of the questionable parameter is adjusted; the physical feasible domain Ω refers to the reasonable value range of the questionable parameter in a physical sense, as well as the reasonable value range of other constitutive model parameters that have physical and / or mathematical relationships with the questionable parameter in a physical sense.

[0021] S440, based on the adjusted values ​​of the questionable parameters and the aforementioned full strain parameter quantization correlation network, recalculate the response characteristic layer parameter values, constitutive control layer parameter values, and displacement prediction values, and compare the updated displacement prediction values ​​with the measured displacement values ​​to obtain the deviation between the updated displacement prediction values ​​and the measured displacement values.

[0022] S450, determine whether the updated deviation exceeds a predetermined threshold. If it is determined that the deviation does not exceed the predetermined threshold, proceed to step S460; otherwise, proceed to step S430.

[0023] S460, based on the adjusted values ​​of the questionable parameters and the recalculated values ​​of the response characteristic layer parameters and constitutive control layer parameters, the constitutive model parameter set of the target soil and rock material is updated to obtain the final constitutive model parameter set Π. e .

[0024] Furthermore, the reverse tracing method includes the following steps:

[0025] S421, Calculate the sensitivity of each source physical layer parameter to the displacement prediction value; the sensitivity is used to measure the degree of influence of changes in the source physical layer parameter on the displacement prediction value;

[0026] S422, Sort the sensitivity of the original physical layer parameters in descending order of their values, select the top n original physical layer parameters in the sensitivity ranking and generate a list of questionable parameters; the list of questionable parameters contains n original physical layer parameters with high sensitivity, where n is an integer greater than or equal to 1.

[0027] Furthermore, when adjusting the values ​​of the questionable parameters, a trust-region oriented update is adopted so that all constitutive model parameters after the update still satisfy the constraints of the physical feasible region Ω.

[0028] Furthermore, when adjusting the value of the questionable parameter, the parameter value is adjusted according to a preset convergence criterion so that the deviation between the predicted displacement value and the measured displacement value tends to decrease.

[0029] This invention also discloses a constitutive model construction system for geotechnical materials, the system comprising:

[0030] The network construction module is configured to: construct a full strain parameter quantification correlation network for the target geotechnical material, which includes an intrinsic physical layer, a response feature layer, and a constitutive control layer. This network is used to establish logical connections between the parameters of the target constitutive model and multiple specified intrinsic physical layer parameters.

[0031] The data acquisition module is configured to: acquire the original physical layer parameter values ​​input by the user;

[0032] The data processing module is configured to: perform calculations based on the aforementioned full-strain parameter quantization correlation network to obtain an initial set of constitutive model parameters Π0 that can fully describe the full-strain mechanical characteristics of the target soil and rock material. The initial set of constitutive model parameters Π0 includes the input intrinsic physical layer parameter values, the calculated target constitutive model parameter values, and other constitutive model parameter values ​​required to determine displacement. These other constitutive model parameters include the initial shear modulus located in the response feature layer. and the slope of the critical state line And the plastic modulus interpolation parameters located in the constitutive control layer The calculation includes using a preset parameter quantization correlation formula to calculate the target constitutive model parameter values ​​of the response feature layer and constitutive control layer layer by layer based on the intrinsic physical layer parameter values; the parameter quantization correlation formula includes a reference shear strain for calculating the response feature layer parameters. The first formula and the boundary surface model shape parameters used to calculate the constitutive control layer parameters The second formula; the first formula is ,in, It is the compression index; Porosity; It has an over-consolidation ratio; and The fitting coefficients are obtained based on experimental fitting calibration; the second formula is... ,in, The rebound index; The initial boundary surface shape parameters are reference values ​​obtained based on the modified Cambridge model yield surface shape. The fitting coefficients are obtained based on experimental fitting calibration; and,

[0033] The predicted displacement value and the deviation between the predicted displacement value and the measured displacement value are calculated based on the aforementioned initial set of constitutive model parameters Π0; and...

[0034] When the deviation exceeds a predetermined threshold, a source-tracing adaptive algorithm is initiated. This algorithm is configured to: identify one or more intrinsic physical layer parameters causing the aforementioned deviation; adjust and iteratively calculate the displacement values ​​of the one or more intrinsic physical layer parameters until the aforementioned deviation does not exceed the predetermined threshold; and determine the constitutive model parameter set Π of the target geotechnical material based on the adjusted intrinsic physical layer parameter values. e .

[0035] Furthermore, it also includes a model dynamic evolution module, which is configured to: for the Nth excavation layer during construction, determine the constitutive model parameter set Π for that layer based on the measured displacement values ​​monitored at that layer. e_N Subsequently, based on the constitutive model parameter set Π of the Nth excavation layer... e_NDetermine the initial set Π of constitutive model parameters for the (N+1)th excavation layer 0_N+1 At this point, the initial set of constitutive model parameters Π for the (N+1)th excavation layer 0_N+1 The initial values ​​of the original physical layer parameters are based on the constitutive model parameter set Π of the Nth excavation layer. e_N The corresponding value in the value is taken; where N is an integer greater than or equal to 1.

[0036] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the aforementioned method.

[0037] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned method.

[0038] Compared with the prior art, this invention has the following advantages and positive effects due to the adoption of the above technical solutions: by constructing a full strain parameter quantification correlation network, the quantification relationship between the original physical layer parameters and the response feature layer and constitutive control layer parameters is established. The displacement prediction value and its deviation from the measured displacement value are calculated through the obtained initial set of constitutive model parameters. Based on the source-tracing adaptive algorithm, the parameters that cause the displacement prediction value to deviate are found and corrected, thereby making the displacement prediction value more accurate.

[0039] The technical solution of this invention overcomes the shortcomings of previous constitutive model parameters lacking physical and mathematical correlations across the entire strain range and relying on experience for calibration. At the same time, it adopts an adaptive mechanism of engineering feedback to dynamically adjust the constitutive model parameter values, overcoming the drawbacks of static model fixation, and achieving better engineering prediction accuracy at a lower cost. Attached Figure Description

[0040] Figure 1 A flowchart of a method for constructing constitutive models of geotechnical materials provided in an embodiment of the present invention.

[0041] Figure 2 The logical flowchart of the source tracing adaptive algorithm provided in the embodiment of the present invention is shown.

[0042] Figure 3 The module structure diagram of the geotechnical material constitutive model construction system provided in the embodiments of the present invention is shown. Detailed Implementation

[0043] The following detailed description, in conjunction with the accompanying drawings and specific embodiments, provides a further detailed account of the constitutive model construction method, system, electronic device, and storage medium for geotechnical materials disclosed in this invention. It should be noted that techniques (including methods and apparatus) known to those skilled in the art may not be discussed in detail, but where appropriate, such known techniques are considered part of the specification. Furthermore, other examples of exemplary embodiments may have different values. The structures, proportions, sizes, etc., depicted in the accompanying drawings are merely illustrative of the content disclosed in this specification for the understanding and reading of those skilled in the art, and are not intended to limit the conditions under which the invention can be implemented.

[0044] In the description of the embodiments of this application, " / " means "or", and "and / or" is used to describe the relationship between related objects, indicating that there can be three relationships. For example, "A and / or B" means: A and B exist alone, B exists alone, and A and B exist simultaneously. In the description of the embodiments of this application, "multiple" refers to two or more.

[0045] The technical concept and solution of the present invention will be described below based on exemplary application scenarios.

[0046] Example

[0047] See Figure 1 The illustration shows a method for constructing a constitutive model of geotechnical materials provided in this embodiment. Specifically, the method includes the following steps:

[0048] S100 constructs a full strain parameter quantification correlation network for target geotechnical materials, comprising an intrinsic physical layer, a response characteristic layer, and a constitutive control layer.

[0049] The full strain parameter quantization correlation network is used to establish logical connections between the target constitutive model parameters and multiple specified intrinsic physical layer parameters.

[0050] In this embodiment, the full strain parameter quantization correlation network adopts a hierarchical architecture, which may include an intrinsic physics layer, a response feature layer, and a constitutive control layer.

[0051] The original physical layer is composed of parameters that characterize the inherent properties of soil and rock materials and can be directly measured, usually by conventional geotechnical tests.

[0052] In this embodiment, the intrinsic physical layer parameters may include porosity. Compression index Rebound Index Internal friction angle φ′ and overconsolidation ratio The methods for measuring the above parameters are existing technologies and will not be described in detail here.

[0053] The response feature layer consists of parameters that describe the key macroscopic mechanical characteristics of soil and rock materials across the entire strain range.

[0054] In this embodiment, the response feature layer parameters may include the initial shear modulus. Slope of the critical state line and reference shear strain Among them, the reference shear strain This represents the shear strain corresponding to a reduction in shear modulus to 70% of its initial value. These parameters can typically be determined experimentally or through calculation and empirical estimation.

[0055] The constitutive control layer consists of core parameters that can directly control the evolution of the full strain response.

[0056] In this embodiment, the constitutive control layer parameters may include boundary surface model shape parameters. and plastic modulus interpolation parameters These parameters can usually be determined experimentally, or through calculation and empirical calculation.

[0057] It should be noted that the initial shear modulus Slope of the critical state line and plastic modulus interpolation parameters The specific values ​​and / or calculation methods are existing technologies and will not be detailed here.

[0058] S200: Collect the user-inputted intrinsic physical layer parameter values, and perform calculations based on the aforementioned full strain parameter quantization correlation network to obtain an initial set of constitutive model parameters Π0 that can fully describe the full strain mechanical characteristics of the target soil and rock material. The initial set of constitutive model parameters Π0 includes the input intrinsic physical layer parameter values, the calculated target constitutive model parameter values, and other constitutive model parameter values ​​required to determine the displacement. The calculation includes using a preset parameter quantization correlation formula to calculate the target constitutive model parameter values ​​of the response feature layer and the constitutive control layer layer by layer based on the intrinsic physical layer parameter values.

[0059] In this step, the other constitutive model parameters include the initial shear modulus located in the response feature layer. and the slope of the critical state line And the plastic modulus interpolation parameters located in the constitutive control layer .

[0060] In this embodiment, the target constitutive model parameters include reference shear strain. and boundary surface model shape parameters The parameter quantization correlation formula includes a method for calculating the reference shear strain. The first formula and the parameters used to calculate the boundary surface model shape parameters The second formula.

[0061] The first formula is constructed based on the laws of one-dimensional compression and cyclic shear tests of soil and rock materials. Specifically, the first formula is:

[0062]

[0063] In the formula, The compression index, Porosity It is an over-consolidation ratio. and The fitting coefficients are obtained from experimental fitting calibration in this embodiment.

[0064] The physical idea behind the first formula is: the stiffness decay characteristic of soil under small strain (from... The characterization of compressibility is primarily controlled by the sensitivity of its microstructure to deformation. The compressibility index... This reflects the compressibility of soil particles under pressure; void ratio This reflects the density of the structure; overconsolidation ratio Reflecting the historical degree of compression of the structure, the introduced Power function term It can reflect the control effect of stress history on microstructure. Based on this, construct... , and A comprehensive physical index that quantifies the sensitivity of the microstructure to deformation, based on a combination of three parameters: and with Through the proportionality coefficient Establish direct mathematical connections.

[0065] The parameters and In this embodiment, the fitting coefficient related to soil properties is determined from experimental data through "objective function minimization." Specifically, the coefficients can be obtained from monotonic or cyclic shear tests. curve( Extracted from (shear modulus) Measured values, and in , and To establish a regression or optimization problem for the input, it is preferable to use least squares or robust regression to solve it, thus improving the prediction. Compared with actual measurement The mean square error is minimized (objective function), thus determining the and The value of .

[0066] Preferably, determination is made on multiple groups of samples of the same soil type. and The value, that is, for the same soil type, can be calculated by increasing the number of samples. and The value is used to improve the generalization of the quantification correlation formula.

[0067] In this embodiment, the value ranges of each parameter in the first formula are as follows: Based on the properties of sandy soil and clay, the preferred value range is 0.6 to 1.4; Based on the properties of common soil types, the preferred value range is 0.1 to 1.0; Based on the properties of normally consolidated, slightly overconsolidated, and severely overconsolidated soils, the preferred value range is 1.0 to 20.0; Based on the experimental results, the fitting calibration was performed, and the value range was (5.0~25.0)× ; The fitting calibration is based on the experimental results, and the value range is -1.0 to 0.

[0068] As an example, and not a limitation, the soil sample test results from a certain region are as follows: =0.7, =0.4, different sampling positions The difference is that, taking the selection of 4 sampling locations as an example, the measured values ​​are different. They are respectively: =1.2, 1.9, 2.5, 4.0. and The fitted value is taken as follows =15× , =-0.5, calculated according to the above formula, four types can be obtained. In the case of They are: 3.22× 2.56× 2.23× 1.76× .

[0069] For common soft soils, existing literature indicates that... The value is in the range of 1.6× ~4.2× Within the specified range, the calculation results of the above formula are consistent with those in existing literature. A description within a reasonable range, and capable of reflecting the following Increase The characteristic of a gradually decreasing, reasonable trend of change.

[0070] The second formula is:

[0071]

[0072] In the formula, The rebound index, The initial boundary surface shape parameters are reference values ​​obtained based on the modified Cambridge model yield surface shape. The fitting coefficients are obtained based on experimental fitting calibration.

[0073] The physical idea behind the second formula is: boundary surface model shape parameters Essentially, this reflects the coupling relationship between the development of plastic strain and shear yielding in soil during the yielding process. The ratio of the compression index to the resilience index... It is a key macroscopic indicator describing the asymmetry of soil loading and unloading behavior (i.e., plastic deformation potential). Based on this, a... relative to the initial boundary surface shape parameters Change and There exists a linear relationship, where, The logarithmic function is better able to characterize the nonlinear features of the above coupling relationship as plastic potential changes.

[0074] The value of is a baseline value obtained based on the modified Cambridge model yield surface shape; the fitting coefficient The values ​​can preferably be obtained by least squares method or robust regression.

[0075] In this embodiment, the preferred value ranges for each parameter in the second formula are as follows: Based on the properties of common soil types, the preferred value range is 0.1 to 1.0; Based on the properties of common soil types, the preferred value range is 0.01 to 0.2; As a constant, we take it to be 2.0; Based on the fitting and calibration of the experimental results, the preferred value range is 0.0~0.5.

[0076] As an example, and not a limitation, the consolidation test results of soil samples from a certain region are as follows: =0.4, Empirical estimation is performed, taking four sampling locations as an example, corresponding to four The values ​​are as follows: =0.04, 0.05, 0.06, 0.08. The fitting coefficients are taken as follows: =0.3, =2.0. Calculations using the above formula yield four possible outcomes. The values ​​are 2.69, 2.62, 2.57, and 2.48, respectively.

[0077] For common soil types, existing literature indicates For values ​​ranging from 2.0 to 3.0, the calculation results of the above formulas all conform to existing literature regarding... A description within a reasonable range. Furthermore, an increased rebound index indicates that the soil is better able to resist permanent deformation after unloading, while... The gradual decrease also reflects the soil's increased resistance to deformation.

[0078] Obtain the result calculated by the first formula The result calculated by the second formula After that, the values ​​are combined with the user-input intrinsic physical layer parameter values ​​and the initial shear modulus. Slope of the critical state line and plastic modulus interpolation parameters The values ​​of these parameters are stored in a parameter set to obtain the initial set of constitutive model parameters Π0, which can fully describe the full strain mechanical characteristics of the target soil and rock material.

[0079] The initial shear modulus Slope of the critical state line and plastic modulus interpolation parameters The value can be directly input by the user (e.g., by collecting the parameter value input by the user through the parameter setting interface as an interactive interface), or it can use the system's default value (in this case, the system has pre-stored the values ​​of the relevant parameters and set the default values), or it can be calculated based on the system's preset calculation formula (in this case, the system has pre-stored the calculation formulas of the relevant parameters), and there are no restrictions here.

[0080] S300, calculate the predicted displacement value and the deviation between the predicted displacement value and the measured displacement value based on the aforementioned initial set of constitutive model parameters Π0.

[0081] Based on the obtained initial set of constitutive model parameters Π0, the displacement response is calculated through incremental finite element equilibrium solution. This solution process is existing technology in this field, and its principle is only briefly described here: Under each loading increment, the strain increment at the finite element integration point is used as input, and the stress and state variables are updated by the full-strain constitutive response operator. Simultaneously, the tangent stiffness under that increment is given by the uniform tangent operator, thus ensuring the stable convergence of the global equilibrium iteration. Subsequently, the nodal displacement response is obtained by solving the structural equilibrium equations, thereby obtaining the displacement prediction value. The specific calculation process will not be elaborated here.

[0082] Then, the deviation between the predicted displacement value and the measured displacement value is calculated.

[0083] S400, when the deviation between the predicted displacement value and the measured displacement value exceeds a predetermined threshold, the source tracing adaptive algorithm is activated.

[0084] In this step, when the deviation between the predicted displacement value and the measured displacement value exceeds a predetermined threshold, it indicates that the displacement calculation accuracy is not within an acceptable range. Therefore, it is inferred that one or more of the original physical layer parameter values ​​are deviated, and the source-tracing adaptive algorithm is activated at this time.

[0085] When the deviation between the predicted displacement value calculated in S300 and the measured displacement value does not exceed a predetermined threshold, it indicates that the displacement calculation accuracy is within an acceptable range. At this time, it can be determined that the intrinsic physical layer parameter values ​​are correct, and the final constitutive model parameter set Π of the target soil and rock material can be determined directly using the parameter values ​​in the aforementioned constitutive model parameter initial set Π0. e .

[0086] In this embodiment, the source-tracing adaptive algorithm is configured to: identify one or more intrinsic physical layer parameters that cause the aforementioned deviation; adjust one or more intrinsic physical layer parameters and iteratively calculate the displacement values ​​until the aforementioned deviation does not exceed a predetermined threshold; and determine the final constitutive model parameter set Π of the target geotechnical material based on the adjusted intrinsic physical layer parameter values. e .

[0087] See Figure 2 As shown, the specific steps of the source tracing adaptive algorithm are illustrated as a preferred typical approach, as follows:

[0088] S410, Construct Displacement Mapping: That is, construct the mapping relationship between the original physical layer parameters and the displacement prediction values.

[0089] In practice, different intrinsic physical layer parameter values ​​correspond to different displacement prediction values. After determining the values ​​of each parameter in the intrinsic physical layer parameters, the corresponding displacement prediction value can be determined based on the mapping relationship. .

[0090] Preferably, mapping rules can be used to construct parameters from the original physical layer. Chain mapping to displacement prediction values ,in, This is the predicted displacement value. This represents the constitutive response operator driven by the parameter set. The numerical solution operators are coupled with engineering boundary conditions. The calculation of the above operators are all existing technologies and will not be described in detail here.

[0091] S420, Tracing the source of questionable parameters: that is, using the reverse tracing method to trace back to one or more questionable parameters in the original physical layer.

[0092] The questionable parameter is a parameter among the original physical layer parameters of the target soil and rock material input by the user, where the numerical deviation causes the aforementioned calculated displacement prediction value to deviate from the measured displacement value by more than a predetermined threshold.

[0093] In this embodiment, during the process of tracing the source of questionable parameters, it is preferable to screen the displacement prediction values ​​according to the sensitivity of the source physical layer parameters. The specific steps can be as follows:

[0094] S421, Calculate sensitivity: that is, calculate the sensitivity of each source physical layer parameter to the displacement prediction value; the sensitivity is used to measure the degree of influence of changes in source physical layer parameters on the displacement prediction value.

[0095] In this process, it can be obtained through chain derivatives or numerical perturbations. For each candidate source parameter Calculate its sensitivity to error. . This is the predicted displacement value. Represents the variable Find the partial derivative. This indicates taking the absolute value.

[0096] S422, Generate a list of questionable parameters: Sort the sensitivity of the original physical layer parameters in descending order of their values, select the top n original physical layer parameters in the sensitivity ranking and generate a list of questionable parameters; the list of questionable parameters contains n original physical layer parameters with high sensitivity, where n is an integer greater than or equal to 1.

[0097] In this step, each intrinsic physical layer parameter—including porosity—is... Compression index Rebound Index Internal friction angle φ′, overconsolidation ratio The sensitivity values ​​of the parameters are sorted from largest to smallest. Then, the sensitivity values ​​of each parameter are compared and analyzed to determine the value of n.

[0098] The value of n is dynamically set by the calculated parameter sensitivity value. At this time, a dynamic setting unit for the questionable parameter list is also included. This unit is configured to: acquire the sensitivity values ​​of all intrinsic physical layer parameters; analyze the sensitivity values ​​of the parameters to obtain the maximum sensitivity value f; calculate the high-sensitivity interval [fd, f] based on a preset sensitivity difference threshold d; compare the sensitivity values ​​of all intrinsic physical layer parameters with the aforementioned high-sensitivity interval to obtain all parameters whose sensitivity values ​​fall within the high-sensitivity interval; count the number w of parameters whose sensitivity values ​​fall within the high-sensitivity interval; configure the value of n to w (n=w), meaning the questionable parameter list contains w parameters; and place these w parameters into the questionable parameter list according to their sensitivity values, from largest to smallest, with the parameter with the highest sensitivity placed first.

[0099] At this point, when n=1, corresponding to the situation where the sensitivity of one parameter is significantly greater than that of other parameters, the generated list of dynamically questionable parameters contains only one parameter. This is used as an example, not a limitation, such as calculating the porosity ratio. If the sensitivity of the particle is significantly greater than that of other parameters, then the value of n is set to 1, and the porosity is adjusted accordingly. Add it to the list of questionable parameters.

[0100] When n=2, this corresponds to the situation where the sensitivity values ​​of the two parameters are similar and significantly greater than those of other parameters. In this case, the generated list of questionable parameters will contain both parameters. This is used as an example, not a limitation, such as porosity. Sensitivity and compression index If the sensitivity of a parameter is located in the high-sensitivity range [fd, f], and the sensitivity of other parameters is not in this range, then the porosity will be... Compression index Also add it to the list of parameters under suspicion.

[0101] When n=3, this corresponds to the situation where the sensitivity values ​​of the three parameters are similar and significantly greater than those of the other parameters. In this case, the generated list of questionable parameters contains all three parameters. Similarly, the value of n can cover the sensitivity distribution of various parameters.

[0102] S430, Adjustment of questionable parameters: that is, under the constraint of the physical feasible domain Ω of geotechnical materials, the value of the questionable parameter is adjusted; the physical feasible domain Ω refers to the reasonable range of values ​​of the questionable parameter in a physical sense, as well as the reasonable range of values ​​of other constitutive model parameters that have physical and / or mathematical relationships with the questionable parameter in a physical sense.

[0103] As an example, and not a limitation, for highly compressible soils, the compression index Greater than 0.167. When the target soil material is a highly compressible soil, the compression index is... When entering the list of questionable parameters and adjusting their values, it should be ensured that the value is greater than 0.167. Furthermore, those skilled in the art should know that it should also be ensured that the value is consistent with the compression index. The values ​​of other parameters that have conversion relationships are also within the reasonable range of these parameters in a physical sense.

[0104] Accordingly, preferably, when adjusting the value of any questionable parameter in the questionable parameter list, a trust-region directional update is adopted so that all constitutive model parameters after the update still satisfy the constraints of the physical feasible region.

[0105] More appropriately, one or more intrinsic parameters in the questionable parameter sorting list can be adjusted within a predefined physical feasible region Ω. This physical feasible region Ω is set by the user or by a system default setting.

[0106] In this embodiment, the physically feasible region Ω can be composed of boundary constraints and coupling constraints. The boundary constraints are used to provide a reasonable range for each parameter, and the coupling constraints are used to provide physical consistency conditions between parameters. That is, within the current parameter... At this point, the priority for adjusting questionable parameters is determined based on sensitivity. Under the constraint of the physical feasible region, the parameters are locally iteratively updated to obtain the parameter update amount. ,make This is used to ensure that the updated parameters still satisfy the physical feasible region Ω. Wherein, This represents a projection or clipping operator. This represents the updated parameter value. The mathematical methods related to projection or clipping operators are existing technology and will not be elaborated here.

[0107] When adjusting questionable parameters, for any parameter in the list, the parameter value can be adjusted according to a preset convergence criterion. The convergence criterion refers to a verification standard that aims to reduce the deviation between the predicted and measured displacement values. For example, the convergence criterion can be that the deviation between the predicted and measured displacement values ​​after the (j+1)th adjustment must be less than the deviation between the predicted and measured displacement values ​​after the j-th adjustment (j is an integer greater than or equal to 1), thereby determining the adjustment direction of the questionable parameter.

[0108] Meanwhile, to avoid local oscillations caused by excessively large adjustment amplitudes even when the direction of adjustment of questionable parameters is correct, the adjustment amplitude of questionable parameters preferably adopts an adaptive rule: that is, if the deviation between the predicted displacement value and the measured displacement value after the (j+1)th adjustment is greater than the deviation between the predicted displacement value and the measured displacement value after the j-th adjustment, then the (j+1)th adjustment is cancelled, the adjustment amplitude of the questionable parameter is reduced, and the calculation is recalculated. If the deviation between the predicted displacement value and the measured displacement value still does not decrease compared to the deviation between the predicted displacement value and the measured displacement value after the j-th adjustment, then the adjustment is cancelled again and the adjustment amplitude is reduced again, until the calculation result decreases compared to the deviation between the predicted displacement value and the measured displacement value after the j-th adjustment.

[0109] When there are multiple parameters in the questionable parameter sorting list, the multiple parameters are adjusted together. The joint adjustment method can be based on existing multi-parameter adjustment methods, which are existing technologies and will not be described in detail here.

[0110] S440, Recalculate: Based on the adjusted values ​​of the questionable parameters and the aforementioned full strain parameter quantization correlation network, recalculate the response characteristic layer parameter values, constitutive control layer parameter values, and displacement prediction values, and compare the updated displacement prediction values ​​with the measured displacement values ​​to obtain the deviation between the updated displacement prediction values ​​and the measured displacement values.

[0111] S450, Determine if threshold is exceeded: In this step, it is necessary to determine whether the updated deviation exceeds the predetermined threshold. If it is determined that the predetermined threshold is not exceeded, proceed to step S460 and end; otherwise, proceed to step S430 until the updated deviation is satisfied that it does not exceed the predetermined threshold, then proceed to step S460 and end.

[0112] S460, the final constitutive model parameter set is obtained, i.e., the final constitutive model parameter set is determined. In this step, the constitutive model parameter set of the aforementioned target soil and rock material is updated based on the adjusted values ​​of the questionable parameters and the recalculated values ​​of the response characteristic layer parameters and constitutive control layer parameters, resulting in the final constitutive model parameter set Π. e .

[0113] See Figure 3 As shown in another embodiment of the present invention, a constitutive model construction system for geotechnical materials is also provided.

[0114] The system may specifically include a network construction module, a data acquisition module, and a data processing module.

[0115] The network construction module is configured to: construct a full strain parameter quantification correlation network for the target soil material, which includes an intrinsic physical layer, a response feature layer, and a constitutive control layer, and is used to establish logical connections between the target constitutive model parameters and multiple specified intrinsic physical layer parameters.

[0116] The data acquisition module is configured to acquire the original physical layer parameter values ​​input by the user.

[0117] The data processing module may specifically include a parameter set processing unit, a displacement calculation unit, and a source-tracing adaptive unit.

[0118] The parameter set processing unit is configured to: perform calculations based on the aforementioned full-strain parameter quantization correlation network to obtain an initial constitutive model parameter set Π0 that can fully describe the full-strain mechanical characteristics of the target soil and rock material. The initial constitutive model parameter set Π0 includes the input intrinsic physical layer parameter values, the calculated target constitutive model parameter values, and other constitutive model parameter values ​​required to determine the displacement. These other constitutive model parameters include the initial shear modulus located in the response feature layer. and the slope of the critical state line And the plastic modulus interpolation parameters located in the constitutive control layer The calculation includes using a preset parameter quantization correlation formula to calculate the target constitutive model parameter values ​​of the response feature layer and constitutive control layer layer by layer based on the original physical layer parameter values.

[0119] In this embodiment, the parameter quantization correlation formula includes a reference shear strain for calculating the response feature layer parameters. The first formula and the boundary surface model shape parameters used to calculate the constitutive control layer parameters The second formula.

[0120] The first formula is ,in, The compression index, Porosity It is an over-consolidation ratio. and The fitting coefficients are obtained based on experimental fitting calibration.

[0121] The second formula is ,in, The rebound index, The initial boundary surface shape parameters are reference values ​​obtained based on the modified Cambridge model yield surface shape. The fitting coefficients are obtained based on experimental fitting calibration.

[0122] The displacement calculation unit is configured to calculate the predicted displacement value and the deviation between the predicted displacement value and the measured displacement value based on the aforementioned constitutive model parameter initial set Π0.

[0123] The source-tracing adaptive unit is configured to: activate the source-tracing adaptive algorithm when the deviation exceeds a predetermined threshold; the source-tracing adaptive algorithm is configured to: identify one or more intrinsic physical layer parameters that cause the aforementioned deviation, adjust one or more intrinsic physical layer parameters and iteratively calculate the displacement values ​​until the aforementioned deviation does not exceed the predetermined threshold, and determine the constitutive model parameter set Π of the target geotechnical material based on the adjusted intrinsic physical layer parameter values. e .

[0124] In this embodiment, the system may further include a model dynamic evolution module.

[0125] The model dynamic evolution module is configured to: for the Nth excavation layer during construction, determine the constitutive model parameter set Π for that layer based on the measured displacement values ​​monitored at that layer. e_N Subsequently, based on the constitutive model parameter set Π of the Nth excavation layer... e_N Determine the initial set Π of constitutive model parameters for the (N+1)th excavation layer 0_N+1 At this point, the initial set of constitutive model parameters Π for the (N+1)th excavation layer 0_N+1 The initial values ​​of the original physical layer parameters are based on the constitutive model parameter set Π of the Nth excavation layer. e_N The corresponding value in the value is taken; where N is an integer greater than or equal to 1.

[0126] In this way, the optimized parameter set of the first excavation layer can be used to predict the deformation of the second excavation layer, and the optimized parameter set of the second excavation layer can be used to predict the deformation of the third excavation layer, and so on. After multiple such adaptive iterations, the model prediction error gradually converges and tends to stabilize throughout the subsequent excavation process.

[0127] Other technical features are described in the preceding embodiments and will not be repeated here.

[0128] Another embodiment of the present invention also provides an electronic device.

[0129] The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described above.

[0130] Other technical features are described in the preceding embodiments and will not be repeated here.

[0131] Another embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method described above.

[0132] The storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0133] Other technical features are described in the preceding embodiments and will not be repeated here.

[0134] In the above description, the disclosure of this invention is not intended to limit itself to these aspects. Rather, within the scope of the objectives of this disclosure, components can be selectively and operationally combined in any number. Furthermore, terms such as “comprising,” “encompassing,” and “having” should be interpreted by default as inclusive or open-ended, rather than exclusive or closed, unless explicitly defined as such. All technical, scientific, or other terms are to be understood by those skilled in the art, unless defined as such. Public terms found in dictionaries should not be interpreted too idealistically or impractically in the context of the relevant technical documents, unless explicitly defined as such in this disclosure. Any modifications or alterations made by those skilled in the art based on the foregoing disclosure are within the scope of the claims.

Claims

1. A method for constructing a constitutive model for soil and rock materials, characterized in that, Includes the following steps: S100, for the target geotechnical material, constructs a full strain parameter quantification correlation network including an intrinsic physical layer, a response feature layer, and a constitutive control layer, which is used to establish logical connections between the parameters of the target constitutive model and the parameters of multiple specified intrinsic physical layers. S200: Collect the user-inputted intrinsic physical layer parameter values ​​and perform calculations based on the aforementioned full-strain parameter quantization correlation network to obtain an initial constitutive model parameter set Π0 that can fully describe the full-strain mechanical characteristics of the target soil and rock material. The initial constitutive model parameter set Π0 includes the input intrinsic physical layer parameter values, the calculated target constitutive model parameter values, and other constitutive model parameter values ​​required to determine displacement. These other constitutive model parameters include the initial shear modulus located in the response feature layer. and the slope of the critical state line And the plastic modulus interpolation parameters located in the constitutive control layer The calculation includes using a preset parameter quantization correlation formula to calculate the target constitutive model parameter values ​​of the response feature layer and constitutive control layer layer by layer based on the intrinsic physical layer parameter values; the parameter quantization correlation formula includes a reference shear strain for calculating the response feature layer parameters. The first formula and the boundary surface model shape parameters used to calculate the constitutive control layer parameters The second formula; The first formula is ,in, It is the compression index; Porosity; It has an over-consolidation ratio; and The fitting coefficients are obtained based on experimental fitting calibration; the second formula is... ,in, The rebound index; The initial boundary surface shape parameters are reference values ​​obtained based on the modified Cambridge model yield surface shape. These are the fitting coefficients, obtained based on experimental fitting calibration. S300, calculate the predicted displacement value and the deviation between the predicted displacement value and the measured displacement value based on the aforementioned initial set of constitutive model parameters Π0; S400, when the deviation between the predicted displacement value and the measured displacement value exceeds a predetermined threshold, the source-tracing adaptive algorithm is initiated; the source-tracing adaptive algorithm is configured to: identify one or more intrinsic physical layer parameters that cause the aforementioned deviation, adjust one or more intrinsic physical layer parameters and iteratively calculate the displacement value until the aforementioned deviation does not exceed the predetermined threshold, and determine the constitutive model parameter set Π of the target geotechnical material based on the adjusted intrinsic physical layer parameter values. e .

2. The method according to claim 1, characterized in that, In step S300, after calculating the deviation between the predicted displacement value and the measured displacement value, if the deviation does not exceed a predetermined threshold, the constitutive model parameter set Π of the target soil and rock material is determined using the parameter values ​​in the aforementioned initial constitutive model parameter set Π0. e ; The specified multiple intrinsic physical layer parameters include porosity. Compression index Rebound Index Internal friction angle φ′ and overconsolidation ratio .

3. The method according to claim 1 or 2, characterized in that, The source tracing adaptive algorithm includes the following steps: S410, Construct the mapping relationship between the original physical layer parameters and the displacement prediction values; S420, using the reverse tracing method to trace back to one or more questionable parameters in the original physical layer; the questionable parameters are those parameters in the original physical layer parameter values ​​of the target soil and rock material input by the user, which, due to numerical deviation, cause the aforementioned calculated displacement prediction value and the measured displacement value to deviate from a predetermined threshold. S430, under the constraint of the physical feasible domain Ω of geotechnical materials, the value of the questionable parameter is adjusted; the physical feasible domain Ω refers to the reasonable value range of the questionable parameter in a physical sense, as well as the reasonable value range of other constitutive model parameters that have physical and / or mathematical relationships with the questionable parameter in a physical sense. S440, based on the adjusted values ​​of the questionable parameters and the aforementioned full strain parameter quantization correlation network, recalculate the response characteristic layer parameter values, constitutive control layer parameter values, and displacement prediction values, and compare the updated displacement prediction values ​​with the measured displacement values ​​to obtain the deviation between the updated displacement prediction values ​​and the measured displacement values. S450, determine whether the updated deviation exceeds a predetermined threshold. If it is determined that the deviation does not exceed the predetermined threshold, proceed to step S460; otherwise, proceed to step S430. S460, based on the adjusted values ​​of the questionable parameters and the recalculated values ​​of the response characteristic layer parameters and constitutive control layer parameters, the constitutive model parameter set of the target soil and rock material is updated to obtain the final constitutive model parameter set Π. e .

4. The method according to claim 3, characterized in that, The reverse tracing method includes the following steps: S421, Calculate the sensitivity of each source physical layer parameter to the displacement prediction value; the sensitivity is used to measure the degree of influence of changes in the source physical layer parameter on the displacement prediction value; S422, Sort the sensitivity of the original physical layer parameters in descending order of their values, select the top n original physical layer parameters in the sensitivity ranking and generate a list of questionable parameters; The list of questionable parameters contains n highly sensitive intrinsic physical layer parameters, where n is an integer greater than or equal to 1.

5. The method according to claim 3, characterized in that, When adjusting the values ​​of the questionable parameters, a trust-region oriented update is used to ensure that all updated constitutive model parameters still satisfy the constraints of the physical feasible region Ω.

6. The method according to claim 3, characterized in that, When adjusting the value of the questionable parameter, the parameter value is adjusted according to a preset convergence criterion so that the deviation between the predicted displacement value and the measured displacement value tends to decrease.

7. A constitutive model construction system for soil and rock materials, characterized in that... The system includes, The network construction module is configured to: construct a full strain parameter quantification correlation network for the target geotechnical material, which includes an intrinsic physical layer, a response feature layer, and a constitutive control layer. This network is used to establish logical connections between the parameters of the target constitutive model and multiple specified intrinsic physical layer parameters. The data acquisition module is configured to: acquire the original physical layer parameter values ​​input by the user; The data processing module is configured to: perform calculations based on the aforementioned full-strain parameter quantization correlation network to obtain an initial set of constitutive model parameters Π0 that can fully describe the full-strain mechanical characteristics of the target soil and rock material. The initial set of constitutive model parameters Π0 includes the input intrinsic physical layer parameter values, the calculated target constitutive model parameter values, and other constitutive model parameter values ​​required to determine displacement. These other constitutive model parameters include the initial shear modulus located in the response feature layer. and the slope of the critical state line And the plastic modulus interpolation parameters located in the constitutive control layer The calculation includes using a preset parameter quantization correlation formula to calculate the target constitutive model parameter values ​​of the response feature layer and constitutive control layer layer by layer based on the intrinsic physical layer parameter values; the parameter quantization correlation formula includes a reference shear strain for calculating the response feature layer parameters. The first formula and the boundary surface model shape parameters used to calculate the constitutive control layer parameters The second formula; the first formula is ,in, It is the compression index; Porosity; It has an over-consolidation ratio; and The fitting coefficients are obtained based on experimental fitting calibration; the second formula is... ,in, The rebound index; The initial boundary surface shape parameters are reference values ​​obtained based on the modified Cambridge model yield surface shape. The fitting coefficients are obtained based on experimental fitting calibration; and, The predicted displacement value and the deviation between the predicted displacement value and the measured displacement value are calculated based on the aforementioned initial set of constitutive model parameters Π0; and... When the deviation exceeds a predetermined threshold, a source-tracing adaptive algorithm is initiated. This algorithm is configured to: identify one or more intrinsic physical layer parameters causing the aforementioned deviation; adjust and iteratively calculate the displacement values ​​of the one or more intrinsic physical layer parameters until the aforementioned deviation does not exceed the predetermined threshold; and determine the constitutive model parameter set Π of the target geotechnical material based on the adjusted intrinsic physical layer parameter values. e .

8. The system according to claim 7, characterized in that, It also includes a model dynamic evolution module, which is configured to: for the Nth excavation layer during construction, determine the constitutive model parameter set Π for that layer based on the measured displacement values ​​monitored at that layer. e_N Subsequently, based on the constitutive model parameter set Π of the Nth excavation layer... e_N Determine the initial set Π of constitutive model parameters for the (N+1)th excavation layer 0_N+1 At this point, the initial set of constitutive model parameters Π for the (N+1)th excavation layer 0_N+1 The initial values ​​of the original physical layer parameters are based on the constitutive model parameter set Π of the Nth excavation layer. e_N The corresponding value in the value is taken; where N is an integer greater than or equal to 1.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1-6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-6.

Images