Roadbed top surface equivalent maximum bearing capacity determination method and device considering seasonal effect

Abstract

The invention provides a roadbed top surface equivalent maximum bearing capacity determination method and device considering a seasonal effect, and relates to the technical field of road engineering.The method specifically comprises the steps that firstly, a sample is manufactured based on a roadbed to be evaluated, and an environmental parameter combination is randomly generated according to the annual seasonal parameter range; constructing a simulation environment test to obtain related parameters and roadbed top surface equivalent maximum bearing capacity, constructing a response prediction model through data driving, weighting roadbed characteristics and response parameters, analyzing to generate a comprehensive evaluation coefficient, comprehensively evaluating roadbed performance, and optimizing an environment parameter combination by using a genetic algorithm in combination with the response prediction model. And constructing a bearing capacity prediction model by means of a long-short-term memory network, and dynamically predicting the equivalent maximum bearing capacity of the roadbed in the current time period. According to the method, the defects of static analysis in the prior art are overcome, and the problem that the result is inaccurate due to inconformity between reality and hypothesis in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of road engineering, and specifically provides a method and device for determining the equivalent maximum bearing capacity of the subgrade top surface considering seasonal effects. Background Art

[0002] The frost heaving and thaw settlement properties of seasonal frozen soil have a great impact on the stability of subgrades in seasonal frozen regions, and the water content of subgrade soil directly affects the degree of frost heaving and thaw settlement of subgrades. With the influence of global climate change, the area of permafrost regions is gradually decreasing, and the area of seasonal frozen soil regions is constantly expanding. Existing research shows that the void ratio and pore size distribution inside subgrade soil in seasonal frozen soil regions will change with seasonal effects, and thus it will inevitably suffer from frost heave and thaw settlement diseases. Therefore, studying the maximum bearing capacity of subgrades in seasonal frozen soil regions has important theoretical and practical significance.

[0003] In the prior art, a method and application for determining the maximum bearing capacity of a subgrade and the equivalent maximum bearing capacity of the subgrade top surface considering seasonal effects, with the authorization announcement number of CN114547747B, includes the following steps: If Error c ≤5%, then the overall maximum bearing capacity of the subgrade meets the convergence requirement, and the estimated finite element maximum bearing capacity and the estimated finite element maximum bearing capacity of each node of the subgrade at different frost heave and thaw stages are obtained; if Error c >5%, then the overall maximum bearing capacity of the subgrade does not meet the convergence requirement, return to step 12 to reset the system parameters of the finite element software to adjust the preset finite element maximum bearing capacity of node j and start iteration until the overall maximum bearing capacity of the subgrade meets the convergence requirement. After the overall maximum bearing capacity of the subgrade meets the convergence requirement, the ultimate settlement value of the subgrade top surface under static load is output through the drawing function of the finite element software.

[0004] However, there are still the following deficiencies. From the above statements, it can be seen that the prior art is based on finite element software and determines the overall maximum bearing capacity and the ultimate settlement value of the subgrade by setting an error threshold and iteratively adjusting the system parameters. This method depends on the assumptions and settings of the finite element model, and for complex and changeable actual situations, the flexibility and adaptability of the model are poor. Once the actual situation does not match the model assumptions, it will lead to problems such as too many iteration times, inaccurate results, or even inability to converge.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The purpose of the present invention is to provide a method and device for determining the equivalent maximum bearing capacity of the subgrade top surface considering seasonal effects, so as to solve the problems raised in the above background art.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for determining the equivalent maximum bearing capacity of the subgrade top surface considering seasonal effects, the specific steps include:

[0009] S1. Based on the subgrade to be evaluated, multiple samples are manufactured, the seasonal parameter range for the whole year is determined, and based on this range, multiple groups of environmental parameter combinations are randomly generated. Based on different environmental parameter combinations, simulation environments are respectively constructed to conduct simulation tests on the samples, and the subgrade characteristic parameters, the subgrade response parameters under the load, and the equivalent maximum bearing capacity of the subgrade top surface are obtained;

[0010] S2. A response prediction model is constructed. Using different environmental parameter combinations as inputs, and the corresponding subgrade characteristic parameters, subgrade response parameters, and equivalent maximum bearing capacity of the subgrade top surface as labels, the response prediction model is trained;

[0011] S3. The subgrade characteristic parameters and the subgrade response parameters are respectively subjected to linear weighted processing to generate a subgrade bearing coefficient and a subgrade deformation coefficient. The subgrade bearing coefficient and the subgrade deformation coefficient are subjected to data processing and correlation analysis to generate a comprehensive evaluation coefficient;

[0012] S4. Using the seasonal parameter range for the whole year as a constraint condition, and the environmental parameter combination as an individual in the initial population, with the maximization of the comprehensive evaluation coefficient as the optimization goal, the environmental parameter combination is iteratively optimized based on the genetic algorithm and the response prediction model to determine the optimal value of the environmental parameter combination. The optimal value of the environmental parameter combination is input into the response prediction model to obtain the optimal value of the corresponding equivalent maximum bearing capacity of the subgrade top surface;

[0013] S5. A bearing capacity prediction model is constructed based on a long short-term memory network. Using the time series data of the environmental parameters in the historical n time periods to construct a simulation environment to conduct simulation tests on the samples, and obtaining the equivalent maximum bearing capacity of the subgrade top surface in the historical n time periods. Using the time series data of the environmental parameters in the historical n time periods and the optimal value of the equivalent maximum bearing capacity of the subgrade top surface as inputs, and the corresponding equivalent maximum bearing capacity of the subgrade top surface as labels, the bearing capacity prediction model is trained;

[0014] S6. Input the time series data of the environmental parameters in the current time period and the optimal value of the equivalent maximum bearing capacity of the subgrade top surface into the trained bearing capacity prediction model to obtain the equivalent maximum bearing capacity of the subgrade top surface in the current time period.

[0015] Furthermore, the environmental parameters include temperature, precipitation, evaporation, wind speed, and sunshine duration. The subgrade characteristic parameters include shear strength, tensile strength, soil water content, and soil density. The subgrade response parameters include stress, strain, settlement, deflection value, and amplitude.

[0016] Furthermore, the individuals in the initial population are constructed as follows:

[0017] Label the initial population as Q, and the initial population Q = {Q1, Q2, …, Q j , …, Q n}, where Q j is the j-th individual in the initial population, j is the index of the individuals in the initial population, and j ∈ [1, n], where n is the number of individuals in the initial population. Q j = {T j , P j , E j , v j , S j}, where T j , P j , E j , v j , S j are the temperature, precipitation, evaporation, wind speed, and sunshine duration of the j-th individual respectively.

[0018] Furthermore, the subgrade characteristic parameters are linearly weighted to generate the subgrade bearing coefficient, and the formula is as follows:

[0019]

[0020] where CZxs j is the subgrade bearing coefficient of the j-th individual. The subgrade bearing coefficient is used to comprehensively evaluate the subgrade bearing capacity from four aspects: shear strength, tensile strength, soil water content, and soil density;

[0021] In the formula, τ j is the shear strength of the j-th individual, f j is the tensile strength of the j-th individual, ω j is the soil water content of the j-th individual, ρ j is the soil density of the j-th individual, ω1 is the lower limit of the ideal value of the soil water content, and ω2 is the upper limit of the ideal value of the soil water content;

[0022] In the formula, α1 is the weight coefficient of the shear strength, α2 is the weight coefficient of the tensile strength, α3 is the weight coefficient of the soil water content, and α4 is the weight coefficient of the soil density. On the basis of α1 + α2 + α3 + α4 = 1, let 0 < α3 < α2 < α4 < α1 < 1.

[0023] Furthermore, the subgrade response parameters are linearly weighted to generate the subgrade deformation coefficient, and the formula is as follows:

[0024] BXxs j = β1·σ j + β2·εj + β3·C j + β4·W j + β5·A j

[0025] Among them, BXxs j is the subgrade deformation coefficient of the j-th individual. The subgrade deformation coefficient is used to comprehensively evaluate the subgrade deformation trend from six aspects: stress, strain, settlement, deflection value, and amplitude;

[0026] In the formula, σ j is the stress of the j-th individual, ε j is the strain of the j-th individual, C j is the settlement of the j-th individual, W j is the deflection value of the j-th individual, A j is the amplitude of the j-th individual;

[0027] In the formula, β1 is the weight coefficient of stress, β2 is the weight coefficient of strain, β3 is the weight coefficient of settlement, β4 is the weight coefficient of deflection value, β5 is the weight coefficient of amplitude. On the basis of β1 + β2 + β3 + β4 + β5 = 1, let 0 < β5 < β2 < β1 < β4 < β3 < 1;

[0028] Perform data processing and correlation analysis on the subgrade bearing coefficient and the subgrade deformation coefficient to generate a comprehensive evaluation coefficient. The basis formula is as follows:

[0029]

[0030] Among them, ZPxs j is the comprehensive evaluation coefficient of the j-th individual. The comprehensive evaluation coefficient is used to combine the subgrade bearing coefficient and the subgrade deformation coefficient to comprehensively evaluate the equivalent bearing capacity of the subgrade top surface;

[0031] In the formula, γ1 is the weight coefficient of the subgrade bearing coefficient of the j-th individual, γ2 is the weight coefficient of the subgrade deformation coefficient of the j-th individual, and the specific values of γ1 and γ2 are determined by the analytic hierarchy process.

[0032] Furthermore, the specific process of step S4 is as follows:

[0033] Seek a balance point between the subgrade bearing coefficient and the subgrade deformation coefficient to maximize the comprehensive evaluation coefficient. Then, with the comprehensive evaluation coefficient ZPxs jMaximization is used as the optimization objective to iteratively optimize the initial population Q, that is, perform selection, crossover, and mutation operations on the individuals in the initial population Q. During the iterative optimization process, constraint conditions need to be set, that is, set the maximum and minimum values of temperature, precipitation, evaporation, wind speed, and sunshine duration respectively. Within the constraint ranges of temperature, precipitation, evaporation, wind speed, and sunshine duration, iteratively optimize the initial population Q. Specifically, select the individuals with the top comprehensive evaluation coefficients as the parents, and through the crossover operation, exchange and combine the genes of the parent individuals to generate new individuals. Then, perform mutation operations on the genes of temperature, precipitation, evaporation, wind speed, and sunshine duration in the newly generated individuals, and repeat the selection, crossover, and mutation operations until the predetermined number of iterations is reached;

[0034] After iteratively optimizing the initial population Q, label the optimal individual as Q j1 ={T j1 ,P j1 ,E j1 ,v j1 ,S j1}, and the optimal values of the environmental parameter combinations are temperature T j1 , precipitation P j1 , evaporation E j1 , wind speed v j1 , and sunshine duration S j1 .

[0035] To achieve the above object, the present invention also provides the following technical solutions:

[0036] A device for determining the equivalent maximum bearing capacity of the subgrade top surface considering seasonal effects, the system is used to execute any one of the above-mentioned methods for determining the equivalent maximum bearing capacity of the subgrade top surface considering seasonal effects, including:

[0037] A data acquisition module, which is used to manufacture multiple samples based on the subgrade to be evaluated, determine the annual seasonal parameter range, and randomly generate multiple groups of environmental parameter combinations based on this range. Build simulation environments based on different environmental parameter combinations to conduct simulation tests on the samples, and obtain subgrade characteristic parameters, subgrade response parameters under load, and the equivalent maximum bearing capacity of the subgrade top surface;

[0038] A training set construction module, which is used to construct a response prediction model, use different environmental parameter combinations as inputs, and the corresponding subgrade characteristic parameters, subgrade response parameters, and equivalent maximum bearing capacity of the subgrade top surface as labels to train the response prediction model;

[0039] A data calculation module, which is used to perform linear weighted processing on the subgrade characteristic parameters and subgrade response parameters respectively to generate a subgrade bearing coefficient and a subgrade deformation coefficient, and perform data processing and correlation analysis on the subgrade bearing coefficient and the subgrade deformation coefficient to generate a comprehensive evaluation coefficient;

[0040] A data optimization module, which uses the seasonal parameter range of the whole year as a constraint condition, and the environmental parameter combination as an individual in the initial population. With the maximization of the comprehensive evaluation coefficient as the optimization goal, it iteratively optimizes the environmental parameter combination based on the genetic algorithm and the response prediction model to determine the optimal value of the environmental parameter combination, and inputs the optimal value of the environmental parameter combination into the response prediction model to obtain the optimal value of the equivalent maximum bearing capacity of the subgrade top surface;

[0041] A bearing capacity prediction model construction module, which constructs a bearing capacity prediction model based on a long short-term memory network, constructs a simulation environment with the time series data of environmental parameters in the historical n time periods to conduct simulation tests on samples, obtains the equivalent maximum bearing capacity of the subgrade top surface in the historical n time periods, and uses the time series data of environmental parameters and the optimal value of the equivalent maximum bearing capacity of the subgrade top surface in the historical n time periods as inputs, and the corresponding equivalent maximum bearing capacity of the subgrade top surface as labels to train the bearing capacity prediction model;

[0042] A test set construction module, which inputs the time series data of environmental parameters and the optimal value of the equivalent maximum bearing capacity of the subgrade top surface in the current time period into the trained bearing capacity prediction model to obtain the equivalent maximum bearing capacity of the subgrade top surface in the current time period.

[0043] Compared with the prior art, the beneficial effects of the present invention are:

[0044] By manufacturing multiple samples based on the subgrade to be evaluated, determining the seasonal parameter range of the whole year and randomly generating multiple groups of environmental parameter combinations, and constructing a simulation environment for testing to obtain relevant parameters and the equivalent maximum bearing capacity of the subgrade top surface, this process fully considers the influence of seasonal changes on the subgrade, changes the situation of the prior art relying solely on the assumptions of the finite element model, can be more in line with the actual complex and changeable environment, and improves the adaptability to the actual situation.

[0045] Constructing a response prediction model and training it with different environmental parameter combinations as inputs and the corresponding subgrade characteristic parameters, subgrade response parameters, and equivalent maximum bearing capacity of the subgrade top surface as labels. The data-driven method reduces the dependence on the assumptions of the finite element model, makes the results more accurate and reliable, and overcomes the problem that the results of the prior art are inaccurate due to the inconsistency between the actual situation and the assumptions.

[0046] Linearly weighting the subgrade characteristic parameters and subgrade response parameters to generate a bearing coefficient and a deformation coefficient, and then performing data processing and correlation analysis to obtain a comprehensive evaluation coefficient, comprehensively considering various factors to evaluate the subgrade performance and bearing capacity, which is more comprehensive and accurate than simply determining the maximum bearing capacity in the prior art.

[0047] Taking the annual seasonal parameter range as a constraint condition, the genetic algorithm and the response prediction model are used to iteratively optimize the environmental parameter combination, determine the optimal value and input it into the model to obtain the optimal value of the equivalent maximum bearing capacity of the subgrade top surface, avoiding the iterative problems caused by the inconsistency between the actual situation and the assumptions in the prior art, and improving the calculation efficiency and result accuracy.

[0048] Based on the long short-term memory network, a bearing capacity prediction model is constructed and trained using the historical environmental parameter time series data and the relevant optimal values, which can capture the long-term dependence relationship of the time series and realize the dynamic prediction of the equivalent maximum bearing capacity of the subgrade top surface in the current time period, making up for the deficiency of only static analysis in the prior art and better adapting to the change of environmental parameters over time. Brief Description of the Drawings

[0049] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0050] Figure 2 It is a block diagram of the module composition of the present invention. Detailed Embodiments

[0051] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the following further elaborates on the present invention in conjunction with specific embodiments.

[0052] It should be noted that unless otherwise defined, the technical terms or scientific terms used in the present invention should have the ordinary meaning understood by those with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar terms used in the present invention do not represent any order, quantity or importance, but are only used to distinguish different components. The terms such as "including" or "comprising" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left", "right" are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0053] Embodiment 1:

[0054] Please refer to Figure 1 , the present invention provides a technical solution:

[0055] A method for determining the equivalent maximum bearing capacity of the subgrade top surface considering seasonal effects, the specific steps include:

[0056] S1. Manufacture multiple samples based on the subgrade to be evaluated, determine the seasonal parameter range for the whole year, randomly generate multiple groups of environmental parameter combinations based on this range, construct simulation environments based on different environmental parameter combinations respectively to conduct simulation tests on the samples, and obtain the subgrade characteristic parameters, the subgrade response parameters under load, and the equivalent maximum bearing capacity of the subgrade top surface.

[0057] On the basis of the above embodiments, the collected environmental parameters, subgrade characteristic parameters, subgrade response parameters under load, and the equivalent maximum bearing capacity of the subgrade top surface are all in units of days, that is, collect the environmental parameters, subgrade characteristic parameters, subgrade response parameters under load, and the equivalent maximum bearing capacity of the subgrade top surface every day.

[0058] On the basis of the above embodiments, obtain the temperature range, precipitation range, evaporation range, wind speed range, and sunshine duration range to be evaluated according to the statistics of the meteorological department, including the daily maximum temperature and daily minimum temperature, daily maximum precipitation and daily minimum precipitation, daily maximum evaporation and daily minimum evaporation, daily maximum wind speed and daily minimum wind speed, daily maximum sunshine duration and daily minimum sunshine duration.

[0059] On the basis of the above embodiments, construct simulation environments based on different environmental parameter combinations to conduct simulation tests on the samples, and obtain the subgrade characteristic parameters. The subgrade characteristic parameters include shear strength, tensile strength, soil moisture content, and soil density. The specific steps are as follows:

[0060] Place the samples in the simulation environment facilities, set the environmental conditions according to different environmental parameter combinations, and set the simulation period to not less than 30 days.

[0061] During the simulation period, select some samples of the subgrade to be evaluated every 7 days, conduct shear strength tests using a direct shear instrument or a triaxial shear instrument, record the shear failure loads of the samples under different vertical pressures under the simulation conditions of different environmental parameter combinations, calculate the shear strength according to the formula, and take the average value of multiple measurements as the shear strength of the subgrade.

[0062] Use a universal material testing machine every 7 days to apply an axial tensile force to the samples of the subgrade to be evaluated, record the load when the samples are broken, calculate the tensile strength of the subgrade according to the sample size, and take the average value of multiple measurements as the tensile strength of the subgrade.

[0063] Adopt the drying method to measure the soil moisture content of the subgrade to be evaluated every 5 days during the simulation period. Weigh a certain mass of wet soil samples, dry them in an oven at 105 - 110 °C until they reach a constant weight, calculate the soil moisture content through the mass difference before and after, and obtain the average value of the soil moisture content measured multiple times as the soil moisture content of the subgrade.

[0064] Using the core cutter method, during the process of simulating different seasons, the soil density of the subgrade to be evaluated is measured every 10 days. Take soil samples with a core cutter, weigh the total mass of the core cutter and the soil sample, subtract the mass of the core cutter, and then divide by the volume of the core cutter to obtain the soil density. Calculate the average value of the soil density measured multiple times as the soil density of the subgrade.

[0065] Based on the above embodiments, a simulation environment is constructed based on different combinations of environmental parameters to conduct simulation tests on the samples, and the subgrade response parameters under load are obtained. The subgrade response parameters include stress, strain, settlement, deflection value, and amplitude. The specific steps are as follows:

[0066] Place the samples in the simulation environment facilities, set the environmental conditions according to different combinations of environmental parameters, and set the simulation period to not less than 30 days;

[0067] After the simulation environment has been running stably for 3 days, start applying load for testing;

[0068] Static load test:

[0069] Use a bearing plate loading device, place a circular bearing plate on the surface of the sample, and apply vertical load step by step according to the predetermined loading level. After each load is applied:

[0070] The selected settlement measurement points are evenly distributed around the periphery and at the center of the bearing plate to obtain the settlement distribution on the surface of the sample. Measure the settlement of the sample surface through a high-precision displacement sensor, and take the average value of multiple measurement points as the settlement of the subgrade;

[0071] Record the change of stress with depth under each load level. Measure the vertical stress at different depths with the stress sensors pre-installed inside the sample, and take the vertical stress at the maximum depth as the stress of the subgrade;

[0072] Paste resistance strain gauges on the surface of the sample and connect them to a strain measuring instrument to measure the vertical and horizontal strains on the surface of the sample under each load level, and take the strain under the ultimate load as the strain of the subgrade;

[0073] Dynamic load test:

[0074] Use a dynamic triaxial testing machine to apply an axial dynamic load in the form of a sine wave to the cylindrical sample, and the load frequency can be set to different frequencies such as 0.1Hz, 1Hz, 5Hz, etc.;

[0075] Measure the vibration acceleration of the sample through an acceleration sensor, and calculate the amplitude according to the integral relationship between acceleration and time.

[0076] Place a falling weight deflectometer on the surface of the sample in the simulation environment. By controlling the mass and falling height of the falling weight, simulate the impact load during vehicle driving. At the moment of the falling weight impact:

[0077] Use a high-precision displacement sensor to measure the deflection value of the sample surface under impact load. The deflection measurement points are arranged at different distances around the impact point to obtain deflection value data, and the average value of multiple measurements is taken as the deflection value of the subgrade.

[0078] Based on the above embodiments, the acquisition method of the equivalent maximum bearing capacity of the subgrade top surface is as follows:

[0079] Conduct a static load test, gradually apply load using a load plate, record the corresponding settlement and deformation, establish a load-settlement curve according to the relationship between load and settlement, and determine the equivalent maximum bearing capacity of the subgrade top surface.

[0080] Based on the above embodiments, after collecting shear strength, tensile strength, soil moisture content, soil density, stress, strain, settlement, deflection value, and amplitude, perform maximum-minimum normalization processing on these parameters respectively, and then use the normalized data for subsequent analysis processing, so that in the subsequent analysis processing, various data can be analyzed and processed under the same dimension, avoiding the problem that some data are ignored due to different dimensions.

[0081] S2. Construct a response prediction model, use different combinations of environmental parameters as inputs, and the corresponding subgrade characteristic parameters, subgrade response parameters, and equivalent maximum bearing capacity of the subgrade top surface as labels to train the response prediction model;

[0082] Based on the above embodiments, the response prediction model is composed of a deep learning network based on a multi-layer perceptron. The deep neural network of the multi-layer perceptron includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer. The first hidden layer, the second hidden layer, and the third hidden layer all have at least two neurons, and all use ReLU as the activation function;

[0083] In this embodiment, the input features of the deep learning network of the multi-layer perceptron include: temperature, precipitation, evaporation, wind speed, and sunshine duration, a total of 5 features.

[0084] The structure of the deep learning network of the multi-layer perceptron is as follows:

[0085] Input layer: Receive the input of 5 features;

[0086] First hidden layer: Has 128 neurons and uses ReLU as the activation function;

[0087] Second hidden layer: Has 64 neurons and also uses the ReLU activation function;

[0088] Third hidden layer: Has 32 neurons and uses the ReLU activation function;

[0089] Output layer: It has 3 neurons, namely the subgrade characteristic parameters, subgrade response parameters, and equivalent maximum bearing capacity of the subgrade top surface.

[0090] The process of training the response prediction model is as follows:

[0091] Using different combinations of environmental parameters as input quantities, and the subgrade characteristic parameters, subgrade response parameters, and equivalent maximum bearing capacity of the subgrade top surface as output labels for training. The mean square error is used as the loss function. When the mean square error is within the range of [0, 0.01], the training of the response prediction model is completed.

[0092] S3. Linearly weight the subgrade characteristic parameters and subgrade response parameters respectively to generate the subgrade bearing coefficient and subgrade deformation coefficient. Perform data processing and correlation analysis on the subgrade bearing coefficient and subgrade deformation coefficient to generate a comprehensive evaluation coefficient;

[0093] Based on the above embodiments, linearly weight the subgrade characteristic parameters to generate the subgrade bearing coefficient. The formula is as follows:

[0094]

[0095] where, CZxs j is the subgrade bearing coefficient of the jth individual. The subgrade bearing coefficient is used to comprehensively evaluate the subgrade bearing capacity from four aspects: shear strength, tensile strength, soil water content, and soil density. And the larger the subgrade bearing coefficient, the greater the subgrade bearing capacity;

[0096] In the formula, τ j is the shear strength of the jth individual, f j is the tensile strength of the jth individual, ω j is the soil water content of the jth individual, ρ j is the soil density of the jth individual, ω1 is the lower limit of the ideal value of the soil water content, and ω2 is the upper limit of the ideal value of the soil water content.

[0097] On the above basis, it should be noted that when the shear strength τ j increases, the ability of the subgrade soil to resist shear failure increases, and then the ability to carry the load increases, thus making the subgrade bearing coefficient CZxs j increase; when the tensile strength f j increases, the ability of the subgrade to resist tensile failure is enhanced, and it can better withstand the tensile action transmitted from the pavement structure and vehicle load, thereby improving the overall bearing stability of the subgrade, and thus making the subgrade bearing coefficient CZxs j increase; when the soil density ρ jThe increase means that the number of soil particles per unit volume increases and the arrangement is more dense, and the interaction between particles is enhanced, which enables the roadbed to withstand greater loads, thereby increasing the roadbed bearing coefficient CZxs j Increase; when ω1≤ω j ≤ω2, soil moisture content ω j In the ideal water content range, the cohesion and friction between soil particles can be maintained in a relatively stable and reasonable state. As the water content increases appropriately within this range, the lubrication between soil particles is enhanced, which is conducive to the uniform distribution of stress and the improvement of the bearing capacity of the roadbed, thereby making the roadbed bearing coefficient CZxs j Increase; when ω j >ω2, soil moisture content ω j If the soil is too high, it will show a soft plastic or even fluid plastic state, the effective stress between particles will decrease, the shear strength will be greatly reduced, and the ability of the roadbed to resist deformation and bear load will drop sharply, thus making the roadbed bearing coefficient CZxs j Decrease, when ω j When <ω1, the soil is too dry, the cohesion between particles is weakened, the overall density and stability of the roadbed deteriorate, and the bearing capacity of the roadbed is also weakened, so that the roadbed bearing coefficient CZxs j Reduce.

[0098] Therefore, when ω1≤ω j When ≤ω2, the roadbed bearing coefficient CZxs j and shear strength τ j , tensile strength f j , soil moisture contentω j , soil density ρ j are positively correlated. j >ω2 or ω j <ω1, the roadbed bearing coefficient CZxs j and shear strength τ j , tensile strength f j , soil density ρ j are positively correlated, and the roadbed bearing coefficient CZxs j and soil moisture contentω j Therefore, the above weighted summation formula is used to characterize the roadbed bearing coefficient CZxs j and shear strength τ j , tensile strength f j , soil moisture contentω j , soil density ρ j The functional relationship between them.

[0099] In the formula, α1 is the weight coefficient of shear strength, α2 is the weight coefficient of tensile strength, α3 is the weight coefficient of soil moisture content, and α4 is the weight coefficient of soil density;

[0100] The magnitude setting of the weight coefficient is as follows:

[0101] Since the shear strength plays a crucial role in the bearing capacity of the subgrade. When external forces such as vehicle loads act on the subgrade, shear force is one of the main factors leading to subgrade failure. If the shear strength of the subgrade is insufficient, shear deformation and failure are likely to occur, such as slope instability and shear cracks on the road surface, which will directly and seriously affect the bearing capacity and stability of the subgrade. Therefore, a relatively large value is usually assigned to α1.

[0102] Soil density reflects the degree of compaction of soil particles. The greater the density, the stronger the interaction force between particles, and the higher the overall strength and stability of the subgrade. Good density can enable the subgrade to better transfer and disperse forces when bearing loads, reducing deformation, and the effect of improving the bearing capacity of the subgrade is relatively significant. Therefore, α4 generally also takes a relatively large value, second only to α1.

[0103] Although the tensile strength plays an important role in preventing the expansion of subgrade cracks, etc., under normal actions such as general vehicle loads, the tensile stress on the subgrade is not the main failure factor compared to the shear force and the force generated by the interaction between soil particles. Only in some special working conditions, such as tensile caused by low-temperature shrinkage and uneven settlement, etc., the influence of the tensile strength will be more prominent. Therefore, the value of α2 will be less than α4.

[0104] Soil water content has an important influence on the bearing capacity of the subgrade, but its effect is two-sided. Only under moderate conditions is it beneficial to improve the bearing capacity, and its influence on the bearing capacity is indirectly achieved by affecting the arrangement of soil particles, etc. Compared with factors such as shear strength and soil density that directly determine the mechanical properties of the subgrade, its contribution to the bearing capacity is relatively small, and it needs to be comprehensively considered according to the specific ideal water content range. Therefore, α3 usually takes a relatively small value.

[0105] To sum up, on the basis of α1 + α2 + α3 + α4 = 1, let 0 < α3 < α2 < α4 < α1 < 1.

[0106] As an implementation manner, the value range of α1 is 0.4 - 1, the value range of α2 is 0.2 - 0.4, the value range of α3 is 0 - 0.2, and the value range of α4 is 0.3 - 0.45. The specific values are set by technicians according to the actual situation and are not limited here.

[0107] On the basis of the above embodiments, the subgrade response parameters are linearly weighted to generate a subgrade deformation coefficient, and the formula is as follows:

[0108] BXxs j = β1·σj + β2·ε j + β3·C j + β4·W j + β5·A j

[0109] Among them, BXxs j is the subgrade deformation coefficient of the j-th individual. The subgrade deformation coefficient is used to comprehensively evaluate the subgrade deformation trend from six aspects of stress, strain, settlement, deflection value, and amplitude. Moreover, the larger the subgrade deformation coefficient, the more obvious the subgrade deformation and the smaller the subgrade bearing capacity;

[0110] In the formula, σ j is the stress of the j-th individual, ε j is the strain of the j-th individual, C j is the settlement of the j-th individual, W j is the deflection value of the j-th individual, A j is the amplitude of the j-th individual;

[0111] On this basis, it should be noted that as the stress σ j increases, the strain also increases accordingly, which in turn causes the deformation of the subgrade. When the stress exceeds the elastic limit of the subgrade material, the subgrade will undergo plastic deformation. The continuous increase of the stress σ j will lead to the continuous accumulation of plastic deformation, the degree of subgrade deformation will intensify, and thus the subgrade deformation coefficient BXxs j will increase; as the strain ε j increases, it means that the degree of deformation inside the subgrade material intensifies. In the elastic stage, the strain and stress are linearly related. As the strain ε j increases, the subgrade gradually deviates from its initial state. When the strain ε j exceeds a certain limit, the internal structure of the material begins to undergo irreversible changes, which will further promote the subgrade as a whole to produce greater deformation, and thus the subgrade deformation coefficient BXxs j will increase; as the settlement C j increases, whether it is static load or dynamic load, when the load borne by the subgrade exceeds its bearing capacity, the soil particles will be further compressed and the pores will decrease, resulting in the gradual sinking of the subgrade surface. A large settlement will not only affect the smoothness of the road surface but may also cause diseases such as road surface cracking and collapse. The continuous increase of the settlement indicates the continuous development of the subgrade deformation in the vertical direction, and thus the subgrade deformation coefficient BXxs j will increase; as the deflection value W j increases, it indicates that under the action of impact load or long-term traffic load, a greater vertical displacement has occurred on the subgrade surface, and thus the subgrade deformation coefficient BXxs j will increase; as the amplitude A jIncreases, indicating that the vibration response of the subgrade under dynamic excitation intensifies. When the amplitude of the subgrade increases, it shows that the vibration amplitude of the subgrade under these dynamic loads exceeds the normal range. Excessive amplitude will cause changes in the interaction between subgrade soil particles, affecting the friction and cohesion between particles, and further causing more obvious deformation of the subgrade, thereby making the subgrade deformation coefficient BXxs j increase. Therefore, the subgrade deformation coefficient BXxs j and stress σ j , strain ε j , settlement C j , deflection value W j , and amplitude A j are all positively correlated. Therefore, the calculation formula for the subgrade deformation coefficient BXxs in the form of weighted summation is set as above. j

[0112] In the formula, β1 is the weight coefficient of stress, β2 is the weight coefficient of strain, β3 is the weight coefficient of settlement, β4 is the weight coefficient of deflection value, and β5 is the weight coefficient of amplitude;

[0113] The magnitudes of the weighting coefficients are set as follows:

[0114] Since the settlement is an intuitive manifestation of the deformation of the subgrade in the vertical direction, it has the most direct impact on the flatness, driving comfort, and safety of the road. Excessive settlement will cause diseases such as potholes and cracks on the road surface, seriously affecting the use function of the road and even possibly causing traffic safety accidents. When evaluating the impact of subgrade deformation on the project, the settlement is usually the most critical factor, so the largest weight is assigned to it.

[0115] The deflection value is the vertical deformation value generated at the wheel gap position on the subgrade surface under the action of the specified standard axle load. It can comprehensively reflect the overall strength and deformation performance of the subgrade. The larger the deflection value, the greater the deformation of the subgrade under the load and the lower the bearing capacity, which has a greater impact on the stability and durability of the road structure. In road engineering, the deflection value is one of the important indicators for measuring the quality and rationality of the subgrade design. Its importance is second only to the settlement, so the weight ranks second.

[0116] Stress is the fundamental cause of subgrade deformation and determines the internal stress state of subgrade materials. Although stress is the source of deformation, its influence on subgrade deformation is reflected through intermediate links such as strain and does not directly reflect the macroscopic deformation of the subgrade like settlement and deflection value. In actual engineering, it is usually necessary to indirectly understand the influence of stress on subgrade deformation by measuring parameters such as strain, so its weight is relatively low and ranks third.

[0117] ​Strain directly reflects the deformation of subgrade materials and is an important index for describing the degree of subgrade deformation. However, it mainly focuses on the microscopic deformation inside subgrade materials. Compared with macroscopic deformation indexes such as settlement and deflection, the direct influence of strain on the overall deformation trend of the subgrade is relatively small, and it is more used to analyze the mechanical properties and deformation mechanisms of subgrade materials. When evaluating subgrade deformation, strain is an important reference index, but its importance is less than that of the previous several parameters, so its weight ranks fourth.

[0118] Amplitude reflects the severity of subgrade vibration under the action of vibration load. Although a large amplitude will cause repeated vibration and displacement of subgrade soil particles, accelerating the rearrangement of soil particles and the change of soil density, and long-term action will lead to cumulative deformation of the subgrade, in general subgrade engineering, vibration load is not the main load form, and its occurrence frequency and influence degree on subgrade deformation are relatively limited. Only in some special engineering environments, such as subgrades near railways, airport runways or large power equipment, will the influence of amplitude be more significant. Therefore, in general cases, the weight of amplitude is relatively the smallest.

[0119] To sum up, on the basis of β1 + β2 + β3 + β4 + β5 = 1, let 0 < β5 < β2 < β1 < β4 < β3 < 1.

[0120] As an implementation method, the value range of β1 is 0.15 - 0.25, the value range of β2 is 0.05 - 0.15, the value range of β3 is 0.35 - 0.6, the value range of β4 is 0.2 - 0.35, and the value range of β5 is 0.02 - 0.08. The specific values are set by technicians according to the actual situation and are not limited here.

[0121] On the basis of the above embodiments, data processing and correlation analysis are carried out on the subgrade bearing coefficient and the subgrade deformation coefficient to generate a comprehensive evaluation coefficient. The formula is as follows:

[0122]

[0123] Among them, ZPxs j is the comprehensive evaluation coefficient of the jth individual. The comprehensive evaluation coefficient is used to combine the subgrade bearing coefficient and the subgrade deformation coefficient to comprehensively evaluate the equivalent bearing capacity of the subgrade top surface. And the comprehensive evaluation coefficient ZPxs j The larger it is, the greater the equivalent bearing capacity of the subgrade top surface;

[0124] It should be noted that through the above description, it can be seen that the subgrade bearing coefficient CZxs j The larger it is, the greater the subgrade bearing capacity, and the subgrade deformation coefficient BXxs j The larger it is, the smaller the subgrade bearing capacity. Therefore, the comprehensive evaluation coefficient ZPxsj is positively correlated with the subgrade bearing coefficient CZxs j and the comprehensive evaluation coefficient ZPxs j is negatively correlated with the subgrade deformation coefficient BXxs j Therefore, the comprehensive evaluation coefficient ZPxs in the form of weighted summation is set j calculation formula;

[0125] In the formula, γ1 is the weight coefficient of the subgrade bearing coefficient of the jth individual, γ2 is the weight coefficient of the subgrade deformation coefficient of the jth individual, and the specific values of γ1 and γ2 are determined by the analytic hierarchy process. The specific logic is as follows:

[0126] Mark the two indexes of the subgrade bearing coefficient and the subgrade deformation coefficient of the individual, determine the relative importance values between them through the nine-scale method, and construct a judgment matrix. Among them, mark the index of the subgrade bearing coefficient as 1 and the index of the subgrade deformation coefficient as 2. The constructed judgment matrix [q uv 2×2 is:

[0127]

[0128] where u and v both represent the indexes of the coefficients, and u ∈ [1, 2], v ∈ [1, 2], indicating the importance of the coefficient with index u relative to the coefficient with index v for the comprehensive evaluation coefficient, q uv The specific values are determined by relevant experts using the 1-9 scoring method. q uv = 9 indicates that the coefficient with index u is extremely important for the comprehensive evaluation coefficient compared with the coefficient with index v, and q uv = 1 indicates that the coefficient with index u is extremely unimportant for the comprehensive evaluation coefficient compared with the coefficient with index v;

[0129] Divide each element value in the judgment matrix by the sum of its columns to obtain a normalized judgment matrix. Calculate the mean value of each row element value in the normalized judgment matrix, and use the mean value of the first row element value as the weight coefficient of the subgrade bearing coefficient of the individual, and use the mean value of the second row element value as the weight coefficient of the subgrade deformation coefficient of the individual. With the constraint that the sum of the scaled values is equal to 1, scale the two weight coefficients proportionally, and use the scaled values as the weights of the corresponding coefficients.

[0130] S4. Take the annual seasonal parameter range as the constraint condition, the environmental parameter combination as the individual in the initial population, take the maximization of the comprehensive evaluation coefficient as the optimization goal, and iteratively optimize the environmental parameter combination based on the genetic algorithm and the response prediction model to determine the optimal value of the environmental parameter combination. Input the optimal value of the environmental parameter combination into the response prediction model to obtain the optimal value of the equivalent maximum bearing capacity of the subgrade top surface;​

[0131] Based on the above embodiments, the individuals in the initial population are constructed as follows:

[0132] The initial population is designated as Q, and the initial population Q = {Q1, Q2, …, Q j , …, Q n}, where Q j is the j-th individual in the initial population, j is the index of the individuals in the initial population, and j ∈ [1, n], where n is the number of individuals in the initial population. Q j = {T j , P j , E j , v j , S j}, where T j , P j , E j , v j , S j are the temperature, precipitation, evaporation, wind speed, and sunshine duration of the j-th individual, respectively.

[0133] Based on the above embodiments, with the maximization of the comprehensive evaluation coefficient as the optimization goal, the environmental parameter combinations are iteratively optimized based on the genetic algorithm and the response prediction model to determine the optimal values of the environmental parameter combinations. The specific process is as follows:

[0134] Seek a balance between the subgrade bearing coefficient and the subgrade deformation coefficient to maximize the comprehensive evaluation coefficient. Then, with the maximization of the comprehensive evaluation coefficient ZPxs j as the optimization goal, perform iterative optimization on the initial population Q, that is, perform selection, crossover, and mutation operations on the individuals in the initial population Q. During the iterative optimization process, set constraint conditions, that is, set the maximum and minimum values of the temperature, precipitation, evaporation, wind speed, and sunshine duration respectively. Within the constraint range of the temperature, precipitation, evaporation, wind speed, and sunshine duration, perform iterative optimization on the initial population Q. Specifically, select the individuals with the top comprehensive evaluation coefficients as the parents. The top refers to the individuals in the top 50% of the comprehensive evaluation coefficients. Through the crossover operation, exchange and combine the genes of the parent individuals to generate new individuals. Then, perform mutation operations on the genes of the temperature, precipitation, evaporation, wind speed, and sunshine duration in the newly generated individuals, and repeat the selection, crossover, and mutation operations until the predetermined number of iterations is reached;

[0135] After performing iterative optimization on the initial population Q, designate the optimal individual as Q j1 = {T j1 , P j1 , E j1 , v j1 , S j1}, the optimal value of the environmental parameter combination is temperature T j1 , precipitation P j1 , evaporation E j1 , wind speed v j1 and sunshine duration S j1 .

[0136] S5. Construct a bearing capacity prediction model based on the long short-term memory network. Use the time series data of environmental parameters in the historical n time periods to construct a simulated environment to conduct simulation tests on the samples, obtain the equivalent maximum bearing capacity of the subgrade top surface in the historical n time periods, use the time series data of environmental parameters in the historical n time periods and the optimal value of the equivalent maximum bearing capacity of the subgrade top surface as inputs, and the corresponding equivalent maximum bearing capacity of the subgrade top surface as labels to train the bearing capacity prediction model;

[0137] On the basis of the above embodiments, the bearing capacity prediction model uses the mean square error to measure the difference between the model prediction value and the true value, selects the Adam optimizer to update the parameters of the model to minimize the loss function (mean square error), determines the number of training epochs and batch size, inputs the training data into the bearing capacity prediction model, and updates the parameters of the model through the backpropagation algorithm according to the set optimizer and loss function. After training, use the test set to evaluate the model, calculate indicators such as the loss value, accuracy, mean square error, and mean absolute error of the model on the test set, analyze the prediction results of the model, and observe the prediction accuracy of the model under different environmental parameters, whether there are biases or abnormal situations.

[0138] S6. Input the time series data of environmental parameters in the current time period and the optimal value of the equivalent maximum bearing capacity of the subgrade top surface into the trained bearing capacity prediction model to obtain the equivalent maximum bearing capacity of the subgrade top surface in the current time period.

[0139] On the basis of the above embodiments, the time period lengths of the historical n time periods and the current time period are the same. For example, the time period is 3 days for both, and the time intervals for collecting parameters in the historical n time periods and the current time period are the same. For example, every 1 day.

[0140] Please refer to Figure 2 , the present invention also provides a technical solution:

[0141] An apparatus for determining the equivalent maximum bearing capacity of the subgrade top surface considering seasonal effects, the apparatus is used to execute any one of the above methods for determining the equivalent maximum bearing capacity of the subgrade top surface considering seasonal effects, and includes:

[0142] A data acquisition module, which is used to manufacture multiple samples based on the subgrade to be evaluated, determine the seasonal parameter range throughout the year, randomly generate multiple groups of environmental parameter combinations based on this range, respectively construct simulation environments based on different environmental parameter combinations to conduct simulation tests on the samples, and obtain subgrade characteristic parameters, subgrade response parameters under load, and the equivalent maximum bearing capacity of the subgrade top surface;

[0143] A training set construction module, which is used to construct a response prediction model, use different environmental parameter combinations as inputs, and the corresponding subgrade characteristic parameters, subgrade response parameters, and equivalent maximum bearing capacity of the subgrade top surface as labels to train the response prediction model;

[0144] A data calculation module, which is used to perform linear weighted processing on both the subgrade characteristic parameters and the subgrade response parameters respectively to generate a subgrade bearing coefficient and a subgrade deformation coefficient, perform data processing and correlation analysis on the subgrade bearing coefficient and the subgrade deformation coefficient to generate a comprehensive evaluation coefficient;

[0145] A data optimization module, which is used to use the seasonal parameter range throughout the year as a constraint condition, the environmental parameter combination as an individual in the initial population, take the maximization of the comprehensive evaluation coefficient as the optimization goal, and iteratively optimize the environmental parameter combination based on the genetic algorithm and the response prediction model to determine the optimal value of the environmental parameter combination, and input the optimal value of the environmental parameter combination into the response prediction model to obtain the optimal value of the corresponding equivalent maximum bearing capacity of the subgrade top surface;

[0146] A bearing capacity prediction model construction module, which is used to construct a bearing capacity prediction model based on a long short-term memory network, construct a simulation environment with the time series data of environmental parameters in the past n time periods to conduct simulation tests on the samples, obtain the equivalent maximum bearing capacity of the subgrade top surface in the past n time periods, use the time series data of environmental parameters in the past n time periods and the optimal value of the equivalent maximum bearing capacity of the subgrade top surface as inputs, and the corresponding equivalent maximum bearing capacity of the subgrade top surface as labels to train the bearing capacity prediction model;

[0147] A test set construction module, which is used to input the time series data of environmental parameters and the optimal value of the equivalent maximum bearing capacity of the subgrade top surface in the current time period into the trained bearing capacity prediction model to obtain the equivalent maximum bearing capacity of the subgrade top surface in the current time period.

[0148] All the above formulas are dimensionless and take their numerical calculations. The formula is a formula obtained by collecting a large amount of data for software simulation to get the closest to the real situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0149] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will appreciate that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0150] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units. They may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0151] As described above, the above are only specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application, and all should be covered by the protection scope of the present application.

Claims

1. A method for determining the equivalent maximum bearing capacity of the roadbed top surface considering seasonal effects, characterized in that: The specific steps include: S1. Based on the roadbed to be evaluated, multiple samples are manufactured, the seasonal parameter range of the whole year is determined, and multiple sets of environmental parameter combinations are randomly generated based on the range. Simulated environments are constructed based on different environmental parameter combinations to simulate the samples, and the characteristic parameters of the roadbed and the roadbed response parameters under load and the equivalent maximum bearing capacity of the roadbed top surface are obtained; S2. Construct a response prediction model, take different combinations of environmental parameters as input, and use the corresponding roadbed characteristic parameters, roadbed response parameters, and roadbed top surface equivalent maximum bearing capacity as labels to train the response prediction model; S3. linearly weighting the roadbed characteristic parameters and the roadbed response parameters to generate a roadbed bearing coefficient and a roadbed deformation coefficient, and performing data processing and correlation analysis on the roadbed bearing coefficient and the roadbed deformation coefficient to generate a comprehensive evaluation coefficient; S4. Taking the seasonal parameter range of the whole year as the constraint condition, the environmental parameter combination as the individual in the initial population, and maximizing the comprehensive evaluation coefficient as the optimization goal, the environmental parameter combination is iteratively optimized based on the genetic algorithm and the response prediction model to determine the optimal value of the environmental parameter combination, and the optimal value of the environmental parameter combination is input into the response prediction model to obtain the corresponding optimal value of the equivalent maximum bearing capacity of the roadbed top surface; S5. Construct a bearing capacity prediction model based on the long short-term memory network, construct a simulation environment with the time series data of environmental parameters in the historical n time periods to simulate the sample, obtain the equivalent maximum bearing capacity of the roadbed top surface in the historical n time periods, use the time series data of environmental parameters in the historical n time periods and the optimal value of the equivalent maximum bearing capacity of the roadbed top surface as input, and use the corresponding equivalent maximum bearing capacity of the roadbed top surface as a label to train the bearing capacity prediction model; S6. Input the time series data of the environmental parameters of the current time period and the optimal value of the equivalent maximum bearing capacity of the roadbed top surface into the trained bearing capacity prediction model to obtain the equivalent maximum bearing capacity of the roadbed top surface in the current time period.

2. The method for determining the equivalent maximum bearing capacity of the roadbed top surface considering seasonal effects according to claim 1 is characterized in that: Environmental parameters include temperature, precipitation, evaporation, wind speed and sunshine duration; roadbed characteristic parameters include shear strength, tensile strength, soil moisture content and soil density; and roadbed response parameters include stress, strain, settlement, deflection and amplitude.

3. The method for determining the equivalent maximum bearing capacity of the roadbed top surface considering seasonal effects according to claim 2 is characterized in that: Construct individuals in the initial population. The specific process is as follows: The initial population is labeled as Q, and the initial population Q={Q1,Q2,…,Q j ,…,Q n }, Q j is the jth individual in the initial population, j is the index of the individual in the initial population, and j∈[1,n], n is the number of individuals in the initial population, Q j ={T j ,P j ,E j ,v j ,S j }, where T j ,P j ,E j ,v j ,S j are the temperature, precipitation, evaporation, wind speed and sunshine duration of the jth individual respectively.

4. The method for determining the equivalent maximum bearing capacity of the roadbed top surface considering seasonal effects according to claim 3 is characterized in that: The roadbed characteristic parameters are linearly weighted to generate the roadbed bearing coefficient based on the following formula: Among them, CZxs j is the jth individual roadbed bearing coefficient, which is used to comprehensively evaluate the roadbed bearing capacity from four aspects: shear strength, tensile strength, soil moisture content and soil density; In the formula, τ j is the shear strength of the jth individual, f j is the tensile strength of the jth individual, ω j is the soil moisture content of the jth individual, ρ j is the soil density of the jth individual, ω1 is the lower limit of the ideal soil moisture content, and ω2 is the upper limit of the ideal soil moisture content; In the formula, α1 is the weight coefficient of shear strength, α2 is the weight coefficient of tensile strength, α3 is the weight coefficient of soil moisture content, and α4 is the weight coefficient of soil density. On the basis of α1+α2+α3+α4=1, let 0<α3<α2<α4<α1<1.

5. The method for determining the equivalent maximum bearing capacity of the roadbed top surface considering seasonal effects according to claim 4 is characterized in that: The roadbed response parameters are linearly weighted to generate the roadbed deformation coefficient based on the following formula: BXxs j =β1·σ j +β2·e j +β3·C j +β4·W j +β5·A j Among them, BXxs j is the roadbed deformation coefficient of the jth individual. The roadbed deformation coefficient is used to comprehensively evaluate the roadbed deformation trend from six aspects: stress, strain, settlement, deflection value and amplitude; In the formula, σ j is the stress of the jth individual, ε j is the strain of the jth individual, C j is the sedimentation of the jth individual, W j is the deflection value of the jth individual, A j is the amplitude of the jth individual; Wherein, β1 is the weight coefficient of stress, β2 is the weight coefficient of strain, β3 is the weight coefficient of settlement, β4 is the weight coefficient of deflection value, and β5 is the weight coefficient of amplitude. On the basis of β1+β2+β3+β4+β5=1, let 0<β5<β2<β1<β4<β3<1; The roadbed bearing coefficient and roadbed deformation coefficient are processed and correlated to generate a comprehensive evaluation coefficient based on the following formula: Among them, ZPxs j is the comprehensive evaluation coefficient of the jth individual, which is used to combine the roadbed bearing coefficient and the roadbed deformation coefficient to comprehensively evaluate the equivalent bearing capacity of the roadbed top surface; Wherein, γ1 is the weight coefficient of the j-th individual roadbed bearing coefficient, γ2 is the weight coefficient of the j-th individual roadbed deformation coefficient, and the specific values ​​of γ1 and γ2 are determined by the hierarchical analysis method.

6. The method for determining the equivalent maximum bearing capacity of the roadbed top surface considering seasonal effects according to claim 5 is characterized in that: The specific process of step S4 is as follows: Find a balance point between the roadbed bearing coefficient and the roadbed deformation coefficient to maximize the comprehensive evaluation coefficient. j Taking maximization as the optimization goal, the initial population Q is iteratively optimized, that is, the individuals in the initial population Q are selected, crossed, and mutated. In the iterative optimization process, constraints are set, that is, the maximum and minimum values ​​of temperature, precipitation, evaporation, wind speed, and sunshine duration are set respectively. Within the constraints of temperature, precipitation, evaporation, wind speed, and sunshine duration, the initial population Q is iteratively optimized. Specifically, individuals with a comprehensive evaluation coefficient in the top are selected as parents. Through crossover operations, the genes of the parent individuals are exchanged and combined to generate new individuals. Then, the genes of temperature, precipitation, evaporation, wind speed, and sunshine duration in the newly generated individuals are mutated, and the selection, crossover, and mutation operations are repeated until the predetermined number of iterations is reached. After iterative optimization of the initial population Q, the optimal individual is marked as Q j1 ={T j1 ,P j1 ,E j1 ,v j1 ,S j1 }, the optimal value of the environmental parameter combination is temperature T j1 , precipitation P j1 , evaporation E j1 , wind speed v j1 and sunshine duration S j1 .

7. A device for determining the equivalent maximum bearing capacity of the top surface of a roadbed considering seasonal effects, the system being used to execute a method for determining the equivalent maximum bearing capacity of the top surface of a roadbed considering seasonal effects as claimed in any one of claims 1 to 6, characterized in that: include: The data acquisition module is used to manufacture multiple samples based on the roadbed to be evaluated, determine the seasonal parameter range throughout the year, and randomly generate multiple sets of environmental parameter combinations based on the range. Based on different environmental parameter combinations, simulation environments are constructed to simulate the samples to obtain the roadbed characteristic parameters and the roadbed response parameters under load and the equivalent maximum bearing capacity of the roadbed top surface; The training set construction module is used to construct the response prediction model. Different combinations of environmental parameters are used as input, and the corresponding roadbed characteristic parameters, roadbed response parameters, and roadbed top surface equivalent maximum bearing capacity are used as labels to train the response prediction model. A data calculation module is used to perform linear weighted processing on the roadbed characteristic parameters and the roadbed response parameters respectively to generate the roadbed bearing coefficient and the roadbed deformation coefficient, and to perform data processing and correlation analysis on the roadbed bearing coefficient and the roadbed deformation coefficient to generate a comprehensive evaluation coefficient; The data optimization module is used to use the seasonal parameter range of the whole year as the constraint condition, the environmental parameter combination as the individual in the initial population, and the maximization of the comprehensive evaluation coefficient as the optimization goal. The environmental parameter combination is iteratively optimized based on the genetic algorithm and the response prediction model to determine the optimal value of the environmental parameter combination, and the optimal value of the environmental parameter combination is input into the response prediction model to obtain the corresponding optimal value of the equivalent maximum bearing capacity of the roadbed top surface; The bearing capacity prediction model construction module is used to construct a bearing capacity prediction model based on a long short-term memory network, to construct a simulation environment with the time series data of environmental parameters in n historical time periods to simulate and test the samples, to obtain the equivalent maximum bearing capacity of the roadbed top surface in n historical time periods, and to train the bearing capacity prediction model with the time series data of environmental parameters in n historical time periods and the optimal value of the equivalent maximum bearing capacity of the roadbed top surface as input and the corresponding equivalent maximum bearing capacity of the roadbed top surface as a label; The test set construction module is used to input the time series data of the environmental parameters of the current time period and the optimal value of the equivalent maximum bearing capacity of the roadbed top surface into the trained bearing capacity prediction model to obtain the equivalent maximum bearing capacity of the roadbed top surface in the current time period.

Images