Geothermal field parameter inversion calculation method, device and system and storage medium

Abstract

The invention discloses a geothermal field parameter inversion calculation method, device and system, and a storage medium. The method comprises the following steps: establishing a geophysical stratified model; through Monte Carlo sampling of preset parameter spaces of the crustal heat generation rate and the heat conductivity, a temperature value is calculated in combination with a heat conduction equation, and a Gaussian likelihood function evaluation model is constructed to predict the matching degree of the temperature and the actually measured temperature; obtaining the optimal estimation of the thermal parameters, a confidence interval and a correlation matrix among the parameters based on the posterior probability distribution; and according to the correlation matrix, taking a high-confidence result generated by Monte Carlo inversion as priori knowledge of a physical guidance neural network PINNs, and finally outputting a thermal parameter spatial distribution prediction result of the target area through parameter initialization constraint, output layer range limitation and a physical regularization loss item. By adopting the technical scheme of the invention, the defects that the existing actually measured geothermal field parameters are rare and the regional characteristics cannot be described and the geophysical joint inversion of the geothermal field parameters cannot be realized in the prior art are overcome.

Description

Technical Field

[0001] The present invention belongs to the technical field of geothermal resource exploration, and in particular relates to a geothermal field parameter inversion calculation method and device, system, and storage medium. Background Art

[0002] As an important clean and renewable energy source, geothermal resources play a crucial role in the transformation of the energy structure. The outward transfer of heat from the Earth's interior is the primary source of geothermal resources, and therefore the thermal structure of the Earth's interior is a key factor influencing the distribution of geothermal resources. However, the Earth's internal geological structure is complex and varied, and traditional single geophysical inversion methods have significant limitations in interpreting the thermal structure. Therefore, accurately obtaining geothermal field parameters such as crustal heat generation rate, thermal conductivity, and temperature distribution is a key step in assessing the potential for geothermal resource development and has important geological significance for geothermal resource exploration.

[0003] Commonly used gravity exploration can obtain information on underground density distribution, but it is difficult to directly link it to heat conduction characteristics; magnetic exploration focuses on the analysis of underground magnetic structure and has a weak intrinsic connection with geothermal parameters; electromagnetic exploration is susceptible to electromagnetic interference and lacks accuracy in the inversion of deep geothermal parameters. Existing methods all have defects in quantifying parameter uncertainty, resulting in poor reliability of geothermal resource assessment results. In recent years, although multi-geophysical joint inversion and data-driven inversion technologies have developed, how to efficiently integrate multi-dimensional geophysical information under complex geological conditions, fully consider physical laws and parameter uncertainties in the inversion process, and the problem of low computational efficiency are still technical difficulties that need to be overcome in the field of geothermal exploration. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a geothermal field parameter inversion calculation method and device, system, and storage medium to overcome the deficiency of the existing technology that it cannot use geophysical joint inversion of thermal structure parameters.

[0005] To achieve the above object, the present invention adopts the following technical solutions: A geothermal field parameter inversion calculation method, comprising: Step S1, obtaining a topographic map, basement depth map, Moho depth map, magnetic basement depth map, and Curie depth map of the target area, and establishing a geophysical layered model; Step S2: Based on the geophysical layered model, Monte Carlo sampling of the preset parameter space of crustal heat generation rate and thermal conductivity is performed, and the temperature value is calculated in combination with the heat conduction equation. A Gaussian likelihood function is constructed to evaluate the degree of match between the model-predicted temperature and the measured temperature; based on the posterior probability distribution, the optimal estimate of the thermal parameters, the confidence interval, and the correlation matrix between the parameters are obtained; Step S3: Based on the correlation matrix, the high-confidence results generated by Monte Carlo inversion are used as the prior knowledge of the physical guided neural network (PINNs). Through parameter initialization constraints, output layer range restrictions, and physical regularization loss terms, the prediction results of the thermal parameter spatial distribution of the target area are finally output.

[0006] Preferably, in step S2, the heat conduction equation calculation and the measured temperature data are combined to construct a Gaussian likelihood function to evaluate the matching degree between the predicted temperature and the measured temperature value of the borehole, that is: ; in, is the measured temperature vector of the borehole, are the geophysical layered model parameters, is the number of data points, is the data covariance matrix , is the temperature residual vector.

[0007] Preferably, the total loss function of the neural network includes: a data fitting loss term, which is used to measure the difference between the PINNs predicted temperature and the Monte Carlo high-likelihood sample temperature; a physical regularization loss term, which is used to make the predicted parameters satisfy the residual constraints of the heat conduction equation; and a geological prior regularization loss term, which is used to force the correlation of the predicted parameters to match the Monte Carlo results.

[0008] The present invention also provides a geothermal field parameter inversion calculation device, comprising: The first processing module is used to obtain a topographic map, a basement depth map, a Moho depth map, a magnetic basement depth map, and a Curie depth map of the target area and establish a geophysical layered model; The second processing module is used to calculate temperature values ​​based on the geophysical layered model through Monte Carlo sampling of the preset parameter space of crustal heat generation rate and thermal conductivity, combined with the heat conduction equation, and construct a Gaussian likelihood function to evaluate the degree of match between the model-predicted temperature and the measured temperature. The module also obtains the optimal estimate of the thermal parameters, confidence intervals, and correlation matrix between the parameters based on the posterior probability distribution. The third processing module is used to use the high-confidence results generated by Monte Carlo inversion as the prior knowledge of the physical guided neural network (PINNs) based on the correlation matrix. Through parameter initialization constraints, output layer range restrictions and physical regularization loss terms, it finally outputs the prediction results of the spatial distribution of thermal parameters in the target area.

[0009] Preferably, the second processing module combines the heat conduction equation calculation with the measured temperature data to construct a Gaussian likelihood function to evaluate the matching degree between the predicted temperature and the measured temperature value of the borehole, namely: ; in, is the measured temperature vector of the borehole, are the geophysical layered model parameters, is the number of data points, is the data covariance matrix , is the temperature residual vector.

[0010] Preferably, the total loss function of the neural network includes: a data fitting loss term, which is used to measure the difference between the PINNs predicted temperature and the Monte Carlo high-likelihood sample temperature; a physical regularization loss term, which is used to make the predicted parameters satisfy the residual constraints of the heat conduction equation; and a geological prior regularization loss term, which is used to force the correlation of the predicted parameters to match the Monte Carlo results.

[0011] The present invention also provides a geothermal field parameter inversion calculation system, comprising: a memory and a processor, wherein the memory stores a computer program run by the processor, and the computer program executes a geothermal field parameter inversion calculation method when run by the processor.

[0012] The present invention also provides a storage medium, on which a computer program is stored, and the computer program executes the geothermal field parameter inversion calculation method when running.

[0013] The present invention obtains basement depth maps, Moho surface depth maps, magnetic basement depth maps and Curie surface depth maps of the target area, and establishes a geophysical layered model with clear thermal boundaries; through Monte Carlo sampling of the parameter space of crustal heat generation rate, thermal conductivity and other parameters, combined with the forward calculation of temperature by heat conduction equation, a Gaussian likelihood function is constructed to evaluate the degree of match between the model predicted temperature and the measured temperature considering the spatial correlation of the parameters; based on the posterior probability distribution, the optimal estimate, confidence interval and correlation matrix of the temperature and thermal parameters are obtained; the results generated by Monte Carlo inversion are used as prior knowledge for physical guided neural networks (PINNs), and through parameter initialization constraints, output layer range restrictions and physical regularization loss terms, the final prediction results of the spatial distribution of geothermal parameters in the target area with probabilistic significance are output. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0015] Figure 1 This is a flow chart of the geothermal field parameter inversion calculation method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] Example 1: like Figure 1 As shown, an embodiment of the present invention provides a geothermal field parameter inversion calculation method, including: Step S1, obtaining a topographic map, basement depth map, Moho depth map, magnetic basement depth map, and Curie depth map of the target area, and establishing a geophysical layered model; Step S2: Based on the geophysical layered model, Monte Carlo sampling of the preset parameter space of crustal heat generation rate and thermal conductivity is performed, and the temperature value is calculated in combination with the heat conduction equation. A Gaussian likelihood function is constructed to evaluate the degree of match between the model-predicted temperature and the measured temperature; based on the posterior probability distribution, the optimal estimate of the thermal parameters, the confidence interval, and the correlation matrix between the parameters are obtained; Step S3: Based on the correlation matrix, the high-confidence results generated by Monte Carlo inversion are used as the prior knowledge of the physical guided neural network (PINNs). Through parameter initialization constraints, output layer range restrictions, and physical regularization loss terms, the prediction results of the thermal parameter spatial distribution of the target area are finally output.

[0019] As an implementation method of the embodiment of the present invention, in step S2, Monte Carlo random sampling is performed on the crustal heat generation rate H and thermal conductivity k: (Lognormal distribution, mean , standard deviation Determined by geological experience); (Uniformly distributed, the value range is determined by the rock type).

[0020] The finite difference method is used to solve the steady-state one-dimensional heat conduction equation: ; Calculate the depth point z j Measured temperature Ti (z j ) and the calculated predicted temperature Tobs(z j ) Residuals: ; Assuming that the error follows a Gaussian distribution, a likelihood function is constructed to evaluate the matching degree between the predicted temperature and the measured temperature value of the borehole. The function expression is: ; in: is the borehole measured temperature vector, are the geophysical layered model parameters (rock thermal conductivity k, basement heat flow Qk, rock heat production rate A), is the number of data points, is the data covariance matrix , describing the uncertainty of temperature observations and the correlation between data points, is the temperature residual vector.

[0021] Calculate the posterior probability distribution and obtain the optimal estimate of the thermal parameters, confidence intervals, and correlation matrix between parameters based on the posterior probability distribution. The posterior distribution of the thermal parameters based on Bayes' theorem is: ; in, is the prior distribution of thermal parameters, is the normalization constant.

[0022] Sampling the posterior distribution via Markov Chain Monte Carlo (MCMC) , Posterior mean estimation (minimum mean square error estimation): ; Maximum a Posteriori Estimation (MAP): ; Parameters (such as base heat flow )of Confidence interval, calculate the posterior sample Quantile of: ; Calculate the covariance matrix and correlation matrix based on the posterior samples: ; in, Indicates that the parameters are positively correlated. Indicates that the parameters are negatively correlated.

[0023] The training set is constructed by selecting high-likelihood samples (confidence level > 90%) from the posterior distribution of the Monte Carlo inversion. Parameter information includes spatial location and thermal parameters (such as thermal conductivity and heat production rate).

[0024] ; in, is the spatial coordinate of the sample point (x i 、y i 、z i ), are the layered model parameters (such as stratum number, lithology code), are the MC inversion thermal parameters (such as thermal conductivity κi and heat production rate Ai).

[0025] Construct an MLP multi-layer neural network.

[0026] ; in, for , that is, the input dimension (coordinates + model parameters), for , that is, the output dimension (temperature + thermal parameters).

[0027] Using the high likelihood value samples of the Monte Carlo posterior distribution from the high likelihood sample set Select samples from the dataset and initialize the network parameters by minimizing the difference between the initial prediction and the high likelihood sample: ; in, is a high likelihood sample set, is the mean of the posterior distribution.

[0028] The output is constrained according to the 95% confidence interval of the Monte Carlo inversion.

[0029] ; Construct a loss function. The total loss function consists of a data fitting loss term, which measures the difference between the PINNs predicted temperature and the Monte Carlo high-likelihood sample temperature; a physical regularization loss term, which ensures that the predicted parameters satisfy the residual constraints of the heat conduction equation; and a geological prior regularization loss term, which forces the correlation of the predicted parameters to match the Monte Carlo results.

[0030] ; in, is the data fitting loss weight, is the data fitting loss term, is the physical regularization loss weight, is the physical regularization loss term, is the geological prior loss weight, is the geological prior loss term.

[0031] in, Data fitting loss term: ; Where N is the total number of sampling points, TPINN is the ground temperature field distribution predicted by the neural network PINN, is the temperature field calculated by extracting the i-th high likelihood value sample from the Monte Carlo posterior distribution.

[0032] in, Physical regularization loss term: ; in, is the number of configuration points, 、 To calculate the thermophysical parameters of each j-th configuration point based on the PINNs prediction results, Predicted temperature field gradients for PINNs.

[0033] The geological prior loss term is: ; in, The KL divergence between the PINNs predicted distribution and the Monte Carlo prior distribution is used to measure the difference in probability distribution; represents the PINNs predicted value of the kth thermophysical parameter; represents the prior mean of the kth parameter obtained by Monte Carlo inversion; represents the prior standard deviation of the kth parameter obtained by Monte Carlo inversion; and They represent the lower and upper bounds of the confidence interval of the kth parameter respectively; Represents a linear rectification function to implement unilateral out-of-bounds penalty; The total number of dimensions representing the thermophysical parameters.

[0034] The Adam optimization algorithm is used to train the neural network. After training, the geological model is input for prediction.

[0035] Use MC Dropout for multiple sampling to enable Dropout prediction, and perform K sampling predictions for each spatial position (x, y, z): ; in Indicates the The parameter vector obtained by subsampling.

[0036] Compute statistics for each grid cell: ;

[0037] Generate a spatial distribution grid map of geothermal parameters in the target area, including the mean, standard deviation, and confidence interval of temperature, thermal conductivity, and heat generation rate.

[0038] Example 2: The embodiment of the present invention further provides a geothermal field parameter inversion calculation device, comprising: The first processing module is used to obtain a topographic map, a basement depth map, a Moho depth map, a magnetic basement depth map, and a Curie depth map of the target area and establish a geophysical layered model; The second processing module is used to calculate temperature values ​​based on the geophysical layered model through Monte Carlo sampling of the preset parameter space of crustal heat generation rate and thermal conductivity, combined with the heat conduction equation, and construct a Gaussian likelihood function to evaluate the degree of match between the model-predicted temperature and the measured temperature. The module also obtains the optimal estimate of the thermal parameters, confidence intervals, and correlation matrix between the parameters based on the posterior probability distribution. The third processing module is used to use the high-confidence results generated by Monte Carlo inversion as the prior knowledge of the physical guided neural network (PINNs) based on the correlation matrix. Through parameter initialization constraints, output layer range restrictions and physical regularization loss terms, it finally outputs the prediction results of the spatial distribution of thermal parameters in the target area.

[0039] As an implementation method of an embodiment of the present invention, the second processing module combines the heat conduction equation calculation and the measured temperature data to construct a Gaussian likelihood function to evaluate the matching degree between the predicted temperature and the measured temperature value of the borehole, that is: ;

[0040] in, is the measured temperature vector of the borehole, are the geophysical layered model parameters, is the number of data points, is the data covariance matrix , is the temperature residual vector.

[0041] As an implementation method of an embodiment of the present invention, the total loss function of the neural network includes: a data fitting loss term, which is used to measure the difference between the PINNs predicted temperature and the Monte Carlo high-likelihood sample temperature; a physical regularization loss term, which is used to make the predicted parameters satisfy the residual constraints of the heat conduction equation; and a geological prior regularization loss term, which is used to force the correlation of the predicted parameters to match the Monte Carlo results.

[0042] Example 3: An embodiment of the present invention further provides a geothermal field parameter inversion calculation system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a geothermal field parameter inversion calculation method when executed by the processor.

[0043] Example 4: An embodiment of the present invention further provides a storage medium having a computer program stored thereon, and the computer program executes the geothermal field parameter inversion calculation method when running.

[0044] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. A geothermal field parameter inversion calculation method, characterized in that: include: Step S1, obtaining a topographic map, basement depth map, Moho depth map, magnetic basement depth map, and Curie depth map of the target area, and establishing a geophysical layered model; Step S2: Based on the geophysical layered model, Monte Carlo sampling of the preset parameter space of crustal heat generation rate and thermal conductivity is performed, and the temperature value is calculated in combination with the heat conduction equation. A Gaussian likelihood function is constructed to evaluate the degree of match between the model-predicted temperature and the measured temperature; based on the posterior probability distribution, the optimal estimate of the thermal parameters, the confidence interval, and the correlation matrix between the parameters are obtained; Step S3: Based on the correlation matrix, the high-confidence results generated by Monte Carlo inversion are used as the prior knowledge of the physical guided neural network (PINNs). Through parameter initialization constraints, output layer range restrictions, and physical regularization loss terms, the prediction results of the thermal parameter spatial distribution of the target area are finally output.

2. The geothermal field parameter inversion calculation method according to claim 1, characterized in that: In step S2, the heat conduction equation calculation and the measured temperature data are combined to construct a Gaussian likelihood function to evaluate the matching degree between the model predicted temperature and the measured borehole temperature value, namely: ； in, is the measured temperature vector of the borehole, are the geophysical layered model parameters, is the number of data points, is the data covariance matrix , is the temperature residual vector.

3. The geothermal field parameter inversion calculation method according to claim 2, characterized in that: The total loss function of the neural network includes: a data fitting loss term, which is used to measure the difference between the PINNs predicted temperature and the Monte Carlo high-likelihood sample temperature; a physical regularization loss term, which is used to make the predicted parameters satisfy the residual constraints of the heat conduction equation; and a geological prior regularization loss term, which is used to force the correlation of the predicted parameters to match the Monte Carlo results.

4. A geothermal field parameter inversion calculation device, characterized in that: include: The first processing module is used to obtain a topographic map, a basement depth map, a Moho depth map, a magnetic basement depth map, and a Curie depth map of the target area and establish a geophysical layered model; The second processing module is used to calculate temperature values ​​based on the geophysical layered model through Monte Carlo sampling of the preset parameter space of crustal heat generation rate and thermal conductivity, combined with the heat conduction equation, and construct a Gaussian likelihood function to evaluate the degree of match between the model-predicted temperature and the measured temperature. The module also obtains the optimal estimate of the thermal parameters, confidence intervals, and correlation matrix between the parameters based on the posterior probability distribution. The third processing module is used to use the high-confidence results generated by Monte Carlo inversion as the prior knowledge of the physical guided neural network (PINNs) based on the correlation matrix. Through parameter initialization constraints, output layer range restrictions and physical regularization loss terms, it finally outputs the prediction results of the spatial distribution of thermal parameters in the target area.

5. The geothermal field parameter inversion calculation device according to claim 4, characterized in that: The second processing module combines the heat conduction equation calculation with the measured temperature data to construct a Gaussian likelihood function to evaluate the matching degree between the predicted temperature and the measured temperature value of the borehole, namely: ； in, is the measured temperature vector of the borehole, are the geophysical layered model parameters, is the number of data points, is the data covariance matrix , is the temperature residual vector.

6. The geothermal field parameter inversion calculation device according to claim 5, characterized in that: The total loss function of the neural network includes: a data fitting loss term, which is used to measure the difference between the PINNs predicted temperature and the Monte Carlo high-likelihood sample temperature; a physical regularization loss term, which is used to make the predicted parameters satisfy the residual constraints of the heat conduction equation; and a geological prior regularization loss term, which is used to force the correlation of the predicted parameters to match the Monte Carlo results.

7. A geothermal field parameter inversion calculation system, characterized in that: include: A memory and a processor, wherein the memory stores a computer program executed by the processor, and when the computer program is executed by the processor, the geothermal field parameter inversion calculation method according to any one of claims 1 to 3 is executed.

8. A storage medium, characterized in that: The storage medium stores a computer program, which, when running, executes the geothermal field parameter inversion calculation method according to any one of claims 1 to 3.

Images