Freeze-thaw soil hydrothermal dynamic parameter inversion method and device and medium

Abstract

The invention discloses a freeze-thaw soil hydrothermal dynamic parameter inversion method and device and a medium. The method comprises the following steps: determining a freeze-thaw soil hydrothermal coupling model and to-be-inverted hydrothermal parameters; establishing a one-dimensional numerical calculation model of the frozen-thawed soil layer; determining a hydrothermal parameter combination scheme; establishing an objective function of errors of measured values and predicted values of the ground temperature and the volumetric water content at different depths; establishing a relationship between the input inversion hydrothermal parameters and the output freeze-thaw soil hydrothermal response by adopting a back propagation neural network; optimizing by utilizing a genetic algorithm to obtain a globally optimal solution of the target function; determining inversion values of the hydrothermal parameters; and substituting the inversion values of the hydrothermal parameters into the dimensional numerical calculation model to obtain predicted values of the ground temperature and the volumetric water content. The method can accurately reflect the evolution characteristics of the soil hydrothermal characteristics along with the time, thereby solving the problem of identifying the hydrothermal parameters of the freeze-thaw soil body under the condition of real-time change of the external environment.

Description

Technical Field

[0001] The present invention belongs to the technical field of geotechnical engineering, and more specifically, relates to a method, device and medium for inverting hydrothermal dynamic parameters of frozen-thawed soil. Background Art

[0002] Essentially, the occurrence of freeze-thaw disasters under the background of climate warming and humidification is the result of the hydrothermal coupling effect of the active layer. Therefore, in order to clarify the freeze-thaw disaster mechanism on the Qinghai-Tibet Plateau, it is urgent to explore the hydrothermal behavior of the active layer under the background of warming and humidification on the Qinghai-Tibet Plateau. Hydrothermal parameters are important parameters for effectively studying the hydrothermal response of frozen-thawed soil. Under the background of warming and humidification on the Qinghai-Tibet Plateau, the hydrothermal parameters of frozen-thawed soil change dynamically with the external environment. How to efficiently obtain hydrothermal dynamic parameters has always been a classic and hot issue in the geotechnical field.

[0003] In recent years, the inversion method based on observational data has become a practical method for more reasonably estimating material parameters. This method needs to solve two key problems. The first challenge is how to determine multiple target parameters simultaneously. Since there are a large number of parameters to be inverted in the hydrothermal coupling model, using the traditional direct search algorithm is likely to cause low computational efficiency and unable to obtain the global optimal solution. Currently, introducing a weight distribution coefficient has become a more common method to achieve the optimal solution of the multi-objective function. By subjectively defining the weight distribution coefficient, the multi-objective problem is transformed into a single-objective problem. Another challenge is the non-uniformity of the values of the coupling model parameters. In the inversion of hydrothermal parameters, due to the different orders of magnitude of unknown parameters such as hydrothermal parameters, the orders of magnitude of the partial derivatives of the corresponding parameters in the objective function vary greatly. Therefore, a reasonable parameter inversion method needs to be sought. The inversion method based on observational data has currently been applied to different fields, and important problems in the geotechnical engineering field such as slope stability, dam seepage, and foundation pit support have been solved by using algorithms such as artificial neural network (ANN), backpropagation neural network (BPNN), and genetic algorithm (GA).

[0004] However, previous studies have mainly focused on the static inversion of seepage field parameters, and the inversion problem of hydrothermal parameters of soil under hydrothermal coupling has not been deeply explored. In addition, the time effect of soil hydrothermal parameters caused by complex climate change has not been considered. Therefore, it is urgent to establish a method for inverting hydrothermal dynamic parameters of frozen-thawed soil in order to accurately describe the time-evolution characteristics of hydrothermal parameters of frozen-thawed soil. Summary of the Invention

[0005] The present invention is provided to solve the above problems existing in the prior art. Therefore, there is a need for a method, device and medium for inverting the hydrothermal dynamic parameters of frozen-thawed soil to overcome the bottleneck problems that the hydrothermal parameters of frozen-thawed soil cannot be predicted and adjusted in real time at present and the time effect of the hydrothermal parameters of frozen-thawed soil cannot be considered. The present invention can accurately reflect the evolution characteristics of soil hydrothermal properties over time, thereby solving the problem of identifying the hydrothermal parameters of frozen-thawed soil under the condition of real-time change of the external environment.

[0006] According to the first aspect of the present invention, there is provided a method for inverting the hydrothermal dynamic parameters of frozen-thawed soil, the method comprising:

[0007] Obtaining the monitoring information of the ground temperature and water content of the frozen-thawed soil at different depths during the change process of the external environment;

[0008] Determining the hydrothermal coupling model of the frozen-thawed soil and the hydrothermal parameters to be inverted;

[0009] Establishing a one-dimensional numerical calculation model of the frozen-thawed soil layer according to the geological condition characteristics of the frozen-thawed soil and the hydrothermal coupling model of the frozen-thawed soil;

[0010] Determining the value range of each hydrothermal parameter according to the geological characteristics, and obtaining a combination scheme of hydrothermal parameters by using orthogonal design; taking the obtained different combinations of hydrothermal parameters as the input parameters of the one-dimensional numerical calculation model of the frozen-thawed soil layer, performing forward calculation, obtaining the ground temperature and volumetric water content at different depths and the corresponding combinations of hydrothermal parameters as a sample set, and dividing the sample set into training samples and test samples;

[0011] Introducing a weight distribution coefficient to establish an objective function for the error between the measured values and predicted values of the ground temperature and volumetric water content at different depths;

[0012] Establishing the relationship between the input inverted hydrothermal parameters and the hydrothermal response of the frozen-thawed soil by using a backpropagation neural network;

[0013] Optimizing by using a genetic algorithm to obtain the global optimal solution of the objective function;

[0014] Selecting multiple characteristic time points at a set time interval. For each characteristic time point, based on the global optimal solution of the objective function at each characteristic time point, obtaining the evolution characteristics of the hydrothermal parameters of the frozen-thawed soil over time, and determining the inverted values of the hydrothermal parameters;

[0015] Substituting the inverted values of the hydrothermal parameters into the one-dimensional numerical calculation model of the frozen-thawed soil layer to obtain the predicted values of the ground temperature and volumetric water content; comparing the predicted values of the ground temperature and volumetric water content with the actual monitoring values to test the accuracy of the inversion results.

[0016] Furthermore, the freeze-thaw soil hydrothermal coupling model includes a moisture field equation, a temperature field equation, and a supplementary equation. The moisture field equation is expressed as:

[0017]

[0018] In the formula, ρ i 、ρ w 、ρ v represent the densities of solid ice, liquid water, and water vapor (kg / m 3 ); θ l 、θ i represent the volumetric water content and the volumetric ice content (m 3 / m 3 ); q l 、q v represent the liquid water flux and the water vapor flux (m / s); n represents the porosity of the freeze-thaw soil; t represents time (s); y represents the coordinate in the y-axis direction (m); h represents the hydraulic head (m); k lh0 (m / s) and k lT0 (m 2 / (K·s)) represent the unit hydraulic conductivity of liquid water under the water potential gradient and the unit hydraulic conductivity of liquid water under the temperature gradient; k vh0 (m / s) and k vT0 (m 2 / (K·s)) represent the unit water vapor migration coefficient under the action of the water potential gradient and the unit water vapor migration coefficient under the action of the temperature gradient; f1, f2, f3, f4 represent the characteristic functions of the hydraulic conductivity of liquid water and water vapor varying with time (s), where the hydraulic conductivity of liquid water k lh under the water potential gradient is expressed by the following formula:

[0019]

[0020] In the formula, k s represents the hydraulic conductivity of liquid water under the water potential gradient; α (m -1 ) and m (unitless) represent the empirical coefficients of the SWCC curve; θ s represents the saturated volumetric water content of the freeze-thaw soil (m 3 / m 3 ); θ r represents the residual volumetric water content of the freeze-thaw soil (m 3 / m 3 );

[0021] The temperature field equation is expressed as:

[0022]

[0023] In the formula, C l and Cv represents the volumetric heat capacity of liquid water and water vapor (J / (m 3 ·K)); C0 represents the volumetric heat capacity of the frozen-thawed soil mass (J / (m 3 ·K)); θ v represents the volumetric content of water vapor (m 3 / m 3 ); L f represents the latent heat of ice-water phase change (J / kg); L w represents the latent heat of evaporation of liquid water (J / kg); T represents the ground temperature (K); λ0 represents the unit thermal conductivity of the frozen-thawed soil mass (W / (m·K)); f5, f6 represent the characteristic functions of the volumetric heat capacity and thermal conductivity of the frozen-thawed soil mass varying with time (s);

[0024] The supplementary equation is expressed as:

[0025]

[0026] where θ0 represents the initial water content of the frozen-thawed soil mass (m 3 / m 3 ); T f represents the freezing point of the frozen-thawed soil mass (K); B represents the empirical coefficient.

[0027] Furthermore, the hydrothermal parameters to be inverted include the parameters to be inverted in the moisture field and the parameters to be inverted in the temperature field;

[0028] The parameters to be inverted in the moisture field include the saturated hydraulic conductivity k of liquid water under the water potential gradient s , the hydraulic conductivity k of liquid water under the temperature gradient lT , the vapor migration coefficient k under the action of the water potential gradient vh , the vapor migration coefficient k under the action of the temperature gradient vT , and the empirical coefficients α and m of the SWCC curve;

[0029] The parameters to be inverted in the temperature field include the volumetric heat capacity C of the soil and the thermal conductivity λ of the frozen-thawed soil mass.

[0030] Furthermore, according to the geological condition characteristics of the frozen-thawed soil mass and the hydrothermal coupling model of the frozen-thawed soil mass, a one-dimensional numerical calculation model of the frozen-thawed soil layer is established using the partial differential equation module. The soil layer depth of the one-dimensional numerical calculation model of the frozen-thawed soil layer is 10 m, the model uses a spatial discretization of 1 cm, and the maximum time step is 24 h.

[0031] Furthermore, a weight distribution coefficient is introduced, and the objective function representing the error between the measured values and predicted values of the ground temperature and volumetric water content at different depths is established as:

[0032]

[0033] In the formula, f represents the objective function; θ i (P) represents the time series data of the simulated volumetric water content at depth i; The time series data of the measured volumetric water content at depth i; T i (P) The time series data of the simulated ground temperature at depth i; The time series data of the measured ground temperature at depth i; w is the weight coefficient; P is the hydrothermal parameter vector to be inverted, expressed as:

[0034] P = [k s , m, α, θ s , k lT , k vh , k vT , λ, C] T (6).

[0035] Furthermore, establishing the relationship between the input inverted hydrothermal parameters and the output hydrothermal response of the frozen-thawed soil body by using a backpropagation neural network specifically includes:

[0036] The backpropagation neural network includes 1 input layer, 2 hidden layers and 1 output layer, and the input layer, hidden layer and output layer all include multiple neurons;

[0037] The backpropagation neural network is trained by using a backpropagation learning algorithm and regularization, and the Sigmoid function is used as the transfer function, and the expression is:

[0038]

[0039] In the formula, sig represents the Sigmoid function, and x represents the network input parameter.

[0040] Furthermore, optimizing by using a genetic algorithm to obtain the global optimal solution of the objective function specifically includes:

[0041] Generate a random hydrothermal parameter population, iteratively call selection, crossover and mutation, set the initial population size to 120, the crossover probability to 0.9, and the mutation probability to 0.1. When running 20 times or more, the optimal solution is obtained.

[0042] Furthermore, selecting multiple characteristic time points at a set time interval specifically includes:

[0043] Select one characteristic time point every 5 days, and a total of 26 characteristic time points are obtained.

[0044] According to the second technical solution of the present invention, a device for inverting the hydrothermal dynamic parameters of a frozen-thawed soil body is provided, and the device includes:

[0045] An acquisition module, configured to acquire monitoring information of the ground temperature and water content of frozen-thawed soil at different depths during the change process of the external environment;

[0046] A determination module, configured to determine a hydrothermal coupling model of frozen-thawed soil and hydrothermal parameters to be inverted;

[0047] A model establishment module, configured to establish a one-dimensional numerical calculation model of the frozen-thawed soil layer according to the geological condition characteristics of the frozen-thawed soil and the hydrothermal coupling model of the frozen-thawed soil;

[0048] A forward calculation module, configured to determine the value range of each hydrothermal parameter according to geological characteristics, and obtain a hydrothermal parameter combination scheme by using orthogonal design; use the obtained different hydrothermal parameter combinations as input parameters of the one-dimensional numerical calculation model of the frozen-thawed soil layer, perform forward calculation, obtain the ground temperature and volumetric water content at different depths and the corresponding hydrothermal parameter combinations as a sample set, and divide the sample set into training samples and test samples;

[0049] A function establishment module, configured to introduce a weight distribution coefficient and establish an objective function for the error between the measured values and predicted values of the ground temperature and volumetric water content at different depths;

[0050] A relationship establishment module, configured to establish a relationship between the input inverted hydrothermal parameters and the hydrothermal response of the frozen-thawed soil by using a backpropagation neural network;

[0051] A function optimization module, configured to perform optimization by using a genetic algorithm to obtain the global optimal solution of the objective function;

[0052] An inversion value determination module, configured to select multiple characteristic time points at a set time interval. For each characteristic time point, based on the global optimal solution of the objective function at each characteristic time point, obtain the time evolution characteristics of the hydrothermal parameters of the frozen-thawed soil and determine the inversion values of the hydrothermal parameters;

[0053] An inversion parameter verification module, configured to substitute the inversion values of the hydrothermal parameters into the one-dimensional numerical calculation model of the frozen-thawed soil layer to obtain predicted values of the ground temperature and volumetric water content; compare the predicted values of the ground temperature and volumetric water content with the actual monitoring values to check the accuracy of the inversion results.

[0054] According to the third technical solution of the present invention, a readable storage medium is provided. The readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the method as described above.

[0055] The present invention has at least the following beneficial effects:

[0056] (1) It can accurately describe the evolution characteristics of hydrothermal parameters of frozen-thawed soil with time;

[0057] (2) It can simultaneously achieve the accurate determination of multi-objective parameters of hydraulic parameters and thermal parameters;

[0058] (3) It can reflect the influence of external climate change on hydrothermal parameters;

[0059] (4) It can improve the prediction accuracy of the hydrothermal behavior of frozen-thawed soil. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It shows a schematic diagram of a backpropagation neural network according to an embodiment of the present invention;

[0061] Figure 2 It shows a flowchart of a method for inverting hydrothermal dynamic parameters of frozen-thawed soil according to an embodiment of the present invention;

[0062] Figure 3 It shows a schematic diagram of changes in the external environment during the monitoring period according to an embodiment of the present invention;

[0063] Figure 4 It shows the inverted values of hydrothermal parameters of frozen-thawed soil according to an embodiment of the present invention;

[0064] Figure 5 It shows a comparison chart of the simulated values and measured values of hydrothermal parameters of frozen-thawed soil according to an embodiment of the present invention;

[0065] Figure 6 It shows a structural diagram of a device for inverting hydrothermal dynamic parameters of frozen-thawed soil according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0066] To enable those skilled in the art to better understand the technical solutions of the present invention, the present invention will be described in detail below in conjunction with the drawings and specific embodiments. The embodiments of the present invention will be further described in detail below in conjunction with the drawings and specific examples, but shall not be construed as a limitation to the present invention. For the steps described herein, if there is no necessity for a front-back relationship between them, the order in which they are described as examples herein should not be regarded as a limitation. Those skilled in the art should know that they can be adjusted in order as long as the logic between them is not destroyed and the entire process cannot be realized.

[0067] In this embodiment, a certain monitoring area on the Qinghai-Tibet Plateau is taken as an example. The geological characteristics of the monitoring area and the characteristics of the established monitoring stations are as follows: The monitoring area belongs to a seasonal frozen soil area, and the soil texture is sandy soil. FDS-100 temperature and humidity sensors are buried at different depths to obtain the ground temperature and volumetric water content of the frozen-thawed soil. The monitoring depths are 0.15, 0.35, 0.70, 1.10, 1.55, and 1.85 m respectively. In this example, the freezing period is selected as the research period, from November 7, 2021 to March 5, 2022. Attached Figure 3 Describes the air temperature and daily rainfall changes during the monitoring period.

[0068] Such as Figure 2 As shown, this embodiment provides a method for inverting the hydrothermal dynamic parameters of frozen-thawed soil. This method is divided into a calculation part and an optimization part, and specifically includes the following steps 1 to 9.

[0069] Step 1: Obtain the monitoring information of the ground temperature and water content of the frozen-thawed soil at different depths during the change of the external environment.

[0070] It should be noted that the method for obtaining the monitoring information of the ground temperature and water content of the frozen-thawed soil at different depths during the change of the external environment can be to obtain it in real time by establishing an automated hydrothermal monitoring system for the frozen-thawed soil. Only as an example, the automated hydrothermal monitoring system for the frozen-thawed soil includes temperature and humidity integrated sensors at different depths, so as to obtain the monitoring information of the ground temperature and water content of the frozen-thawed soil at different depths during the change of the external environment.

[0071] Step 2: Determine the hydrothermal coupling model of the frozen-thawed soil and the hydrothermal parameters to be inverted.

[0072] For the moisture field, its formula is:

[0073]

[0074] In the formula, ρ i , ρ w , ρ v represent the density of solid ice, liquid water, and water vapor (kg / m 3 ); θ l , θ i represent the volumetric water content and volumetric ice content (m 3 / m 3 ); q l , q v represent the liquid water flux and water vapor flux (m / s); n represents the porosity of the frozen-thawed soil; t represents time (s); y represents the coordinate in the y-axis direction (m); h represents the hydraulic head (m); k lh0 (m / s) and k lT0 (m 2 / (K·s)) represents the unit hydraulic conductivity of liquid water under the water potential gradient and the unit hydraulic conductivity of liquid water under the temperature gradient (m 2 / (K·s)); k vh0 (m / s) and k vT0 (m 2 / (K·s)) represents the unit water vapor migration coefficient under the action of the water potential gradient and the unit water vapor migration coefficient under the action of the temperature gradient; f1, f2, f3, f4 represent the characteristic functions (s) of the hydraulic conductivity of liquid water and water vapor changing with time. Among them, the hydraulic conductivity k of liquid water under the water potential gradient lh can be expressed by the following formula:

[0075]

[0076] In the formula, k s represents the hydraulic conductivity of liquid water under the water potential gradient; α (m -1 ) and m represent the SWCC curve empirical coefficients; θ s represents the saturated volume water content of the frozen-thawed soil mass (m 3 / m 3 ); θ r represents the residual volume water content of the frozen-thawed soil mass (m 3 / m 3 ).

[0077] The temperature field equation is expressed as:

[0078]

[0079] In the formula, C l and C v represent the volume heat capacities of liquid water and water vapor (J / (m 3 ·K)); C0 represents the unit volume heat capacity of the frozen-thawed soil mass (J / (m 3 ·K)); θ v represents the volume content of water vapor (m 3 / m 3 ); L f represents the latent heat of ice-water phase change (J / kg); L w represents the latent heat of evaporation of liquid water (J / kg); T represents the ground temperature (K); λ0 represents the unit thermal conductivity of the frozen-thawed soil mass (W / (m·K)); f5, f6 represent the characteristic functions (s) of the volume heat capacity and thermal conductivity of the frozen-thawed soil mass changing with time.

[0080] The supplementary equation is:

[0081]

[0082] In the formula, θ0 represents the initial water content of the frozen-thawed soil mass (m 3 / m3 ); T f represents the freezing point (K) of the frozen-thawed soil; B represents the empirical coefficient.

[0083] Therefore, the parameters to be inverted in the moisture field include the saturated hydraulic conductivity k of liquid water under the water potential gradient s (m / s), the hydraulic conductivity k of liquid water under the temperature gradient lT (m 2 / (K·s)), the vapor migration coefficient k of water vapor under the action of the water potential gradient vh (m / s), the vapor migration coefficient k of water vapor under the action of the temperature gradient vT (m 2 / (K·s)), as well as the SWCC curve empirical coefficients α and m. The parameters to be inverted in the temperature field include the volumetric heat capacity C of the soil (J / (m 3 ·K)) and the thermal conductivity λ of the frozen-thawed soil (W / (m·K)).

[0084] Step 3: According to the geological condition characteristics of the frozen-thawed soil and the determined hydrothermal coupling model, use the partial differential equation (PDEs) module of COMSOL Multiphysics 5.6 to establish a one-dimensional numerical calculation model of a 10-m deep frozen-thawed soil layer.

[0085] Step 4: Determine the value range of each hydrothermal parameter according to the geological characteristics, and obtain a representative hydrothermal parameter combination scheme by using the orthogonal design in Table 1. Then, use the numerical model established in Step 2, with each parameter combination generated by the orthogonal design as the input parameters, to perform forward calculations. The numerical model uses a 1-cm spatial discretization, and the maximum time step is 24 h. The ground temperature and volumetric water content at different depths obtained and the corresponding hydrothermal parameter combinations are used as the sample set, and the sample set is divided into training samples and test samples. The ratio of training samples to test samples is 7:3.

[0086] Table 1 Settings of Orthogonal Design of Hydrothermal Parameters

[0087]

[0088]

[0089] Step 5: Introduce the weight distribution coefficient, and establish an objective function for the error between the measured values and predicted values of the ground temperature and volumetric water content at different depths. The formula is as follows:

[0090]

[0091] In the formula, f represents the objective function; θ i (P) represents the time series data of the simulated value of the volumetric water content at depth i; Time series data of the measured volumetric water content at depth i; T i (P) Time series data of the simulated ground temperature at depth i; Time series data of the measured ground temperature at depth i; w represents the weight coefficient, with a value of 0.06; P is the hydrothermal parameter vector to be inverted, which can be expressed as:

[0092] P = [k s , m, α, θ s , k lT , k vh , k vT , λ, C] T (6)

[0093] Step 6: Establish the relationship between the input inverted hydrothermal parameters and the output hydrothermal response (objective function) of the frozen-thawed soil mass using a backpropagation neural network (BPNN), as Figure 1 shown.

[0094] Step 7: Use the genetic algorithm (GA) for optimization to obtain the global optimal solution of the objective function.

[0095] Step 8: To obtain the time-evolution characteristics of the hydrothermal parameters of the frozen-thawed soil mass, select some typical characteristic time points at certain time intervals during the monitoring period. During the entire monitoring process, select a characteristic time point every 5 days. A total of 26 representative characteristic time points are obtained. For each characteristic time point, use orthogonal design, numerical simulation, BPNN, and GA to solve the established objective function, and the objective function of each characteristic time point is calculated according to the requirements of Steps 1 to 7. The inverted values of the hydrothermal parameters of the frozen-thawed soil mass are as Figure 4 shown.

[0096] Step 9: Compare the predicted results with the actual measured soil parameters to verify the accuracy of the inversion results. The results are as Figure 5 shown. It can be seen that the inversion results are accurate, conform to the characteristics of the on-site project, and obtain acceptable accurate hydrothermal parameter values of the frozen-thawed soil mass, and the inversion operation ends.

[0097] It should be noted that the method of comparing the predicted results with the actual measured soil parameters to verify the inversion results can be to calculate the difference between the predicted value and the measured value, and compare the calculated difference with the set threshold. When the difference is within the set threshold range, it indicates that the inversion result is accurate. If the calculated difference is not within the set threshold range, it is determined that the inversion result is deviated, and the inversion is performed again, or the parameters of the backpropagation neural network are adjusted and then the inversion is performed again until an accurate predicted value is obtained.

[0098] The embodiment of the present invention provides a device for inverting the hydrothermal dynamic parameters of a frozen-thawed soil mass, asFigure 6 As shown, the device 600 includes:

[0099] An acquisition module 601, configured to acquire monitoring information of the ground temperature and water content of frozen-thawed soil at different depths during the change process of the external environment;

[0100] A determination module 602, configured to determine a hydrothermal coupling model of the frozen-thawed soil and hydrothermal parameters to be inverted;

[0101] A model establishment module 603, configured to establish a one-dimensional numerical calculation model of the frozen-thawed soil layer according to the geological condition characteristics of the frozen-thawed soil and the hydrothermal coupling model of the frozen-thawed soil;

[0102] A forward calculation module 604, configured to determine the value range of each hydrothermal parameter according to the geological characteristics, and obtain a hydrothermal parameter combination scheme by using orthogonal design; use the obtained different hydrothermal parameter combinations as input parameters of the one-dimensional numerical calculation model of the frozen-thawed soil layer, perform forward calculation, obtain the ground temperature and volumetric water content at different depths and the corresponding hydrothermal parameter combinations as a sample set, and divide the sample set into training samples and test samples;

[0103] A function establishment module 605, configured to introduce a weight distribution coefficient and establish an objective function for the error between the measured values and predicted values of the ground temperature and volumetric water content at different depths;

[0104] A relationship establishment module 606, configured to establish a relationship between the input inverted hydrothermal parameters and the hydrothermal response of the frozen-thawed soil by using a backpropagation neural network;

[0105] A function optimization module 607, configured to perform optimization by using a genetic algorithm to obtain the global optimal solution of the objective function;

[0106] An inversion value determination module 608, configured to select multiple characteristic time points at a set time interval. For each characteristic time point, based on the global optimal solution of the objective function at each characteristic time point, obtain the evolution characteristics of the hydrothermal parameters of the frozen-thawed soil over time, and determine the inversion values of the hydrothermal parameters;

[0107] An inversion parameter verification module 609, configured to substitute the inversion values of the hydrothermal parameters into the one-dimensional numerical calculation model of the frozen-thawed soil layer to obtain predicted values of the ground temperature and volumetric water content; compare the predicted values of the ground temperature and volumetric water content with the actual monitoring values to test the accuracy of the inversion results.

[0108] In some embodiments, the hydrothermal coupling model of the frozen-thawed soil includes a moisture field equation, a temperature field equation, and a supplementary equation. The moisture field equation is expressed as:

[0109]

[0110] In the formula, ρ i , ρ w , ρ v represent the density of solid ice, liquid water, and water vapor (kg / m 3 ); θ l , θ i represent the volumetric water content and the volumetric ice content (m 3 / m 3 ); q l , q v represent the liquid water flux and the water vapor flux (m / s); n represents the porosity of the freeze-thaw soil; t represents time (s); y represents the coordinate in the y-axis direction (m); h represents the hydraulic head (m); k lh0 (m / s) and k lT0 (m 2 / (K·s)) represent the unit hydraulic conductivity of liquid water under the water potential gradient and the unit hydraulic conductivity of liquid water under the temperature gradient; k vh0 (m / s) and k vT0 (m 2 / (K·s)) represent the unit water vapor migration coefficient under the action of the water potential gradient and the unit water vapor migration coefficient under the action of the temperature gradient; f1, f2, f3, f4 represent the characteristic functions (s) of the hydraulic conductivity of liquid water and water vapor changing with time, where the hydraulic conductivity of liquid water k lh is expressed by the following formula:

[0111]

[0112] In the formula, k s represents the hydraulic conductivity of liquid water under the water potential gradient; α (m -1 ) and m (unitless) represent the empirical coefficients of the SWCC curve; θ s represents the saturated volumetric water content of the freeze-thaw soil (m 3 / m 3 ); θ r represents the residual volumetric water content of the freeze-thaw soil (m 3 / m 3 );

[0113] The temperature field equation is expressed as:

[0114]

[0115] In the formula, C l and C v represent the volumetric heat capacities of liquid water and water vapor (J / (m 3 ·K)); C0 represents the unit volume heat capacity of the freeze-thaw soil (J / (m 3 ·K)); θv represents the volume content of water vapor (m 3 / m 3 ); L f represents the latent heat of phase change of ice and water (J / kg); L w represents the latent heat of evaporation of liquid water (J / kg); T represents the ground temperature (K); λ0 represents the unit thermal conductivity of frozen-thawed soil (W / (m·K)); f5, f6 represent the characteristic functions (s) of the volumetric heat capacity and thermal conductivity of frozen-thawed soil changing with time;

[0116] The supplementary equation is expressed as:

[0117]

[0118] where θ0 represents the initial water content of frozen-thawed soil (m 3 / m 3 ); T f represents the freezing point (K) of frozen-thawed soil; B represents an empirical coefficient.

[0119] In some embodiments, the hydrothermal parameters to be inverted include the parameters to be inverted in the moisture field and the parameters to be inverted in the temperature field;

[0120] The parameters to be inverted in the moisture field include the saturated hydraulic conductivity k of liquid water under the water potential gradient s , the hydraulic conductivity k of liquid water under the temperature gradient lT , the water vapor migration coefficient k under the action of the water potential gradient vh , the water vapor migration coefficient k under the action of the temperature gradient vT , and the SWCC curve empirical coefficients α and m;

[0121] The parameters to be inverted in the temperature field include the volumetric heat capacity C of the soil and the thermal conductivity λ of the frozen-thawed soil.

[0122] In some embodiments, the model establishment module is further configured to establish a one-dimensional numerical calculation model of the frozen-thawed soil layer according to the geological condition characteristics of the frozen-thawed soil and the hydrothermal coupling model of the frozen-thawed soil. The soil layer depth of the one-dimensional numerical calculation model of the frozen-thawed soil layer is 10 m, the model adopts a spatial discretization of 1 cm, and the maximum time step is 24 h.

[0123] In some embodiments, the function establishment module is further configured to introduce a weight distribution coefficient and establish an objective function for the error between the measured and predicted values of the ground temperature and volumetric water content at different depths, which is expressed as:

[0124]

[0125] where f represents the objective function; θ i(P) represents the time series data of the simulated volumetric water content at depth i; The time series data of the measured volumetric water content at depth i; T i (P) The time series data of the simulated ground temperature at depth i; The time series data of the measured ground temperature at depth i; w is the weight coefficient; P is the hydrothermal parameter vector to be inverted, expressed as:

[0126] P = [k s , m, α, θ s , k lT , k vh , k vT , λ, C] T (6).

[0127] In some embodiments, the backpropagation neural network includes 1 input layer, 2 hidden layers and 1 output layer, and the input layer, hidden layers and output layer all include multiple neurons;

[0128] The backpropagation neural network is trained using the backpropagation learning algorithm and regularization, and the Sigmoid function is used as the transfer function, with the expression:

[0129]

[0130] In the formula, sig represents the Sigmoid function, and x represents the network input parameter.

[0131] In some embodiments, the function optimization module is further configured to:

[0132] Generate a random population of hydrothermal parameters, iteratively call selection, crossover and mutation, set the initial population size to 120, the crossover probability to 0.9, and the mutation probability to 0.1. When running 20 times or more, the optimal solution is obtained.

[0133] In some embodiments, the inversion value determination module is further configured to select a characteristic time point every 5 days, and a total of 26 characteristic time points are obtained.

[0134] It should be noted that the device described in this embodiment belongs to the same technical concept as the method described above, and can achieve the same technical effects, which will not be elaborated here.

[0135] The embodiment of the present invention provides a readable storage medium, and the readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the methods described in the above various embodiments.

[0136] The above description is intended to be illustrative and not restrictive. For example, the above examples (or one or more aspects thereof) may be used in combination with each other. For instance, those of ordinary skill in the art may use other embodiments when reading the above description. Additionally, in the above detailed description, various features may be grouped together to simplify the present invention. This should not be construed as an intention that the features of an unclaimed invention are necessary for any claim. On the contrary, the subject matter of the present invention may be less than all the features of a particular embodiment of the invention. Thus, the following claims are hereby incorporated into the detailed description by way of example or embodiment, where each claim stands on its own as a separate embodiment, and it is contemplated that these embodiments may be combined with each other in various combinations or permutations. The scope of the present invention should be determined with reference to the appended claims and the full scope of equivalents to which those claims are entitled.

Claims

1. A method for inversion of hydrothermal dynamic parameters of frozen-thaw soil, characterized in that: The method comprises: Obtain monitoring information on ground temperature and water content of frozen-thawed soil at different depths during changes in the external environment; Determine the hydrothermal coupling model of frozen-thaw soil and the hydrothermal parameters to be inverted; According to the geological conditions of the frozen-thaw soil body and the water-heat coupling model of the frozen-thaw soil body, a one-dimensional numerical calculation model of the frozen-thaw soil layer is established; Determine the value range of each hydrothermal parameter according to geological characteristics, and use orthogonal design to obtain a hydrothermal parameter combination scheme; use the obtained different hydrothermal parameter combinations as input parameters of the one-dimensional numerical calculation model of the frozen-thaw soil layer, perform forward calculations, obtain ground temperatures and volumetric water contents at different depths and corresponding hydrothermal parameter combinations as sample sets, and divide the sample sets into training samples and test samples; The weight distribution coefficient is introduced to establish the objective function of the error between the measured and predicted values ​​of ground temperature and volume water content at different depths; Back propagation neural network is used to establish the relationship between the input inversion hydrothermal parameters and the output hydrothermal response of frozen-thaw soil. Genetic algorithm is used for optimization to obtain the global optimal solution of the objective function; Select multiple characteristic time points according to the set time interval, and for each characteristic time point, obtain the time evolution characteristics of the hydrothermal parameters of the frozen-thawed soil based on the global optimal solution of the objective function at each characteristic time point, and determine the inversion value of the hydrothermal parameters; The inverted values ​​of the hydrothermal parameters are substituted into the one-dimensional numerical calculation model of the frozen-thaw soil layer to obtain the predicted values ​​of ground temperature and volumetric water content; the predicted values ​​of ground temperature and volumetric water content are compared with the actual monitored values ​​to verify the accuracy of the inversion results.

2. The method according to claim 1, characterized in that The freeze-thaw soil water-heat coupling model includes a moisture field equation, a temperature field equation and a supplementary equation. The moisture field equation is expressed as: In the formula, ρ i , w , v Indicates the density of solid ice, liquid water, and water vapor (kg / m 3 );θ l ,θ i Indicates volumetric water content and volumetric ice content (m 3 / m 3 );q l ,q v represents liquid water flux and water vapor flux (m / s); n represents the porosity of the frozen-thawed soil; t represents time (s); y represents the coordinate in the y direction of the coordinate axis (m); h represents the water head (m); k lh0 (m / s) and k lT0 (m 2 / (K·s)) represents the unit hydraulic conductivity of liquid water under water potential gradient and the unit hydraulic conductivity of liquid water under temperature gradient; k vh0 (m / s) and k vT0 (m 2 / (K·s)) represents the unit water vapor migration coefficient under the water potential gradient and the unit water vapor migration coefficient under the temperature gradient; f1, f2, f3, f4 represent the characteristic functions of the water conductivity of liquid water and water vapor over time (s), where the liquid water conductivity k under the water potential gradient is lh It is expressed by the following formula: In the formula, k s represents the hydraulic conductivity of liquid water under water potential gradient; α(m -1 ) and m(unitless) represent the empirical coefficients of the SWCC curve; θ s represents the saturated volumetric water content of frozen-thawed soil (m 3 / m 3 );θ r Represents the residual volume water content of frozen-thawed soil (m 3 / m 3 ); The temperature field equation is expressed as: In the formula, C l and C v The volume heat capacity of liquid water and water vapor (J / (m 3 ·K)); C0 represents the heat capacity per unit volume of frozen-thawed soil (J / (m 3 ·K));θ v Indicates the volume content of water vapor (m 3 / m 3 );L f represents the latent heat of ice-water phase change (J / kg); L w represents the latent heat of evaporation of liquid water (J / kg); T represents the ground temperature (K); λ0 represents the unit thermal conductivity of the frozen-thawed soil (W / (m·K)); f5, f6 represent the characteristic functions of the volume heat capacity and thermal conductivity of the frozen-thawed soil over time (s); The supplementary equation is expressed as: Where θ0 represents the initial water content of the frozen-thawed soil (m 3 / m 3 );T f represents the freezing point of frozen-thawed soil (K); B represents the empirical coefficient.

3. The method according to claim 2, characterized in that The hydrothermal parameters to be inverted include parameters to be inverted in the moisture field and parameters to be inverted in the temperature field; The parameters to be inverted in the water field include the saturated hydraulic conductivity k of liquid water under the water potential gradient s , Liquid water conductivity k under temperature gradient lT , water vapor transport coefficient k under the action of water potential gradient vh , water vapor transport coefficient k under temperature gradient vT And the SWCC curve empirical coefficients α and m; The parameters to be inverted in the temperature field include the volume heat capacity C of the soil and the thermal conductivity λ of the frozen-thawed soil.

4. The method according to claim 1, characterized in that: According to the geological conditions of the frozen-thaw soil and the hydrothermal coupling model of the frozen-thaw soil, a one-dimensional numerical calculation model of the frozen-thaw soil layer is established using a partial differential equation module. The soil layer depth of the one-dimensional numerical calculation model of the frozen-thaw soil layer is 10m, the model adopts a spatial discretization of 1cm, and the maximum time step is 24h.

5. The method according to claim 2, characterized in that: Introducing the weight distribution coefficient, the objective function of the error between the measured and predicted values ​​of ground temperature and volume water content at different depths is established as follows: Where f represents the objective function; θ i (P) represents the time series data of the simulated value of volumetric water content at depth i; Time series data of measured volumetric water content at depth i; T i (P) Time series data of simulated ground temperature at depth i; The time series data of the measured ground temperature at depth i; w is the weight coefficient; P is the hydrothermal parameter vector to be inverted, expressed as: P=[k s ,m,a,i s ,k lT ,k vh ,k vT ,λ,C] T (6)。 6. The method according to claim 1, characterized in that The back propagation neural network is used to establish the relationship between the input inversion hydrothermal parameters and the output hydrothermal response of the frozen-thaw soil, specifically including: The back propagation neural network includes an input layer, two hidden layers and an output layer, and the input layer, the hidden layer and the output layer all include a plurality of neurons; The back propagation neural network is trained using a reverse learning algorithm and regularization, with the Sigmoid function as the transfer function, expressed as: In the formula, sig represents the Sigmoid function, and x represents the network input parameter.

7. The method according to claim 1, characterized in that The optimization using genetic algorithm to obtain the global optimal solution of the objective function specifically includes: A random population of hydrothermal parameters was generated, and selection, crossover, and mutation were iteratively called. The initial population size was set to 120, the crossover probability was 0.9, and the mutation probability was 0.

1. The optimal solution was obtained when running 20 times or more.

8. The method according to claim 1, characterized in that The selecting of multiple characteristic time points according to the set time interval specifically includes: A characteristic time point is selected every 5 days, and a total of 26 characteristic time points are obtained.

9. A method and device for inversion of hydrothermal dynamic parameters of frozen-thaw soil, characterized in that: The device comprises: An acquisition module is configured to acquire monitoring information of ground temperature and water content of frozen-thawed soil at different depths during changes in the external environment; A determination module configured to determine a hydrothermal coupling model of frozen-thaw soil and hydrothermal parameters to be inverted; A model building module is configured to build a one-dimensional numerical calculation model of the frozen-thaw soil layer according to the geological conditions of the frozen-thaw soil body and the hydrothermal coupling model of the frozen-thaw soil body; The forward calculation module is configured to determine the value range of each hydrothermal parameter according to geological characteristics, and obtain a hydrothermal parameter combination scheme by orthogonal design; the obtained different hydrothermal parameter combinations are used as input parameters of the one-dimensional numerical calculation model of the frozen-thaw soil layer, and forward calculation is performed to obtain ground temperatures and volumetric water contents at different depths and corresponding hydrothermal parameter combinations as sample sets, and the sample sets are divided into training samples and test samples; A function building module is configured to introduce a weight distribution coefficient and build an objective function of the error between the measured value and the predicted value of the ground temperature and volume water content at different depths; A relationship building module is configured to use a back propagation neural network to establish a relationship between an input inversion hydrothermal parameter and an output freeze-thaw soil hydrothermal response; A function optimization module is configured to optimize using a genetic algorithm to obtain a global optimal solution of the objective function; The inversion value determination module is configured to select a plurality of characteristic time points at a set time interval, and for each characteristic time point, obtain the time evolution characteristics of the hydrothermal parameters of the frozen-thawed soil body based on the global optimal solution of the objective function at each characteristic time point, and determine the inversion value of the hydrothermal parameters; The inversion parameter verification module is configured to substitute the inversion value of the hydrothermal parameter into the one-dimensional numerical calculation model of the frozen-thaw soil layer to obtain the predicted values ​​of ground temperature and volumetric water content; compare the predicted values ​​of ground temperature and volumetric water content with the actual monitored values ​​to verify the accuracy of the inversion result.

10. A readable storage medium, characterized in that: The readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the method according to any one of claims 1 to 8.

Images