A real-time correction method and system for groundwater contaminant concentration field
Patent Information
- Application Number
- CN202610986254.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-03
- Publication Date
- 2026-09-25
AI Technical Summary
[0004]鉴于以上现有技术的不足，本发明实施例的目的在于提供一种地下水污染物浓度场的实时校正方法，能够解决现有技术存在的同化结果不稳定且不准确的技术问题
[0039]在本发明实施例中，通过本征正交分解将高维物理空间降维构建潜空间，依托统一投影规则完成观测数据与浓度场的空间转换，实现观测信息有效对齐。借助卡尔曼滤波在低维潜空间完成状态更新，规避了高维空间运算易出现的伪相关、滤波发散问题，同时解决稀疏观测与高维模型维度不匹配的缺陷，能够充分发挥观测数据的约束作用，有效提升数据同化结果的稳定性与准确性，保障地下水污染物浓度场实时校正效果。
Smart Images
Figure CN122817795A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of environmental engineering technology, and in particular to a method and system for real-time correction of groundwater pollutant concentration fields. Background Technology
[0002] Accurate prediction of groundwater pollutant transport is crucial for environmental governance and emergency early warning. Existing technologies typically employ ensemble Kalman filtering based on physical grid space for data assimilation: the study area is discretized into hundreds of thousands of grid cells to generate initial ensemble members, time-lapse is performed using a numerical simulator, and after acquiring sparse monitoring well data, a massive background error covariance matrix is calculated on the entire physical grid space. Subsequently, the Kalman gain is solved to update the concentration values of all grid points.
[0003] However, this method operates directly in the original high-dimensional physical space. When the physical space grid is large (high-dimensionality) and the observation data is extremely sparse and accompanied by noise, the assimilation process suffers from the "curse of dimensionality" because it needs to calculate and store a covariance matrix that is proportional to the square of the number of grids. This results in low computational efficiency and a high risk of memory overflow. At the same time, a small number of observations cannot provide sufficient constraints and are prone to spurious correlations, making the assimilation results unstable and inaccurate. Summary of the Invention
[0004] In view of the shortcomings of the prior art, the purpose of this invention is to provide a real-time correction method for groundwater pollutant concentration field, which can solve the technical problems of unstable and inaccurate assimilation results in the prior art.
[0005] A first aspect of this invention provides a real-time correction method for groundwater pollutant concentration fields, comprising:
[0006] The historical groundwater concentration field snapshot matrix of the region to be corrected is subjected to intrinsic orthogonal decomposition to obtain latent variables projected onto the latent space; the latent space is obtained by dimensionality reduction of the physical space of the region to be corrected.
[0007] Observational data of the area to be calibrated are collected through multiple monitoring wells distributed in the area to be calibrated;
[0008] The observation data is converted into a latent space observation vector for the latent space;
[0009] Based on the latent space observation vector, the latent variables are updated using Kalman filtering to obtain latent space analysis values;
[0010] The latent space analysis values are inversely mapped to the physical space to obtain the corrected high-dimensional concentration distribution field.
[0011] Optionally, the historical groundwater concentration field snapshot matrix of the area to be corrected is subjected to intrinsic orthogonal decomposition to obtain latent variables projected onto the latent space, specifically including:
[0012] Multiple historical groundwater concentration field data for the area to be corrected were collected.
[0013] Multiple intrinsic modes were determined from the aforementioned historical groundwater concentration field data;
[0014] The multiple intrinsic modes are spanned to obtain the latent space corresponding to the physical space of the region to be corrected;
[0015] Construct a projection matrix based on the multiple intrinsic modes;
[0016] Based on the projection matrix, the physical space is projected onto the latent space to obtain latent variables.
[0017] Optionally, determining multiple intrinsic modes from the multiple historical groundwater concentration field data specifically includes:
[0018] Multiple high-dimensional groundwater concentration field data are arranged in columns to obtain a concentration field snapshot matrix;
[0019] The concentration field snapshot matrix is subjected to singular value decomposition to obtain multiple initial intrinsic modes and the energy value of each initial intrinsic mode;
[0020] A preset number of eigenmodes are selected from multiple initial eigenmodes; wherein the energy values of the selected eigenmodes are all greater than the energy values of the unselected initial eigenmodes.
[0021] Optionally, the step of projecting the physical space onto the latent space based on the projection matrix to obtain latent variables specifically includes:
[0022] Determine the high-dimensional concentration field vector from the physical space;
[0023] Based on the high-dimensional concentration field vector and the projection matrix, latent variables projected onto the latent space are obtained.
[0024] Optionally, converting the observation data into a latent space observation vector specifically includes:
[0025] The observation data is converted into a pseudo-observation field based on the mosaic interpolation method;
[0026] The pseudo-observation field is mapped into the latent space based on the projection matrix to obtain the latent space observation vector.
[0027] Optionally, the step of converting the observation data into a pseudo-observation field based on mosaic interpolation specifically includes:
[0028] The polygonal influence zone of each monitoring well is determined from the observation data based on the mosaic interpolation method;
[0029] The observation data of any monitoring well is assigned to all grid nodes within the polygonal influence area of that monitoring well to obtain a pseudo-observation field.
[0030] Optionally, the step of inversely mapping the latent space analysis values to the physical space to obtain the corrected high-dimensional concentration distribution field specifically includes:
[0031] The reconstruction matrix is determined based on the projection matrix;
[0032] Based on the reconstruction matrix, the latent space analysis values are inversely mapped to the physical space to obtain the corrected high-dimensional concentration distribution field.
[0033] Optionally, determining the reconstruction matrix based on the projection matrix specifically includes:
[0034] The projection matrix is transposed to obtain the reconstructed matrix.
[0035] A second aspect of this invention provides a real-time correction system for groundwater pollutant concentration fields, comprising: a processor and a memory;
[0036] The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the real-time correction method for the groundwater pollutant concentration field as described in the first aspect.
[0037] A third aspect of the present invention provides a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the real-time correction method for groundwater pollutant concentration field as described in the first aspect.
[0038] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:
[0039] In this embodiment of the invention, the high-dimensional physical space is reduced to a latent space through intrinsic orthogonal decomposition. A unified projection rule is used to complete the spatial transformation between observation data and the concentration field, achieving effective alignment of observation information. Kalman filtering is used to complete state updates in the low-dimensional latent space, avoiding the pseudo-correlation and filter divergence problems that easily occur in high-dimensional space operations. It also solves the defect of dimensionality mismatch between sparse observations and high-dimensional models, fully leveraging the constraints of observation data, effectively improving the stability and accuracy of data assimilation results, and ensuring the real-time correction effect of the groundwater pollutant concentration field. Attached Figure Description
[0040] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.
[0041] Figure 1 This is a flowchart illustrating a real-time correction method for groundwater pollutant concentration field provided in an embodiment of the present invention.
[0042] Figure 2 This is a schematic diagram of the structure of a real-time correction system for groundwater pollutant concentration field provided in an embodiment of the present invention. Detailed Implementation
[0043] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.
[0044] The real-time correction method for groundwater pollutant concentration field provided by the present invention will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.
[0045] Reference manual attached Figure 1 The diagram shows a flowchart of a real-time correction method for groundwater pollutant concentration field provided by an embodiment of the present invention.
[0046] This invention provides a method for real-time correction of groundwater pollutant concentration fields, which may include the following steps:
[0047] S1: Perform eigenorthogonal decomposition on the historical groundwater concentration field snapshot matrix of the area to be corrected to obtain latent variables projected onto the latent space.
[0048] In this embodiment, the latent space is obtained by dimensionality reduction of the physical space of the region to be corrected. The latent space is a low-dimensional compressed feature space hidden behind the original high-dimensional data. Latent variables are a set of specific coordinate values obtained by projecting the physical space onto the latent space. Proper Orthogonal Decomposition (POD) is a linear method for data dimensionality reduction by extracting principal components.
[0049] By performing singular value decomposition (SVD) on the snapshot matrix of historical simulation data, the top r eigenmodes (i.e., basis vectors) that best represent the characteristics of the physical field are extracted. These basis vectors are then used to construct a projection matrix, linearly projecting the tens of thousands of dimensions of the concentration field C in physical space onto a low-dimensional vector z (i.e., latent variables). In this way, the physical field is compressed, and the process of truncating low-order modes naturally filters out high-frequency random noise in the initial field.
[0050] As an optional implementation, S1 may perform eigenorthogonal decomposition on the historical groundwater concentration field snapshot matrix of the area to be corrected to obtain latent variables projected onto the latent space in the following ways:
[0051] Multiple historical groundwater concentration field data for the area to be corrected were collected.
[0052] Multiple intrinsic modes were determined from the aforementioned historical groundwater concentration field data;
[0053] The multiple intrinsic modes are spanned to obtain the latent space corresponding to the physical space of the region to be corrected;
[0054] Construct a projection matrix based on the multiple intrinsic modes;
[0055] Based on the projection matrix, the physical space is projected onto the latent space to obtain latent variables.
[0056] This implementation method extracts eigenmodes from historical concentration field data, constructs a latent space, and generates a corresponding projection matrix, accurately preserving the core features of the concentration field. Relying on the same set of eigenmodes for spatial construction and data projection ensures consistent and coherent transformation rules, guaranteeing no deviation in data characteristics. Transforming high-dimensional data into low-dimensional latent variables simplifies subsequent calculations and makes the data representation more realistic, laying a solid foundation for subsequent observation matching and concentration field correction, resulting in a smoother process and more reliable results.
[0057] As an optional implementation, the method for determining multiple intrinsic modes from the multiple historical groundwater concentration field data may include:
[0058] Multiple high-dimensional groundwater concentration field data are arranged in columns to obtain a concentration field snapshot matrix;
[0059] The concentration field snapshot matrix is subjected to singular value decomposition to obtain multiple initial intrinsic modes and the energy value of each initial intrinsic mode;
[0060] A preset number of eigenmodes are selected from multiple initial eigenmodes; wherein the energy values of the selected eigenmodes are all greater than the energy values of the unselected initial eigenmodes.
[0061] This implementation method first integrates various concentration field data into a snapshot matrix, and then decomposes it to obtain different intrinsic modes and their corresponding energy levels. The main modes are selected according to energy level, prioritizing the retention of core information that reflects changes in the concentration field while discarding useless interference noise. This simplifies the data, reduces the computational burden, and ensures the authenticity and effectiveness of the extracted features, making the subsequently constructed latent space more closely reflect reality and laying a solid foundation for the entire concentration field correction work.
[0062] In this embodiment, the preset quantity can be determined in advance based on experience.
[0063] As an optional implementation, the method of projecting the physical space onto the latent space based on the projection matrix to obtain latent variables may include:
[0064] Determine the high-dimensional concentration field vector from the physical space;
[0065] Based on the high-dimensional concentration field vector and the projection matrix, latent variables projected onto the latent space are obtained.
[0066] This implementation method directly transforms the high-dimensional concentration field vector using a projection matrix to quickly obtain the corresponding latent variables. The transformation process is simple and direct, with clear operational logic. It relies on unified projection rules to complete spatial mapping, fully preserving the original characteristics of the concentration field. Transforming high-dimensional data into low-dimensional latent variables effectively reduces the data dimensionality, significantly alleviating subsequent computational burden while ensuring accurate data correspondence, facilitating smoother subsequent filtering updates and concentration field correction.
[0067] S2: Collect observation data of the area to be calibrated through multiple monitoring wells distributed in the area to be calibrated.
[0068] S3: Convert the observation data into a latent space observation vector for the latent space.
[0069] In this embodiment, for monitoring well data that is sparse and irregularly distributed in physical space, the proposed solution first employs Voronoi tessellation to construct a polygonal influence zone centered on each observation point, transforming discrete "point" observations into a pseudo-observation field Cv covering the entire grid. Subsequently, the same projection matrix is used to map Cv into the latent space, yielding the latent space observation vector Zv. This step solves the technical problem of mismatch between observation operators in high- and low-dimensional spaces in traditional methods, achieving feature alignment of observation information in the latent space.
[0070] As an optional implementation, S3 may convert the observation data into a latent space observation vector of the latent space in the following ways:
[0071] The observation data is converted into a pseudo-observation field based on the mosaic interpolation method;
[0072] The pseudo-observation field is mapped into the latent space based on the projection matrix to obtain the latent space observation vector.
[0073] This implementation method uses mosaicking interpolation to piece together fragmented observation data into a complete pseudo-observation field, solving the problems of sparse monitoring points and uneven data distribution. Then, a unified projection matrix is used to complete spatial mapping, placing the observation data and concentration field data in the same latent space, ensuring data matching. The entire conversion process is simple and practical, fully utilizing measured observation information while adapting to subsequent filtering operations, allowing the observation data to effectively play a constraining role and improving the overall correction effect.
[0074] Optionally, the method of converting the observation data into a pseudo-observation field based on mosaic interpolation may include:
[0075] The polygonal influence zone of each monitoring well is determined from the observation data based on the mosaic interpolation method;
[0076] The observation data of any monitoring well is assigned to all grid nodes within the polygonal influence area of that monitoring well to obtain a pseudo-observation field.
[0077] This implementation method, by dividing the influence area of each monitoring well and assigning the measured observation values to the corresponding grid nodes, quickly generates a complete pseudo-observation field. This effectively fills the data gaps caused by the sparse monitoring points, allowing discrete observation data to cover the entire area to be corrected. The interpolation method is simple and practical, accurately restoring the observation information, ensuring the consistency and uniformity of the entire field data, facilitating subsequent spatial projection, and allowing the observation data to smoothly participate in subsequent concentration field correction work.
[0078] S4: Based on the latent space observation vector, use Kalman filtering to update the latent variables and obtain the latent space analysis values.
[0079] In this embodiment, the Kalman filter can be an ensemble Kalman filter (EnKF), a serialized data assimilation algorithm that estimates the system error covariance using a sample set.
[0080] In this embodiment, the assimilation process is performed entirely in a closed loop within the latent space. The system calculates the background error covariance matrix Pz within the latent space. To prevent filter instability due to modal energy differences, this invention introduces a scaling factor α to regularize the latent space observation error covariance matrix. By solving for the Kalman gain in the latent space, the predicted latent variable Z(f) is updated to obtain the analysis value Z(a).
[0081] S5: Inversely map the latent space analysis values to the physical space to obtain the corrected high-dimensional concentration distribution field.
[0082] In this embodiment, Z(a) is finally mapped back to the physical space using the inverse transformation (reconstruction) of POD to obtain the final corrected high-dimensional concentration distribution field C(a).
[0083] As an optional implementation, S5 can be performed by inversely mapping the latent space analysis values to the physical space to obtain the corrected high-dimensional concentration distribution field in the following ways:
[0084] The reconstruction matrix is determined based on the projection matrix;
[0085] Based on the reconstruction matrix, the latent space analysis values are inversely mapped to the physical space to obtain the corrected high-dimensional concentration distribution field.
[0086] This implementation method derives the reconstruction matrix from the original projection matrix without requiring recalculation of modes, making the process simple and efficient. Using the accompanying matrix for inverse mapping fully restores the original characteristics of the concentration field, avoiding data distortion. The high-dimensional concentration distribution field is reconstructed from low-dimensional analysis values, successfully completing the spatial transformation and ensuring the correction results closely match the actual scenario. This also makes the entire correction process seamless, resulting in accurate and reliable output concentration field data.
[0087] Optionally, the reconstruction matrix can be determined based on the projection matrix in the following ways:
[0088] The projection matrix is transposed to obtain the reconstructed matrix.
[0089] This implementation method simplifies the process by obtaining the reconstructed matrix simply by transposing the projection matrix. It is computationally efficient, requires minimal computation, and eliminates the need for additional modal data solving, significantly simplifying the process. The transposed reconstructed matrix exhibits a high degree of matching with the projection matrix, ensuring consistency in spatial mapping rules and preserving data characteristics throughout the inverse transformation process, minimizing the risk of deviation. This approach improves overall computational efficiency while maintaining the stability of spatial transformation, providing reliable support for subsequent reconstruction of high-dimensional concentration fields.
[0090] In this embodiment, physical interpretability means that the core dimensionality reduction (POD) and assimilation (EnKF) processes of this technology follow the characteristic decomposition principle of statistical fluid dynamics and have clear physical meaning.
[0091] Experimental data support the conclusion that, in tests on strongly heterogeneous permeable fields, the latent space data assimilation framework offers an order-of-magnitude improvement in computational efficiency compared to physical space assimilation, while maintaining extremely high prediction accuracy even with limited observations.
[0092] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:
[0093] In this embodiment of the invention, the high-dimensional physical space is reduced to a latent space through intrinsic orthogonal decomposition. A unified projection rule is used to complete the spatial transformation between observation data and the concentration field, achieving effective alignment of observation information. Kalman filtering is used to update the state in the low-dimensional latent space, avoiding the pseudo-correlation and filter divergence problems that easily occur in high-dimensional space operations. It also addresses the dimensionality mismatch between sparse observations and high-dimensional models, fully leveraging the constraints of observation data to effectively improve the stability and accuracy of data assimilation results and ensure real-time correction of the groundwater pollutant concentration field.
[0094] Reference manual attached Figure 2 The diagram shows a schematic of the structure of a real-time correction system for groundwater pollutant concentration field provided in an embodiment of the present invention.
[0095] This invention provides a real-time correction system 20 for groundwater pollutant concentration fields, comprising: a processor 201 and a memory 202;
[0096] The memory 202 stores programs or instructions that can run on the processor 201. When the program or instructions are executed by the processor 201, they implement the steps of the above-mentioned real-time correction method for groundwater pollutant concentration field and achieve the same technical effect. To avoid repetition, the present invention will not repeat the above-mentioned steps.
[0097] It should be understood that the processor 201 in this embodiment of the invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.
[0098] It should also be understood that the memory 202 in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DR RAM).
[0099] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.
[0100] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
[0101] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.
[0102] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0103] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.
[0104] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0105] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0106] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0107] This invention provides a readable storage medium comprising: storing a program or instructions on the readable storage medium, wherein when the program or instructions are executed by a processor, the program or instructions implement the steps of the above-described real-time correction method for groundwater pollutant concentration field, and can achieve the same technical effect. To avoid repetition, this invention will not elaborate further.
[0108] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention.
Claims
1. A method for real-time correction of groundwater pollutant concentration field, characterized in that, include: The historical groundwater concentration field snapshot matrix of the region to be corrected is subjected to intrinsic orthogonal decomposition to obtain latent variables projected onto the latent space; wherein, the latent space is obtained by dimensionality reduction of the physical space of the region to be corrected. Observational data of the area to be calibrated are collected through multiple monitoring wells distributed in the area to be calibrated; The observation data is converted into a latent space observation vector for the latent space; Based on the latent space observation vector, the latent variables are updated using Kalman filtering to obtain latent space analysis values; The latent space analysis values are inversely mapped to the physical space to obtain the corrected high-dimensional concentration distribution field.
2. The real-time correction method for groundwater pollutant concentration field according to claim 1, characterized in that, The historical groundwater concentration field snapshot matrix of the area to be corrected is subjected to intrinsic orthogonal decomposition to obtain latent variables projected onto the latent space, specifically including: Multiple historical groundwater concentration field data for the area to be corrected were collected. Multiple intrinsic modes were determined from the aforementioned historical groundwater concentration field data; The multiple intrinsic modes are spanned to obtain the latent space corresponding to the physical space of the region to be corrected; Construct a projection matrix based on the multiple intrinsic modes; Based on the projection matrix, the physical space is projected onto the latent space to obtain latent variables.
3. The real-time correction method for groundwater pollutant concentration field according to claim 2, characterized in that, The determination of multiple intrinsic modes from the multiple historical groundwater concentration field data specifically includes: Multiple high-dimensional groundwater concentration field data are arranged in columns to obtain a concentration field snapshot matrix; The concentration field snapshot matrix is subjected to singular value decomposition to obtain multiple initial intrinsic modes and the energy value of each initial intrinsic mode; A preset number of eigenmodes are selected from multiple initial eigenmodes; wherein the energy values of the selected eigenmodes are all greater than the energy values of the unselected initial eigenmodes.
4. The real-time correction method for groundwater pollutant concentration field according to claim 3, characterized in that, The step of projecting the physical space onto the latent space based on the projection matrix to obtain latent variables specifically includes: Determine the high-dimensional concentration field vector from the physical space; Based on the high-dimensional concentration field vector and the projection matrix, latent variables projected onto the latent space are obtained.
5. The real-time correction method for groundwater pollutant concentration field according to claim 2, characterized in that, The step of converting the observation data into a latent space observation vector specifically includes: The observation data is converted into a pseudo-observation field based on the mosaic interpolation method; The pseudo-observation field is mapped into the latent space based on the projection matrix to obtain the latent space observation vector.
6. The real-time correction method for groundwater pollutant concentration field according to claim 5, characterized in that, The process of converting the observation data into a pseudo-observation field based on mosaic interpolation specifically includes: The polygonal influence zone of each monitoring well is determined from the observation data based on the mosaic interpolation method; The observation data of any monitoring well is assigned to all grid nodes within the polygonal influence area of that monitoring well to obtain a pseudo-observation field.
7. The real-time correction method for groundwater pollutant concentration field according to claim 2, characterized in that, The step of inversely mapping the latent space analysis values to the physical space to obtain the corrected high-dimensional concentration distribution field specifically includes: The reconstruction matrix is determined based on the projection matrix; Based on the reconstruction matrix, the latent space analysis values are inversely mapped to the physical space to obtain the corrected high-dimensional concentration distribution field.
8. The real-time correction method for groundwater pollutant concentration field according to claim 7, characterized in that, The process of determining the reconstruction matrix based on the projection matrix specifically includes: The projection matrix is transposed to obtain the reconstructed matrix.
9. A real-time correction system for groundwater pollutant concentration field, characterized in that, include: Processor and memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the real-time correction method for the groundwater pollutant concentration field as described in any one of claims 1 to 8.
10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the real-time correction method for the groundwater pollutant concentration field as described in any one of claims 1 to 8.