A multi-physics collaborative property inversion method for mining areas based on interactive guidance

Abstract

The present invention discloses a multi-physics field collaborative physical property inversion method for a mineral cluster area based on interactive guidance, which belongs to the field of mineral exploration technology. The technical key points are as follows: Step 1: Subdividing the underground space into grids according to the horizontal position distribution of its observation points; Step 2: Calculating the forward kernel matrix of each geophysical field to obtain the mapping relationship between the observation point and the underground discrete grid unit under the unit physical property, and establishing the observation data through the forward kernel matrix. d To Model m relationship; Step three: obtain the three-dimensional discrete grids corresponding to the observation points of different physical fields, and realize the forward and inverse calculation of the three-dimensional discrete grids of each physical field by interactively guiding the calculation process. This has the advantages of solving the problems in the existing technology of strong dependence on rock physical property measurement and statistical relationship, and the inability of structural coupling mode to adapt to the sensitivity differences of different physical property parameters, and realizing high-precision and high-resolution multi-physical field physical property modeling of mining areas.

Description

Technical Field

[0001] The present invention relates to the technical field of mineral exploration, and in particular to a multi-physical field collaborative physical property inversion method for a mineral cluster area based on interactive guidance. Background Art

[0002] Geophysical field inversion technology is an important means of clarifying the spatial structure of underground targets and the spatial distribution of material physical parameters, and plays a vital role in mineral exploration and resource evaluation. However, geophysical data inversion is inherently a mathematically ill-posed problem. Especially under the complex geological conditions of mineralized areas, the inversion results of a single physical field (such as gravity, geomagnetic field, and others) often suffer from significant interpretation ambiguity and multiple solutions. Because different physical fields have different response characteristics to underground targets: the gravity field is extremely sensitive to density, and the geomagnetic field reflects the distribution of magnetization intensity, a single data set cannot fully characterize the three-dimensional structure and physical properties of geological bodies. To improve the reliability of underground material detection, multi-physics collaborative inversion methods have emerged. Collaborative inversion combines data from multiple geophysical fields (such as gravity, geomagnetic field, and others) and utilizes their complementary information and coupled constraints to invert the distribution of underground material physical parameters (density, magnetization, and others). Existing collaborative inversion method technologies have mainly developed into two categories: physical property-based coupling constraints and structural coupling constraints. The former is based on the correlation between the physical property parameters of rocks and minerals corresponding to different geophysical sites, and the mutual collaborative constraints promote the consistent interpretation of underground targets.

[0003] However, this method requires a large number of rock and mineral sample sampling and statistics at different locations and depths in the interpretation area to provide a priori rock physical correlation. In practice, the inversion accuracy is limited by accurate prior information and the operation is complex. The latter flexibly assumes that the distribution structure of different physical property parameters of the same underground medium is the same or similar. It usually constrains the inversion model to be the same or similar by measuring the directional parallelism or intensity consistency of the changes in different physical parameters. However, there are differences in the sensitivity of different physical fields of underground targets. The commonly used cross-gradient collaborative inversion method cannot flexibly adapt to the different characteristics of different physical structures, and is prone to interlaced artifacts, making it difficult to accurately characterize the edge structure and physical property distribution of the geological body.

[0004] In response to the above bottlenecks, the present invention discloses a collaborative physical property inversion method for multi-physical fields in mining areas based on interactive guidance, and proposes a new type of nonlinear structural coupling constraint method. It dynamically guides the distribution of each physical property model in the collaborative inversion process of different physical parameters, and uses their respective information to guide each other to achieve joint reconstruction of different physical property parameter models, thereby effectively obtaining higher-resolution geological target edge features and more accurate underground physical parameter distribution. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the purpose of the embodiments of the present invention is to provide a multi-physical field collaborative physical property inversion method for a mining area based on interactive guidance to solve the problems in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A multi-physics field collaborative physical property inversion method for a mineral cluster area based on interactive guidance includes the following steps:

[0008] Step 1: Divide the underground space into grids according to the horizontal position distribution of the observation points;

[0009] Step 2: Calculate the forward kernel matrix of each geophysical field, obtain the mapping relationship between the observation point and the underground discrete grid unit under the unit physical property, and establish the observation data through the forward kernel matrix d To Model m relationship;

[0010] Step 3: Obtain the three-dimensional discrete grids corresponding to different physical field observation points, and implement the three-dimensional discrete grid forward and inverse calculations of each physical field through interactively guided calculation processes.

[0011] As a further solution of the present invention, the observation data in step 2 d To Model m The relationship is:

[0012] (1)

[0013] in, is the forward kernel matrix of each geophysical field, Represents different geophysical field observation data, are model parameters that respectively characterize the physical properties of the underground medium.

[0014] As a further solution of the present invention, the forward and inverse calculations in step 3 adopt the Tikhonov regularized inversion framework, and construct a multi-physics field collaborative property inversion objective function:

[0015] (2)

[0016] in, is the fitting error function of each physical field data, is the stability function of each physical property model, is the interactive guidance constraint function of the multi-property model, represents the square operation of the vector's two-norm, is the regularization parameter that stabilizes the model functions of different physical properties, is the constraint weight parameter corresponding to the update of different physical property models, is the model weight matrix, Guide coupling functions for multiphysics interactions.

[0017] As a further solution of the present invention, the maximum number of iterations of the inversion calculation of the inversion objective function is , given the target fitting accuracy of different physical field data and the initial model If it is determined that there is prior geological and drilling data information, it is directly assigned to the initial model in vector form.

[0018] As a further solution of the present invention, if it is determined that the prior geological and drilling data information is not available, the initial model is assigned a uniform physical property distribution.

[0019] As a further solution of the present invention, after the initial model is assigned, the inversion cyclic iterative calculation process is entered. The collaborative inversion process is decomposed into independent inversion processes of multiple physical field data. During the synchronization process, information transmission and mutual constraints between multiple physical property models are achieved through interactive guidance of coupling functions, which are in the following form:

[0020] (3)

[0021] Interactively guide the coupling function with three physics fields C For example, the definition is:

[0022] (4)

[0023] in, A and B It is a nonlinear function that extracts the structural characteristics of each physical field and is expressed as follows:

[0024] (5).

[0025] As a further solution of the present invention, the regularization parameter of the stable different physical property model function in the forward and inverse calculations is The selection method is as follows:

[0026] (6)

[0027] in, are the corresponding numbers of different physical field data, is the iteration number, Representative The next moment The data fitting error value of the physical field, It is The iteration time The stable function value of a physical property model, It is the attenuation coefficient, which can control the attenuation rate of the parameter between 0 and 1.

[0028] As a further solution of the present invention, the weight parameters of the corresponding constraint items are updated by different physical property models in the forward and inverse calculations. The selection method is as follows:

[0029] (7)

[0030] in, q is the proportional coefficient, the rate of change of the control parameter, j is the iteration number, Representative j The iteration time i The data fitting error value of the physical field, It is j The iteration time i The stable function value of a physical property model, For the j The interaction guide coupling function of the multi-physics field at the iteration moment.

[0031] As a further solution of the present invention, the inversion calculation continues to perform the independent inversion process of different physical fields, and updates the sub-objective function , and then judge the data fitting difference of each physical field to determine whether the inversion process converges to the target fitting accuracy If it is greater than the target fitting accuracy, the iterative cycle update will continue, otherwise the final inverted physical property model will be output and the inversion will end.

[0032] In summary, the embodiments of the present invention have the following beneficial effects compared with the prior art:

[0033] This method is of great significance for the comprehensive interpretation and modeling of multi-geophysical field data in mining areas. The resolution of the inversion imaging is higher than that of independent inversion of a single physical field, and the recovered physical property values ​​are closer to the actual physical properties. In addition, the algorithm is highly stable and more accurate than the commonly used cross-gradient collaborative inversion method for multi-physical field data. It does not produce local edge distortion and other effects, making it suitable for the actual needs of complex geological structures.

[0034] In order to more clearly illustrate the structural features and effects of the present invention, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a technical flow chart of the interactively guided multi-physical field collaborative physical property inversion method for a mining area in an embodiment of the invention.

[0036] Figure 2It is a diagram of ore body models of different burial depths and sizes and their geophysical observation field data in the embodiment of the invention.

[0037] Figure 3 This is a comparison example effect diagram of the single physical field independent inversion and the commonly used cross-gradient collaborative inversion in the embodiment of the invention. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0039] The specific implementation of the present invention is described in detail below with reference to specific embodiments.

[0040] In one embodiment, a multi-physics collaborative physical property inversion method for a mineral cluster area based on interactive guidance is described. Figures 1 to 3 , including the following steps:

[0041] Step 1: Divide the underground space into grids according to the horizontal position distribution of the observation points;

[0042] Step 2: Calculate the forward kernel matrix of each geophysical field, obtain the mapping relationship between the observation point and the underground discrete grid unit under the unit physical property, and establish the observation data through the forward kernel matrix d To Model m relationship;

[0043] Step 3: Obtain the three-dimensional discrete grids corresponding to different physical field observation points, and implement the three-dimensional discrete grid forward and inverse calculations of each physical field through interactively guided calculation processes.

[0044] For further information, see Figures 1 to 3 , the observation data in step 2 d To Model m The relationship is:

[0045] (1)

[0046] in, is the forward kernel matrix of each geophysical field, Represents different geophysical field observation data, are model parameters that respectively characterize the physical properties of the underground medium.

[0047] For further information, see Figures 1 to 3 The forward and inverse calculations in step 3 adopt the Tikhonov regularized inversion framework and construct a multi-physics field collaborative property inversion objective function:

[0048] (2)

[0049] in, is the fitting error function of each physical field data, is the stability function of each physical property model, is the interactive guidance constraint function of the multi-property model, represents the square operation of the vector's two-norm, is the regularization parameter that stabilizes the model functions of different physical properties, is the constraint weight parameter corresponding to the update of different physical property models, is the model weight matrix, Guide coupling functions for multiphysics interactions.

[0050] For further information, see Figures 1 to 3 , the maximum number of iterations of the inversion calculation of the inversion objective function , given the target fitting accuracy of different physical field data and the initial model If it is determined that there is prior geological and drilling data information, it is directly assigned to the initial model in vector form.

[0051] For further information, see Figures 1 to 3 If it is determined that the prior geological and drilling data information is not available, the initial model is assigned a uniform physical property distribution.

[0052] For further information, see Figures 1 to 3 After the initial model is assigned, the inversion cycle iterative calculation process begins. The collaborative inversion process is decomposed into independent inversion processes of multiple physical field data. During the synchronization process, the information transfer and mutual constraints between the multi-property models are realized through interactive guidance of the coupling function, which is in the following form:

[0053] (3)

[0054] Interactively guide the coupling function with three physics fields C For example, the definition is:

[0055] (4)

[0056] in, A and B It is a nonlinear function that extracts the structural characteristics of each physical field and is expressed as follows:

[0057] (5).

[0058] For further information, see Figures 1 to 3 , the regularization parameter of the stable different physical property model function in the forward and inverse calculations The selection method is as follows:

[0059] (6)

[0060] in, are the corresponding numbers of different physical field data, is the iteration number, Representative The next moment The data fitting error value of the physical field, It is The iteration time The stable function value of a physical property model, It is the attenuation coefficient, which can control the attenuation rate of the parameter between 0 and 1.

[0061] For further information, see Figures 1 to 3 , the corresponding constraint weight parameters of different physical property models in the forward and inverse calculations are updated The selection method is as follows:

[0062] (7)

[0063] in, q is the proportional coefficient, the rate of change of the control parameter, j is the iteration number, Representative j The iteration time i The data fitting error value of the physical field, It is j The iteration time i The stable function value of a physical property model, For the j The interaction guide coupling function of the multi-physics field at the iteration moment.

[0064] For further information, see Figures 1 to 3 , the inversion calculation continues to perform the independent inversion process of different physical fields, updating the sub-objective function , and then judge the data fitting difference of each physical field to determine whether the inversion process converges to the target fitting accuracy If it is greater than the target fitting accuracy, the iterative cycle update will continue, otherwise the final inverted physical property model will be output and the inversion will end.

[0065] In this embodiment, after geophysical field exploration using different methods and principles is conducted in an actual mineral cluster area, a multi-physics collaborative inversion overall objective function is constructed based on the different types of geophysical field exploration data obtained. Iterative loop solving and interactive guidance constraint updating of each physical property model are performed, ultimately obtaining a higher-resolution physical property model than a single physical field inversion. The detailed implementation scheme is as follows:

[0066] 1. Taking two different geophysical field detection data as an example, the underground space grid can be divided according to the horizontal position distribution of their observation points. Usually, the horizontal size of the grid corresponds to the point-line distance of the observation point, and the vertical size does not exceed twice the horizontal size, so as to obtain the three-dimensional discrete grid corresponding to the observation points of different physical fields for subsequent forward and inversion calculations. Before the inversion calculation, it is necessary to first complete the forward kernel matrix calculation of each geophysical field, that is, to obtain the mapping relationship between the observation point and the underground discrete grid unit under the unit physical property, and establish the observation data through the forward kernel matrix. d To Model m Relationship:

[0067] (1)

[0068] in, is the forward kernel matrix of each geophysical field, Represents observation data of different geophysical fields, are model parameters that respectively characterize the physical properties of the underground medium.

[0069] 2. Based on the Tikhonov regularized inversion framework, the objective function of multi-physics collaborative property inversion is constructed as follows:

[0070] (2)

[0071] in, is the fitting error function of each physical field data, is the stability function of each physical property model, is the interactive guidance constraint function of the multi-property model, represents the square operation of the vector's two-norm, is the regularization parameter that stabilizes the model functions of different physical properties, is the constraint weight parameter corresponding to the update of different physical property models, is the model weight matrix, is the interactive guidance coupling function of multiple physical fields. This method does not require manual experience to determine λ i and μ i Instead of weight parameters in the inversion process, they are stably, flexibly and automatically selected in an adaptive manner during the inversion process. See step 4 for details.

[0072] 3. Manually set the maximum number of iterations for collaborative inversion calculation , given the target fitting accuracy of different physical field data and the initial model The initial model is given in the following ways: 1) If there is prior geological, drilling data, and other information, the initial model can be directly assigned in vector form; 2) If there is no additional prior information, the initial model can be assigned according to the uniform physical property distribution. Once all the above information is initialized, the inversion iterative calculation process can be entered. The collaborative inversion process can be decomposed into independent inversion processes for multiple physical field data. During the synchronization process, the information transfer and mutual constraints between the multi-property models are achieved through interactive guidance of coupling functions, as shown below:

[0073] (3)

[0074] In order to maintain generality, the three physical fields are used to guide the coupling function C For example, the definition of is:

[0075] (4)

[0076] in, A and B It is a nonlinear function that extracts the structural characteristics of each physical field and is expressed as follows:

[0077] (5)

[0078] 4. Regularization parameters for stabilizing different physical property model functions during the iteration process Constraint weight parameters corresponding to different physical property model updates It has a great impact on the stability of the inversion process and the recovery effect of the inversion model. In order to avoid the difficulty and subjective interference of human judgment, this method constructs an adaptive parameter selection method, which is as follows.

[0079] Regularization parameters for stabilizing model functions with different physical properties λ i Selection method:

[0080] (6)

[0081] in, are the corresponding numbers of different physical field data, is the iteration number, Representative The next moment The data fitting error value of the physical field, It is The iteration time The stable function value of a physical property model, It is the attenuation coefficient, which can control the attenuation rate of the parameter between 0 and 1.

[0082] Constraint weight parameters corresponding to different physical property model updates The selection method is:

[0083] (7)

[0084] in, q is the proportional coefficient, which controls the rate of change of the parameter. For the j The first iteration moment is the interactive guidance coupling function of the multi-physics field, and the other symbols have the same meanings as in formula (6).

[0085] 5. Continue to perform the independent inversion process of different physical fields and update the sub-objective function , and then judge the data fitting difference of each physical field to determine whether the inversion process converges to the target fitting accuracy If it is greater than the target fitting accuracy, it returns to step 4 to continue iterative cycle update; otherwise, it outputs the final inverted physical property model and the inversion ends.

[0086] Simulate a set of density and magnetization intensity distribution models corresponding to the ore body model and the corresponding gravity field and magnetic field two physical field observation data, such as Figure 2 As shown, Figure 2 (a) Density distribution corresponding to the ore body model; (b) Gravity field observation data corresponding to the density model; (c) Magnetization intensity distribution corresponding to the ore body model; (d) Magnetic field observation data corresponding to the magnetization intensity model. The model consists of two massive geological bodies of different sizes and different burial depths. The observed gravity field in this model is relatively sensitive to the larger geological bodies in the deep part, with a larger anomaly amplitude. The observed magnetic field is relatively sensitive to the smaller geological bodies in the shallow part, with a larger anomaly amplitude and a higher resolution. The method of the present invention is used to calculate the density and magnetization intensity distribution of the target object in the underground space by observing the gravity field and magnetic field anomaly data, and the following is obtained: Figure 3 The inversion result of Figure 3 The white wireframe in the middle is the actual position corresponding to the geological model. Figure 3 (a) Density model obtained from independent inversion of gravity field data; (b) Magnetization intensity model obtained from independent inversion of magnetic field data; (c) Density model obtained from commonly used coordinated inversion of gravity and magnetic field based on cross gradients; (d) Magnetization intensity model obtained from commonly used coordinated inversion of gravity and magnetic field based on cross gradients; (e) Density model obtained by the method of the present invention; (f) Magnetization intensity model obtained by the method of the present invention.

[0087] The results show that the new method significantly improves the resolution of independent inversion of gravity and magnetic field data, and the recovered density and magnetization amplitudes are closer to the actual physical property range of the model. Compared with the commonly used cross-gradient-based collaborative inversion method, the method of the present invention can more clearly restore the boundaries of geological bodies, and the model boundaries do not exhibit morphological distortions that interfere with the judgment of interpreters. This shows that the method can integrate the sensitive characteristics of each physical field, achieve complementary advantages, and has high stability, providing methodological and technical support for application in actual mining areas.

[0088] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A multi-physics field collaborative physical property inversion method for a mining area based on interactive guidance, characterized in that: The following steps are involved: Step 1: Divide the underground space into grids according to the horizontal position distribution of the observation points; Step 2: Calculate the forward kernel matrix of each geophysical field, obtain the mapping relationship between the observation point and the underground discrete grid unit under the unit physical property, and establish the observation data through the forward kernel matrix d To Model m relationship; Step 3: Obtain the three-dimensional discrete grids corresponding to different physical field observation points, and implement the forward and inverse calculations of the three-dimensional discrete grids of each physical field through interactive guidance of the calculation process; Among them, the observation data in step 2 d To Model m The relationship is: （1） in, is the forward kernel matrix of each geophysical field, Represents different geophysical field observation data, are model parameters corresponding to the physical properties of the underground medium; The forward and inverse calculations in step 3 adopt the Tikhonov regularized inversion framework and construct a multi-physics field collaborative property inversion objective function: （2） in, is the fitting error function of each physical field data, is the stability function of each physical property model, is the interactive guidance constraint function of the multi-property model, represents the square operation of the vector's two-norm, is the regularization parameter that stabilizes the model functions of different physical properties, is the constraint weight parameter corresponding to the update of different physical property models, is the model weight matrix, Guide coupling functions for multiphysics interactions.

2. The interactive guidance-based multi-physics field collaborative physical property inversion method for a mineral concentration area according to claim 1 is characterized in that: The maximum number of iterations of the inversion calculation of the inversion objective function , given the target fitting accuracy of different physical field data and the initial model If it is determined that there is prior geological and drilling data information, it is directly assigned to the initial model in vector form.

3. The interactive guidance-based multi-physics field collaborative physical property inversion method for a mineral concentration area according to claim 2 is characterized in that: If it is determined that the prior geological and drilling data information is not available, the initial model is assigned a uniform physical property distribution.

4. The interactive guidance-based multi-physics field collaborative physical property inversion method for a mineral concentration area according to claim 3 is characterized in that: After the initial model is assigned, the inversion cycle iterative calculation process begins. The collaborative inversion process is decomposed into independent inversion processes of multiple physical field data. During the synchronization process, the information transfer and mutual constraints between the multi-property models are realized through interactive guidance of coupling functions, which are in the following form: （3） Interactively guide the coupling function with three physics fields C For example, the definition of is: （4） in, A and B It is a nonlinear function that extracts the structural characteristics of each physical field and is expressed as follows: （5）。 5. The interactive guidance-based multi-physics field collaborative physical property inversion method for a mineral concentration area according to claim 4 is characterized in that: The regularization parameter of the stable different physical property model function in the forward and inverse calculations The selection method is as follows: （6） in, are the corresponding numbers of different physical field data, is the iteration number, Representative The next moment The data fitting error value of the physical field, It is The iteration time The stable function value of a physical property model, It is the attenuation coefficient, which can control the attenuation rate of the parameter between 0 and 1.

6. The interactive guidance-based multi-physics field collaborative physical property inversion method for a mineral concentration area according to claim 5 is characterized in that: The constraint item weight parameters corresponding to the update of different physical property models in the forward and inverse calculations The selection method is as follows: （7） in, q is the proportional coefficient, the rate of change of the control parameter, j is the iteration number, Representative j The iteration time i The data fitting error value of the physical field, It is j The iteration time i The stable function value of a physical property model, For the j The interaction guide coupling function of the multi-physics field at the iteration moment.

7. The interactive guidance-based multi-physics field collaborative physical property inversion method for a mineral concentration area according to claim 6 is characterized in that: The inversion calculation continues to perform the independent inversion process of different physical fields, updating the sub-objective function , and then judge the data fitting difference of each physical field to determine whether the inversion process converges to the target fitting accuracy If it is greater than the target fitting accuracy, the iterative cycle update will continue, otherwise the final inverted physical property model will be output and the inversion will end.

Images