Cold region groundwater level true value reconstruction method and device based on dual latent space decoupling

Abstract

This invention proposes a method and apparatus for reconstructing the true groundwater level in cold regions based on dual-latency space decoupling. When the water level deviation exhibits continuous shifts or synchronous changes with the freeze-thaw cycle, the observation distortion trend at the k-th sampling time is determined based on the displacement characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label at the k-th sampling time. The target window data is input into a dual-latency space decoupling network, which extracts the true hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time. Based on these variables, the reconstructed water level at the k-th sampling time is determined. By using the dual-latency space decoupling network to separately represent the true hydrodynamic changes and observation distortion changes, the aliasing between the two within the model is reduced, improving the reliability and interpretability of the groundwater level true reconstruction results.

Description

Technical Field

[0001] This invention relates to the field of data processing, and more specifically, to a method and apparatus for reconstructing the true value of groundwater level in cold regions based on dual-potential space decoupling. Background Technology

[0002] Groundwater level is an important indicator reflecting the dynamic changes of the groundwater system and is widely used in hydrological analysis, engineering evaluation, and ecological environment monitoring. Current groundwater level monitoring typically uses methods such as pressure level gauges, float level gauges, cable-type measuring devices, or manual calibration to obtain well water level data.

[0003] In cold environments such as permafrost regions, seasonally frozen soil regions, and high-altitude mountainous areas, monitoring wells and the surrounding soil are easily affected by alternating freezing and thawing, leading to phenomena such as frost heave, thaw settlement, wellhead tilting, well deformation, and changes in the relative elevation of sensors. Due to these factors, the raw groundwater level sequence output by monitoring equipment often includes not only the actual groundwater dynamic changes but also observational distortions caused by well displacement, sensor offset, and low-temperature environmental responses, resulting in reduced accuracy of groundwater level observations. Summary of the Invention

[0004] The purpose of this invention is to provide a method and apparatus for reconstructing the true value of groundwater level in cold regions based on dual-submersible space decoupling, so as to improve the above-mentioned problems.

[0005] To achieve the above objectives, the technical solutions adopted in the embodiments of the present invention are as follows: In a first aspect, embodiments of the present invention provide a method for reconstructing the true value of groundwater level in cold regions based on dual-latency space decoupling, including: Acquire the basic observation sequence of the monitoring well in the cold region during the observation period. The basic observation sequence includes any one or more of the following: original groundwater level sequence, well pressure sequence, temperature observation sequence, and displacement state observation sequence. The freeze-thaw state characteristics at the k-th sampling time are determined based on the temperature observation sequence, where the freeze-thaw state characteristics include freezing increment and thawing increment; The displacement state characteristics and displacement change rate at the k-th sampling time are extracted from the displacement state observation sequence; Based on the freezing increment, melting increment, and displacement change rate at the k-th sampling time, construct the freeze-thaw stage label at the k-th sampling time; When the water level deviation shows a continuous shift or changes synchronously with the freeze-thaw stage, the trend of observation distortion at the k-th sampling time is determined based on the displacement state characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label at the k-th sampling time. The target window data is input into the dual-potential space decoupling network. The dual-potential space decoupling network extracts the real hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time. The target window data includes the original groundwater level, hydrostatic reference water level, freeze-thaw stage label, and observation distortion trend at each sampling time in the target window. The target window is from the (k-ω+1)-th sampling time to the k-th sampling time, where ω represents the window length. The reconstructed water level at the k-th sampling time is determined based on the actual hydrodynamic latent variables and the observed distortion latent variables at the k-th sampling time.

[0006] Secondly, embodiments of the present invention provide a groundwater level truth reconstruction device for cold regions based on dual-submersible space decoupling, the device comprising: The first processing unit is used to acquire the basic observation sequence of the cold region monitoring well during the observation period. The basic observation sequence includes any one or more of the following: original groundwater level sequence, well pressure sequence, temperature observation sequence, and displacement state observation sequence. The second processing unit is used to determine the freeze-thaw state characteristics at the k-th sampling time based on the temperature observation sequence, wherein the freeze-thaw state characteristics include freezing increment and thawing increment; The second processing unit is also used to extract the displacement state features and displacement change rate at the kth sampling time from the displacement state observation sequence; The second processing unit is also used to construct a freeze-thaw stage label for the k-th sampling time based on the freeze increment, thaw increment, and displacement change rate at the k-th sampling time. The second processing unit is also used to determine the observation distortion trend at the k-th sampling time based on the displacement state characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label when the water level deviation shows a continuous shift or changes synchronously with the freeze-thaw stage. The second processing unit is also used to input the target window data into the dual-potential space decoupling network. The dual-potential space decoupling network extracts the real hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time. The target window data includes the original groundwater level, hydrostatic reference water level, freeze-thaw stage label and observation distortion change trend at each sampling time in the target window. The target window is from the (k-ω+1)-th sampling time to the k-th sampling time, where ω represents the window length. The second processing unit is also used to determine the reconstructed water level at the k-th sampling time based on the actual hydrodynamic latent variables and the observed distortion latent variables at the k-th sampling time.

[0007] Thirdly, embodiments of the present invention provide a storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method.

[0008] Fourthly, embodiments of the present invention provide an electronic device, the electronic device comprising: a processor and a memory, the memory being used to store one or more programs; when the one or more programs are executed by the processor, the above-described method is implemented.

[0009] Compared to existing technologies, the present invention provides a method and apparatus for reconstructing the true groundwater level in cold regions based on dual-latency space decoupling. When the water level deviation exhibits continuous shifts or synchronous changes with the freeze-thaw cycle, the method determines the observation distortion trend at the k-th sampling time based on the displacement characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label. The target window data is input into a dual-latency space decoupling network, which extracts the true hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time. Based on these variables, the reconstructed water level at the k-th sampling time is determined. By using a dual-latency space decoupling network to separately express the true hydrodynamic changes and observation distortion changes, the aliasing between the two within the model is reduced, improving the reliability and interpretability of the groundwater level true reconstruction results.

[0010] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0013] Figure 2 This is a flowchart illustrating the method for reconstructing the true value of groundwater level in cold regions based on dual-potential space decoupling, as provided in an embodiment of the present invention.

[0014] Figure 3 This is a schematic diagram of a unit for a groundwater level truth reconstruction device in cold regions based on dual-submerged space decoupling, provided in an embodiment of the present invention.

[0015] In the diagram: 10-Processor; 11-Memory; 12-Bus; 13-Communication interface; 801-First processing unit; 802-Second processing unit. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments 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, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0017] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0018] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0019] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0020] In the description of this invention, it should be noted that the terms "upper," "lower," "inner," "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the product of this invention is usually placed when in use. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0021] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set" and "connection" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0022] The following detailed description of some embodiments of the present invention is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0023] Existing groundwater level data processing methods mostly employ smoothing, interpolation, filtering, regression correction, or ordinary time-series prediction models, primarily used to reduce random noise, supplement missing data, or predict water level change trends. However, groundwater level observation errors in cold regions are usually related to freeze-thaw cycles, well displacement, and changes in sensor installation status, exhibiting staged and structural characteristics rather than simple random noise. Therefore, conventional processing methods struggle to effectively distinguish between actual groundwater level changes and observed distortions, easily mistaking deviations caused by freeze-thaw cycles, thaw settlement, or sensor displacement for genuine groundwater level fluctuations.

[0024] In addition, monitoring sites in cold regions can typically obtain information such as air temperature, soil temperature, well pressure, cable condition, wellhead inclination, displacement observations, and manually calibrated water levels. This information can reflect whether there is distortion in water level observations from aspects such as freeze-thaw environment, equipment condition, and hydrostatic pressure relationship. However, existing methods lack an effective mechanism to use the above information to construct a proxy for observation distortion and further separate the true hydrodynamic components from the observation distortion components.

[0025] Therefore, it is necessary to propose a groundwater level true value reconstruction method in cold regions, which can combine freeze-thaw state, displacement state and hydrostatic pressure relationship to separate the real hydrodynamic changes and observation distortion changes in the original groundwater level observation sequence, thereby obtaining a more reliable groundwater level true value sequence.

[0026] This invention provides an electronic device, which may be a server device, a computer device, or a mobile phone device. Please refer to... Figure 1 This is a schematic diagram of the structure of an electronic device. The electronic device includes a processor 10, a memory 11, and a bus 12. The processor 10 and the memory 11 are connected via the bus 12. The processor 10 is used to execute executable modules, such as computer programs, stored in the memory 11.

[0027] Processor 10 can be an integrated circuit chip with signal processing capabilities. In its implementation, each step of the cold-region groundwater level truth reconstruction method based on dual-latency space decoupling can be completed through integrated logic circuits in the hardware or instructions in software within processor 10. Processor 10 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0028] The memory 11 may include high-speed random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage.

[0029] Bus 12 can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. Figure 1 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus 12 or one type of bus 12.

[0030] The memory 11 is used to store programs, such as the program corresponding to the cold region groundwater level truth reconstruction device based on dual-space decoupling. The cold region groundwater level truth reconstruction device based on dual-space decoupling includes at least one software functional module that can be stored in the memory 11 in the form of software or firmware or embedded in the operating system (OS) of the electronic device. After receiving the execution instruction, the processor 10 executes the program to implement the cold region groundwater level truth reconstruction method based on dual-space decoupling.

[0031] The electronic device provided in this embodiment of the invention may further include a communication interface 13. The communication interface 13 is connected to the processor 10 via a bus.

[0032] It should be understood that, Figure 1 The structure shown is only a partial schematic diagram of the electronic device; the electronic device may also include components that are larger than... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown. Figure 1 The components shown can be implemented using hardware, software, or a combination thereof.

[0033] The present invention provides a method for reconstructing the true groundwater level in cold regions based on dual-potential space decoupling, which can be applied, but is not limited to, to... Figure 1 For the specific process of the electronic devices shown, please refer to [link / reference]. Figure 2 The groundwater level true value reconstruction method in cold regions based on dual-potential space decoupling includes: S10 to S70, which are described in detail below.

[0034] S10: Obtain the basic observation sequence of the cold region monitoring wells during the observation period.

[0035] The basic observation sequence includes any one or more of the following: original groundwater level sequence, well pressure sequence, temperature observation sequence, and displacement state observation sequence.

[0036] Original groundwater level sequence This represents the groundwater level observation value output by the monitoring equipment deployed in the monitoring well. If the monitoring equipment directly outputs the water level depth, it is first converted to the water level elevation under a unified benchmark.

[0037] Well pressure sequence This indicates a reference level for groundwater level based on hydrostatic pressure.

[0038] Temperature observation sequence It indicates the frozen, thawed, or stable state of the environment where the monitoring well is located, and can be an air temperature sequence, a shallow soil temperature sequence, or a combination of both.

[0039] Displacement state observation sequence This indicates the state change of the well body, wellhead, or water level measurement sensor (used to acquire raw groundwater level sequences) deployed in the monitoring well relative to a reference elevation. It can be one or more of the following: cable length change, cable tension change, wellhead tilt angle change, inertial measurement unit attitude change, wellhead displacement, or surface displacement.

[0040] Each basic observation sequence is preprocessed to obtain a unified time series sample set: Preprocessing includes any one or more of the following: time series unification processing, outlier removal processing, missing test completion processing, and standardization processing. The unified time series sample set is represented as:

[0041] in, This represents the original groundwater level at the k-th sampling time in the preprocessed original groundwater level sequence. This represents the well pressure at the k-th sampling time in the preprocessed well pressure sequence. This represents the freeze-thaw state at the k-th sampling time in the preprocessed temperature observation sequence. This represents the displacement state at the k-th sampling time in the preprocessed displacement state observation sequence. This represents the k-th sampling time, where K is the sampling time number and N is the total number of sampling points; Time series unification processing is used to convert various basic observation sequences to a unified time base and according to a preset time step. A unified time series sample is formed. The unified sampling time is represented as follows:

[0042] in, Indicates the starting sampling time (i.e., the starting time of observation). Indicates the sampling time step.

[0043] Missing data completion processing is used to ensure that different observation sequences have usable data at the same sampling time; outlier removal processing is used to remove isolated observations that obviously do not conform to the physical conditions on site; standardization processing is used to eliminate dimensional differences between different types of observations.

[0044] S20, determine the freeze-thaw state characteristics at the k-th sampling time based on the temperature observation sequence, wherein the freeze-thaw state characteristics include freezing increment and thawing increment.

[0045] S30 extracts the displacement state features and displacement change rate at the k-th sampling time from the displacement state observation sequence.

[0046] S40, construct the freeze-thaw stage label at the k-th sampling time based on the freeze increment, thaw increment, and displacement change rate at the k-th sampling time; S50, when the water level deviation shows a continuous shift or changes synchronously with the freeze-thaw stage, the observation distortion trend at the k-th sampling time is determined based on the displacement state characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label at the k-th sampling time.

[0047] Among them, persistent offset refers to the offset across multiple consecutive sampling times. The deviation of the average value from 0 is greater than the set threshold. If the water level deviation follows the same trend as the freezing increment or thawing increment, it is considered to be a synchronous change, indicating that the original groundwater level observation may be affected by changes in the sensor's relative elevation or well condition. The water level deviation is the deviation between the original groundwater level and the hydrostatic reference water level.

[0048] When the water level deviation does not show a continuous shift and does not change synchronously with the freeze-thaw cycle, the original groundwater level can be directly taken as the true water level.

[0049]

[0050] in, This refers to the water level deviation, which is the inconsistency between the original groundwater level and the hydrostatic reference water level. This represents the original groundwater level at the k-th sampling time in the preprocessed original groundwater level sequence. The reference water level for hydrostatic pressure at the k-th sampling time is denoted as .

[0051] The trend of observed distortion is not directly equivalent to the actual observed distortion component. Instead, it is used to represent the possible trend of observed distortion and guides the learning of latent variables of observed distortion in the subsequent dual-latency decoupling network.

[0052] S60, input the target window data into the dual-latency space decoupling network. The dual-latency space decoupling network extracts the real hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time. The target window data includes the original groundwater level, hydrostatic reference water level, freeze-thaw stage label, and observation distortion trend at each sampling time in the target window. The target window is from the (k-ω+1)-th sampling time to the k-th sampling time, where ω represents the window length.

[0053] To address the formation mechanism of groundwater level observation errors in cold regions, the offsets caused by frost heave, thaw settlement, well displacement, and changes in the relative elevation of sensors are treated as structural observation distortions rather than simply as random noise, thus improving the targeting of groundwater level correction in cold regions. An observation distortion proxy sequence is constructed using freeze-thaw state characteristics, displacement state characteristics, and hydrostatic pressure inconsistency characteristics, so that the model has clear physical and environmental guidance information when separating observation distortion components.

[0054] If k-ω+1 is less than 1, then the target window is from the first sampling time to the kth sampling time.

[0055] In this embodiment of the invention, hydrostatic pressure reference water level, freeze-thaw state characteristics, and displacement state characteristics are used. These three intermediate quantities are used to subsequently determine whether there is a physical inconsistency in the water level observation, whether it is in the freeze-thaw disturbance stage, and whether there is relative displacement of the well body or sensor.

[0056] S70. Based on the actual hydrodynamic latent variables and the observed distortion latent variables at the k-th sampling time, determine the reconstructed water level at the k-th sampling time.

[0057] The groundwater level true value reconstruction method in cold regions based on dual-space decoupling provided in this embodiment of the invention expresses the real hydrodynamic changes and the observed distortion changes respectively through a dual-space decoupling network, reducing the aliasing of the two within the model and improving the reliability and interpretability of the groundwater level true value reconstruction results.

[0058] Based on the preceding text, and in addition to the content of S20, this embodiment of the invention also provides an optional implementation method, please refer to the following. S20, determining the freeze-thaw state characteristics at the k-th sampling time based on the temperature observation sequence, includes: S21 and S22, as detailed below.

[0059] S21, Obtain the temperature observation value at the k-th sampling time from the temperature observation sequence, and determine the criterion temperature at the k-th sampling time based on the temperature observation value at the k-th sampling time. .

[0060] Temperature observations include air temperature and / or shallow soil temperature in the environment where the monitoring well is located. Criterion: Temperature The temperature value used to determine whether the well is in a frozen or thawing state. When the temperature observation values ​​include the air temperature and shallow soil temperature of the environment where the monitoring well is located, the criterion temperature is the temperature value obtained by weighting and fusing the air temperature and shallow soil temperature. The temperature observation values ​​include the air temperature or shallow soil temperature of the environment where the monitoring well is located, and the air temperature or shallow soil temperature of the environment where the monitoring well is located can be directly used as the criterion temperature.

[0061] S22, based on the criterion temperature at the k-th sampling time. Freezing reference temperature and sampling time step The freeze-thaw state characteristics at the k-th sampling time are determined.

[0062]

[0063]

[0064] in, This represents the freeze increment at the k-th sampling time, i.e., the degree of freeze triggering at the k-th sampling time. This represents the melting increment at the k-th sampling time, i.e., the degree of melting triggering at the k-th sampling time.

[0065] Based on the preceding text, and in addition to the content in S30, this embodiment of the invention also provides an optional implementation method, please refer to the following text. S30, extracting the displacement state features and displacement change rate at the k-th sampling time from the displacement state observation sequence, includes: S31 and S32, which are specifically described below.

[0066] S31, extract the displacement state features at the k-th sampling time from the displacement state observation sequence.

[0067] The formula for the displacement state characteristics is:

[0068] in, This represents the displacement state characteristics at the k-th sampling time. This is the displacement feature extraction function.

[0069] The displacement state at the k-th sampling time in the displacement state observation sequence When it is a single observation, The standardized mapping function; the displacement state at the k-th sampling time in the displacement state observation sequence. When multiple observations are included, This is a weighted fusion function or feature mapping function.

[0070] S32, Based on the displacement state characteristics at the kth sampling time and the displacement state characteristics at the (k-1)th sampling time, determine the displacement change rate at the kth sampling time.

[0071]

[0072] in, This represents the rate of change of displacement at the k-th sampling time, i.e., the intensity of the change in displacement state. This represents the displacement state characteristics at the (k-1)th sampling time.

[0073] Optionally, the formula for the freeze-thaw stage label is:

[0074] in, This represents the freeze-thaw stage label at the k-th sampling time. This indicates the freezing of the disturbance phase. This indicates the melting and disturbance stage. Indicates a stable phase. Indicates a transitional phase. This represents the freeze increment at the k-th sampling time. This represents the melting increment at the k-th sampling time. This represents the rate of change of displacement at the k-th sampling time. Indicates the threshold for freezing incremental values. Indicates the melting increment threshold. This represents the threshold for the rate of change of displacement.

[0075] The freezing disturbance stage indicates that the monitoring well and the surrounding medium may be in a period of enhanced frost heave; the thawing disturbance stage indicates that the monitoring well and the surrounding medium may be in a period of enhanced thawing or displacement release; the stable stage indicates that the intensity of freeze-thaw triggering and displacement changes are relatively weak; the transition stage indicates that the freezing, thawing or displacement changes have not yet met the stability criteria.

[0076] The thresholds for freezing increment, melting increment, and displacement change rate can be determined based on statistical quantiles of historical observation data, field experience values, or manually labeled samples.

[0077] In an optional implementation, the freeze-thaw stage labels can be further subdivided into the initial freezing period, stable freezing period, initial thawing period, thawing and settling recovery period, and freeze-free stable period.

[0078] Distortion in groundwater level observations in cold regions primarily stems from changes in the wellbore, wellhead, or sensor elevation relative to a reference elevation during freeze-thaw cycles. When the wellbore experiences frost heave and rise or thaw subsidence and fall, the sensor's installation position may change accordingly; similarly, when the wellhead tilts or the cable condition changes, the water level output by the monitoring equipment may also shift. These shifts differ from actual groundwater dynamic changes but are superimposed on the original groundwater level observation sequence. Therefore, embodiments of this invention can construct a proxy sequence for observation distortion, in which... The observation distortion trend at the k-th sampling time represents the observation distortion trend at the k-th sampling time in the original groundwater level observation sequence.

[0079] Building upon the preceding text, this embodiment of the invention also provides an optional implementation method for the content in S50, as detailed below. Based on the displacement state characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label at the k-th sampling time, the observation distortion change trend at the k-th sampling time is determined, including S51 and S52, which are specifically described below.

[0080] S51, determine the stage gating weight at the kth sampling time based on the freeze-thaw stage label at the kth sampling time.

[0081] The formula for stage gating weights is: ; in, This represents the stage gating weight corresponding to the i-th freeze-thaw stage label. This represents the total number of label types for the freeze-thaw stage. The stage gating weights are used to reflect the differences in the importance of different distortion sources under different freeze-thaw stages. For example, in the freezing disturbance stage, displacement state features and freezing trigger features have a higher weight in terms of their impact on observation distortion; in the thawing disturbance stage, displacement changes, thawing trigger features, and hydrostatic pressure inconsistency features have a higher weight in terms of their impact on observation distortion; in the stable stage, the weights of various distortion proxy features decrease or remain stable.

[0082] S52, based on the stage gating weight at the kth sampling time, perform weighted fusion calculation on the displacement state characteristics, water level deviation, freezing increment, and melting increment at the kth sampling time to determine the observation distortion change trend at the kth sampling time.

[0083] Proxy feature set Each symbol is multiplied by its corresponding stage gating weight, and then summed to determine the trend of observation distortion at the k-th sampling time.

[0084] Based on the foregoing, this invention also provides an optional implementation method for determining the hydrostatic pressure reference water level, which is described below.

[0085] Based on the well pressure at the kth sampling time and the atmospheric pressure at the kth sampling time, determine the well gauge pressure at the kth sampling time (after atmospheric pressure correction).

[0086] Well pressure sequence The well pressure is corrected for atmospheric pressure to obtain the well gauge pressure:

[0087] in, The gauge pressure inside the well at the k-th sampling time is corrected for atmospheric pressure. Let K be the well pressure at the k-th sampling time. Let be the atmospheric pressure at the k-th sampling time.

[0088] Based on the well gauge pressure at the kth sampling time, the water density at the kth sampling time, the elevation correction constant, and the gravitational acceleration, the hydrostatic reference water level at the kth sampling time is determined.

[0089]

[0090] in, Let be the hydrostatic reference water level at the k-th sampling time, and let its water level benchmark be the original groundwater level sequence. The water level benchmark is consistent with the elevation; Let the density of the water be the density at the k-th sampling time. It is the acceleration due to gravity; This represents the elevation correction constant, which is the correction constant for the sensor installation elevation or reference elevation.

[0091] When acquiring water temperature or conductivity data from the monitoring well, the water density at the k-th sampling time... Corrections can be made based on standard water density, water temperature, or conductivity; however, if water temperature or conductivity data is unavailable, Take the constant corresponding to the standard water density.

[0092] It should be noted that the hydrostatic pressure reference water level It is not used directly as the final true value, but rather as a reference quantity to determine whether the original groundwater level observation value satisfies the pressure-water level physical relationship.

[0093] Optionally, the dual-submersible space decoupling network includes one or more of the following: a shared temporal coding module, a real hydrodynamic submersible space coding module, an observation distortion submersible space coding module, a real water level decoding module, and a distortion component decoding module. The shared temporal coding module can employ one of the following: a recurrent neural network, a gated recurrent unit network, a long short-term memory network, a temporal convolutional network, or a Transformer encoder; the real hydrodynamic submersible space coding module and the observation distortion submersible space coding module are implemented using independent coding branches.

[0094] The target window data is represented as follows:

[0095] in, This represents the original groundwater level at the j-th sampling time in the preprocessed original groundwater level sequence. Let the hydrostatic reference water level be the value at the j-th sampling time. This represents the freeze-thaw stage label at the j-th sampling time. This represents the trend of observation distortion at the j-th sampling time.

[0096] Building upon the preceding text, this embodiment of the invention also provides an optional implementation method for the content in S60, as detailed below. The dual-latency spatial decoupling network extracts the true hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time, including S61, S62, and S63, which are specifically described below.

[0097] S61 extracts common temporal representations from the target window data through a shared temporal coding module.

[0098] The formula for representing common time series is:

[0099] in, This represents the common temporal sequence representation at the k-th sampling time. To share the timing coding function, This represents the target window data.

[0100] S62 uses the real hydrodynamic latent space coding module to analyze and process the common time series representation to obtain the real hydrodynamic latent variables at the k-th sampling time.

[0101] The formula for the actual hydrodynamic latent variables is:

[0102] in, Let K be the true hydrodynamic latent variable at the k-th sampling time. For real hydrodynamic submersible space coding module, This represents the common temporal sequence representation at the k-th sampling time. This represents the hydrostatic pressure reference water level sequence within the target window.

[0103] S63 uses the observation distortion latent space coding module to analyze and process the common time series representation to obtain the observation distortion latent variable at the kth sampling time.

[0104]

[0105] in, Let be the latent variable representing the observation distortion at the k-th sampling time. For the latent space encoding function of observation distortion, This represents the observation distortion proxy sequence within the target window.

[0106] Real hydrodynamic latent variables Used to express the actual hydrodynamic changes in groundwater system recharge, runoff, and discharge; observe distorted latent variables. This is used to represent unrealistic hydrodynamic changes caused by frost heave, thaw settlement, or relative displacement of the well body or sensors. By encoding in two separate latent spaces, the aliasing between real water level changes and observed distortions can be reduced within the model.

[0107] Based on the preceding text, regarding the content of S70, this embodiment of the invention also provides an optional implementation method, please refer to the following. S70, determining the reconstructed water level at the k-th sampling time based on the actual hydrodynamic latent variables and the observed distortion latent variables at the k-th sampling time, includes: S71, S72, and S73, which are specifically described below.

[0108] S71 uses the real water level decoding module to decode the real hydrodynamic latent variables at the kth sampling time to obtain the groundwater level decoding result.

[0109] The formula for decoding the groundwater level is:

[0110] in, This represents the decoding result of the groundwater level at the k-th sampling time. This represents the true hydrodynamic latent variable at the k-th sampling time. This indicates the actual water level decoding module.

[0111] S72, the latent variable of the observation distortion at the k-th sampling time is decoded by the distortion component decoding module to obtain the estimation result of the observation distortion component.

[0112] The formula for estimating the observed distortion components is:

[0113] in, This represents the estimation result of the observation distortion component at the k-th sampling time. Denotes the latent variable representing the observation distortion at the k-th sampling time. This indicates the distortion component decoding module.

[0114] S73, based on the estimation results of the observation distortion components at k sampling times and the decoding results of the groundwater level, the water level value is reconstructed to obtain the reconstructed water level at the kth sampling time.

[0115] The reconstructed water level is represented as follows: ; Through the above reconstruction relationship, the model simultaneously satisfies two objectives during training: on the one hand, the output of the true water level decoding module is as close as possible to the actual changes in groundwater level; on the other hand, the output of the distortion component decoding module can explain the offset components in the original groundwater level observations caused by factors such as freeze-thaw displacement.

[0116] In one alternative implementation, Xu Yongtao jointly trains the dual-latency space decoupling network. The joint loss function is:

[0117] in, To reconstruct the loss, For latent space decoupling loss, For agent consistency loss, For hydrostatic pressure uniformity loss, For manual calibration of anchor point loss; , , , These are non-negative weighting coefficients.

[0118] The reconstruction loss is used to constrain the sum of the groundwater level true value reconstruction result and the observation distortion component estimation result to restore the original groundwater level observation value.

[0119]

[0120] This loss ensures that the model will not arbitrarily generate true sequences or distorted components that deviate from the original observation data.

[0121] Latent space decoupling loss is used to reduce the correlation between real hydrodynamic latent variables and observed distortion latent variables, thereby reducing their overlap.

[0122] Arrange the real hydrodynamic latent variables at all sampling times into a matrix Arrange the latent variables of observation distortion at all sampling times into a matrix. Taking the mean of both, we get... and The latent space decoupling loss is:

[0123] in, This represents the Frobenius norm.

[0124] Through this loss, the real hydrodynamic potential space and the observation distortion potential space are constrained to be as independent as possible in terms of statistical correlation, so that the real groundwater level change and the observation distortion change can be expressed separately.

[0125] Observation Distortion Surrogate Sequence It represents the trend of observed distortion, rather than the precise magnitude of distortion. Therefore, this invention employs a consistency constraint on the trend of change to ensure that the estimation results of the observed distortion components are consistent with the observed distortion surrogate sequence in the direction of temporal change.

[0126] The first difference of the estimation results of the observation distortion component is defined as follows:

[0127] The first difference of the observed distortion surrogate sequence is defined as follows:

[0128] The proxy consistency loss is:

[0129] in, This is a scaling factor used to match the dimensional or amplitude differences between the observed distortion surrogate sequence and the observed distortion component.

[0130] This loss causes the distortion component decoding module to prioritize learning changes consistent with the freeze-thaw displacement distortion trend, rather than misinterpreting real groundwater dynamic changes as observational distortion.

[0131] The hydrostatic consistency loss is used to ensure that the reconstructed groundwater level remains physically consistent with the hydrostatic reference level obtained from well pressure conversion.

[0132] in, The reliability weight of the hydrostatic pressure reference water level has the following value range:

[0133] When pressure observation is stable and sensor status is reliable Take the larger value; when pressure observation is abnormal or sensor status is unreliable, Take the smaller value.

[0134] This loss is used to prevent the reconstruction results of the true groundwater level from deviating from the basic hydrostatic pressure relationship.

[0135] When there is a manually measured water level, the manually measured water level is used as a sparse true value anchor point to constrain the groundwater level true value reconstruction result.

[0136] Manually measured water levels are recorded as follows:

[0137] The manual calibration time mask is recorded as:

[0138] This represents the set of times when water levels were manually measured.

[0139] The loss of manually calibrated anchor points is:

[0140] in, To prevent constants with a denominator of zero.

[0141] When there is no manual water level calibration, let:

[0142] This loss is used to provide a reliable anchor point for the model when there is limited manually calibrated data.

[0143] By jointly training the model with reconstruction constraints, latent space decoupling constraints, proxy consistency constraints, hydrostatic pressure consistency constraints, and manual calibration anchor point constraints, continuous groundwater level true value reconstruction results can be obtained even with limited manual calibration data.

[0144] In one optional implementation, after the model training is complete, the basic observation sequence for the time period to be processed is input into the trained dual-potential space decoupling network to output the groundwater level true value reconstruction sequence:

[0145] And the observed distortion component sequence:

[0146] Among them, the groundwater level true value reconstruction sequence is used to represent the groundwater level change after removing the freeze-thaw displacement distortion; the observation distortion component sequence is used to represent the offset change in the original groundwater level observation value caused by freeze-thaw settlement, well displacement or sensor relative elevation change.

[0147] In an optional implementation, a reconstructed reliability score can also be output based on the observation distortion component, proxy consistency error, hydrostatic consistency error, and manual calibration deviation.

[0148] Define the proxy consistency error as

[0149] Define the hydrostatic pressure consistency error as:

[0150] The manual calibration deviation is defined as:

[0151] The reconstruction credibility score can then be expressed as:

[0152] in, To reconstruct the credibility score, the value range is: to ; , , These are non-negative weighting coefficients.

[0153] when The closer When the value is higher, the reliability of the reconstructed groundwater level at the corresponding time is higher; when... The closer The higher the time, the greater the uncertainty of reconstruction at that time.

[0154] In an optional implementation, an abnormal observation period marker can also be output:

[0155] in, This indicates a high risk of observation distortion at the corresponding time point. This indicates that the anomaly flagging condition was not met at the corresponding time. , , , These are the observation distortion component threshold, the proxy consistency error threshold, the hydrostatic pressure consistency error threshold, and the manual calibration deviation threshold, respectively.

[0156] The abnormal observation period marker is used to indicate that the original groundwater level observation value during that period may be significantly affected by frost heave, thaw settlement, well displacement, or changes in the relative elevation of the sensor. It can be used for subsequent data screening, manual verification, or equipment maintenance.

[0157] The output observation distortion components, reconstruction confidence scores, and abnormal observation period markers can be used for subsequent groundwater level data quality evaluation, abnormal period screening, and monitoring equipment maintenance judgment.

[0158] Please see Figure 3 , Figure 3 The present invention provides a groundwater level truth reconstruction device for cold regions based on dual-space decoupling. Optionally, the groundwater level truth reconstruction device for cold regions based on dual-space decoupling is applied to the electronic equipment described above.

[0159] A groundwater level truth reconstruction device for cold regions based on dual-space decoupling includes: The first processing unit 801 is used to acquire the basic observation sequence of the cold region monitoring well during the observation period. The basic observation sequence includes any one or more of the following: original groundwater level sequence, well pressure sequence, temperature observation sequence, and displacement state observation sequence. The second processing unit 802 is used to determine the freeze-thaw state characteristics at the kth sampling time based on the temperature observation sequence, wherein the freeze-thaw state characteristics include freezing increment and thawing increment. The second processing unit 802 is also used to extract the displacement state features and displacement change rate at the kth sampling time from the displacement state observation sequence; The second processing unit 802 is also used to construct a freeze-thaw stage label for the k-th sampling time based on the freeze increment, thaw increment, and displacement change rate at the k-th sampling time. The second processing unit 802 is also used to determine the observation distortion trend at the k-th sampling time based on the displacement state characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label when the water level deviation shows a continuous shift or changes synchronously with the freeze-thaw stage. The second processing unit 802 is also used to input the target window data into the dual-potential space decoupling network. The dual-potential space decoupling network extracts the real hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time. The target window data includes the original groundwater level, hydrostatic reference water level, freeze-thaw stage label and observation distortion change trend at each sampling time in the target window. The target window is from the (k-ω+1)-th sampling time to the k-th sampling time, where ω represents the window length. The second processing unit 802 is also used to determine the reconstructed water level at the k-th sampling time based on the real hydrodynamic latent variables and the observed distortion latent variables at the k-th sampling time.

[0160] It should be noted that the groundwater level truth reconstruction device for cold regions based on dual-submersible space decoupling provided in this embodiment can execute the method flow shown in the above-described method flow embodiment to achieve the corresponding technical effects. For the sake of brevity, any parts not mentioned in this embodiment can be referred to the corresponding content in the above-described embodiments.

[0161] This invention also provides a storage medium storing computer instructions and programs. When read and executed, these instructions and programs perform the above-described method for reconstructing the true groundwater level in cold regions based on dual-space decoupling. The storage medium may include memory, flash memory, registers, or a combination thereof.

[0162] The following provides an electronic device, which may be a server device, a computer device, or a mobile phone device, such as... Figure 1 As shown, the above-described method for reconstructing the true value of groundwater level in cold regions based on dual-space decoupling can be implemented. Specifically, the electronic device includes: a processor 10, a memory 11, and a bus 12. The processor 10 can be a CPU. The memory 11 is used to store one or more programs. When one or more programs are executed by the processor 10, the method for reconstructing the true value of groundwater level in cold regions based on dual-space decoupling described in the above embodiment is executed.

[0163] In summary, the present invention provides a method and apparatus for reconstructing the true groundwater level in cold regions based on dual-latency space decoupling. When the water level deviation exhibits continuous shifts or synchronous changes with the freeze-thaw cycle, the observation distortion trend at the k-th sampling time is determined based on the displacement state characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label at the k-th sampling time. The target window data is input into a dual-latency space decoupling network, which extracts the true hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time. Based on these variables, the reconstructed water level at the k-th sampling time is determined. By using a dual-latency space decoupling network to separately express the true hydrodynamic changes and observation distortion changes, the aliasing of these two variables within the model is reduced, improving the reliability and interpretability of the groundwater level true reconstruction results.

[0164] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0165] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for reconstructing the true groundwater level in cold regions based on dual-potential space decoupling, characterized in that, The method includes: Acquire the basic observation sequence of the monitoring well in the cold region during the observation period. The basic observation sequence includes any one or more of the following: original groundwater level sequence, well pressure sequence, temperature observation sequence, and displacement state observation sequence. The freeze-thaw state characteristics at the k-th sampling time are determined based on the temperature observation sequence, where the freeze-thaw state characteristics include freezing increment and thawing increment; The displacement state characteristics and displacement change rate at the k-th sampling time are extracted from the displacement state observation sequence; Based on the freezing increment, melting increment, and displacement change rate at the k-th sampling time, construct the freeze-thaw stage label at the k-th sampling time; When the water level deviation shows a continuous shift or changes synchronously with the freeze-thaw stage, the trend of observation distortion at the k-th sampling time is determined based on the displacement state characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label at the k-th sampling time. The target window data is input into the dual-potential space decoupling network. The dual-potential space decoupling network extracts the real hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time. The target window data includes the original groundwater level, hydrostatic reference water level, freeze-thaw stage label, and observation distortion trend at each sampling time in the target window. The target window is from the (k-ω+1)-th sampling time to the k-th sampling time, where ω represents the window length. The reconstructed water level at the k-th sampling time is determined based on the actual hydrodynamic latent variables and the observed distortion latent variables at the k-th sampling time.

2. The method for reconstructing the true groundwater level in cold regions based on dual-submersible space decoupling as described in claim 1, characterized in that, The determination of the freeze-thaw state characteristics at the k-th sampling time based on the temperature observation sequence includes: The temperature observation value at the k-th sampling time is obtained from the temperature observation sequence. Based on the temperature observation value at the k-th sampling time, the criterion temperature at the k-th sampling time is determined. ; Based on the criterion temperature at the k-th sampling time Freezing reference temperature and sampling time step The freeze-thaw state characteristics at the k-th sampling time are determined.

3. The method for reconstructing the true groundwater level in cold regions based on dual-submersible space decoupling as described in claim 1, characterized in that, The extraction of displacement state features and displacement change rate at the k-th sampling time from the displacement state observation sequence includes: Extracting displacement state features at the k-th sampling time from the displacement state observation sequence: Based on the displacement state characteristics at the k-th sampling time and the displacement state characteristics at the (k-1)-th sampling time, the displacement change rate at the k-th sampling time is determined.

4. The method for reconstructing the true groundwater level in cold regions based on dual-submersible space decoupling as described in claim 1, characterized in that, The formula for the freeze-thaw stage label is: in, This represents the freeze-thaw stage label at the k-th sampling time. This indicates the freezing of the disturbance phase. This indicates the melting and disturbance stage. Indicates a stable phase. Indicates a transitional phase. This represents the freeze increment at the k-th sampling time. This represents the melting increment at the k-th sampling time. This represents the rate of change of displacement at the k-th sampling time. Indicates the threshold for freezing incremental values. Indicates the melting increment threshold. This represents the threshold for the rate of change of displacement.

5. The method for reconstructing the true groundwater level in cold regions based on dual-submersible space decoupling as described in claim 1, characterized in that, The process of determining the observation distortion trend at the k-th sampling time based on the displacement state characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label includes: Based on the freeze-thaw stage label at the kth sampling time, determine the stage gating weight at the kth sampling time; Based on the stage gating weight at the k-th sampling time, the displacement state characteristics, water level deviation, freezing increment, and melting increment at the k-th sampling time are weighted and fused to determine the observation distortion trend at the k-th sampling time.

6. The method for reconstructing the true groundwater level in cold regions based on dual-submersible space decoupling as described in claim 1, characterized in that, The dual-latency spatial decoupling network extracts the true hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time, including: Common temporal representations are extracted from the target window data using a shared temporal coding module; The common temporal representation is analyzed and processed by the real hydrodynamic latent space coding module to obtain the real hydrodynamic latent variables at the k-th sampling time. The common temporal representation is analyzed and processed by the observation distortion latent space coding module to obtain the observation distortion latent variable at the k-th sampling time.

7. The method for reconstructing the true groundwater level in cold regions based on dual-submersible space decoupling as described in claim 1, characterized in that, The process of determining the reconstructed water level at the k-th sampling time based on the actual hydrodynamic latent variables and the observed distortion latent variables at the k-th sampling time includes: The groundwater level decoding result is obtained by decoding the real hydrodynamic latent variables at the k-th sampling time through the real water level decoding module. The latent variable of observation distortion at the k-th sampling time is decoded by the distortion component decoding module to obtain the estimation result of the observation distortion component. The water level value is reconstructed based on the estimation results of the observation distortion components at k sampling times and the decoding results of the groundwater level, so as to obtain the reconstructed water level at the kth sampling time.

8. A device for reconstructing the true value of groundwater level in cold regions based on dual-submersible space decoupling, characterized in that, The device includes: The first processing unit is used to acquire the basic observation sequence of the cold region monitoring well during the observation period. The basic observation sequence includes any one or more of the following: original groundwater level sequence, well pressure sequence, temperature observation sequence, and displacement state observation sequence. The second processing unit is used to determine the freeze-thaw state characteristics at the k-th sampling time based on the temperature observation sequence, wherein the freeze-thaw state characteristics include freezing increment and thawing increment; The second processing unit is also used to extract the displacement state features and displacement change rate at the kth sampling time from the displacement state observation sequence; The second processing unit is also used to construct a freeze-thaw stage label for the k-th sampling time based on the freeze increment, thaw increment, and displacement change rate at the k-th sampling time. The second processing unit is also used to determine the observation distortion trend at the k-th sampling time based on the displacement state characteristics, water level deviation, freezing increment, thawing increment, and freeze-thaw stage label when the water level deviation shows a continuous shift or changes synchronously with the freeze-thaw stage. The second processing unit is also used to input the target window data into the dual-potential space decoupling network. The dual-potential space decoupling network extracts the real hydrodynamic latent variables and observation distortion latent variables at the k-th sampling time. The target window data includes the original groundwater level, hydrostatic reference water level, freeze-thaw stage label and observation distortion change trend at each sampling time in the target window. The target window is from the (k-ω+1)-th sampling time to the k-th sampling time, where ω represents the window length. The second processing unit is also used to determine the reconstructed water level at the k-th sampling time based on the actual hydrodynamic latent variables and the observed distortion latent variables at the k-th sampling time.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method as described in any one of claims 1-7.

10. An electronic device, characterized in that, include: Processor and memory, the memory being used to store one or more programs; When the one or more programs are executed by the processor, the method as described in any one of claims 1-7 is implemented.

Images