A mine groundwater multivariate dynamic monitoring platform

Abstract

This invention relates to the field of pollution monitoring technology, specifically to a multivariate dynamic monitoring platform for groundwater in mining areas. This platform includes a data acquisition module for acquiring sample environmental hydrological disturbance data sequences and sample hydrological and water quality data sequences; an evaluation value acquisition module for acquiring complex and sensitive evaluation values ​​of environmental hydrological disturbance responses corresponding to each hydrological and water quality type; and a gain compensation and decision tree model construction module for compensating the traditional information gain of the corresponding hydrological and water quality type based on the complex and sensitive evaluation values ​​of environmental hydrological disturbance responses, obtaining a target information gain, and using the target information gain as a node splitting criterion to conduct supervised learning training on the sample hydrological and water quality data combination to generate a binary classification decision tree model for groundwater pollution identification and monitoring in mining areas. Furthermore, this invention can improve the accuracy of the decision tree model for pollution identification and monitoring.

Description

Technical Field

[0001] This invention relates to the field of pollution monitoring technology, specifically to a multivariate dynamic monitoring platform for groundwater in mining areas. Background Technology

[0002] The groundwater system in mining areas is an important carrier for maintaining regional ecological balance and production safety. Its water quality dynamics are directly related to the safety of the surrounding environment, the health of residents, and the sustainability of mining operations. Since mining production activities can easily introduce pollutants such as heavy metals and acidic wastewater, and these pollutants can migrate and spread through groundwater runoff, it is usually necessary to monitor groundwater pollution in mining areas during production activities in order to ensure the ecological safety of mining areas and promote the green and sustainable development of mining operations.

[0003] Existing technologies typically utilize decision tree models for monitoring and identifying groundwater pollution in mining areas. This involves collecting multi-dimensional hydrological and water quality datasets, including pH, conductivity, and metal ion concentration, as the basic input for model training. During model construction, the optimal features and segmentation thresholds are recursively selected based on information gain or the Gini coefficient to partition the dataset, ultimately constructing a tree-like classification structure with a clear discrimination path. This enables effective identification of groundwater pollution in mining areas. The core of existing decision tree model construction relies on the information gain calculated using traditional methods based on the hydrological and water quality datasets. However, different hydrological and water quality types exhibit structural differences in their responses to different environmental hydrological disturbances, and these differences can lead to… Decision tree models constructed using information gain calculated in the traditional way tend to rely on variables (hydrological and water quality) that are highly resistant to interference but have limited discriminative information when recursively selecting features. This can lead to the neglect of key variables that are highly sensitive, noisy, or have a high response to environmental hydrological disturbances. In other words, the hydrological and water quality, which are weakly resistant to environmental hydrological disturbances, are weakened in feature selection based on information gain calculated using the traditional method. Consequently, decision tree models constructed solely based on information gain calculated in the traditional way have poor pollution identification and monitoring capabilities, as well as poor model robustness and generalization performance, resulting in low accuracy in pollution identification and monitoring. Therefore, how to reconstruct the determination logic of information gain to improve the accuracy of the model in pollution identification and monitoring has become an urgent problem to be solved. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a multivariate dynamic monitoring platform for groundwater in mining areas, the specific technical solution of which is as follows: One embodiment of the present invention provides a multivariate dynamic monitoring platform for groundwater in mining areas, the platform comprising: The data acquisition module is used to acquire sample environmental hydrological disturbance data sequences and sample hydrological and water quality data sequences; The evaluation value acquisition module is used to obtain the collaborative characteristic value between environmental hydrological disturbance types and the driving characteristic value of environmental hydrological disturbance types on hydrological and water quality types based on the correlation between sample environmental hydrological disturbance data sequences and the correlation between sample hydrological and water quality data sequences and sample environmental hydrological disturbance data sequences. Based on the collaborative characteristic value, the driving characteristic value is corrected to obtain the target driving characteristic value of environmental hydrological disturbance type on hydrological and water quality type. Based on the target driving characteristic value, the complex sensitivity evaluation value of environmental hydrological disturbance response corresponding to each hydrological and water quality type is obtained. The gain compensation and decision tree model construction module is used to compensate the traditional information gain of the corresponding hydrological and water quality type based on the complex and sensitive evaluation value of the environmental hydrological disturbance response corresponding to each hydrological and water quality type, to obtain the target information gain, and to use the target information gain as the node splitting criterion to conduct supervised learning training on the combination of sample hydrological and water quality data to generate a binary classification decision tree model for the identification and monitoring of groundwater pollution in mining areas.

[0005] Beneficial effects: This invention includes a data acquisition module for acquiring sample environmental hydrological disturbance data sequences and sample hydrological and water quality data sequences; an evaluation value acquisition module for obtaining collaborative characteristic values ​​between environmental hydrological disturbance types and driving characteristic values ​​of environmental hydrological disturbance types on hydrological and water quality types based on the correlation between sample environmental hydrological disturbance data sequences and the correlation between sample hydrological and water quality data sequences and sample environmental hydrological disturbance data sequences, and correcting the driving characteristic values ​​based on the collaborative characteristic values ​​to obtain target driving characteristic values ​​of environmental hydrological disturbance types on hydrological and water quality types, and obtaining complex sensitivity evaluation values ​​of environmental hydrological disturbance response corresponding to each hydrological and water quality type based on the target driving characteristic values; and a gain compensation and decision tree model construction module for compensating the traditional information gain of the corresponding hydrological and water quality type based on the complex sensitivity evaluation values ​​of environmental hydrological disturbance response corresponding to each hydrological and water quality type to obtain the target information gain, and using the target information gain as a node splitting criterion to conduct supervised learning training on the sample hydrological and water quality data combination to generate a binary classification decision tree model for identification and monitoring of groundwater pollution in mining areas. Furthermore, this invention compensates for the traditional information gain of hydrological and water quality types based on the environmental hydrological disturbance response complexity and sensitivity assessment value corresponding to each hydrological and water quality type. This can restore the discrimination weight of hydrological and water quality variables with high environmental sensitivity and high response complexity in the decision tree model, or restore the competitiveness of hydrological and water quality with weak resistance to environmental hydrological disturbances in the selection of split features. In this way, it can improve the accuracy of the decision tree model in pollution identification and monitoring, as well as improve the robustness and generalization performance of the model. Attached Figure Description

[0006] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0007] Figure 1 This is a structural block diagram of a multivariate dynamic monitoring platform for groundwater in mining areas according to the present invention. Detailed Implementation

[0008] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the protection scope of the embodiments of the present invention.

[0009] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art.

[0010] This embodiment provides a multivariate dynamic monitoring platform for groundwater in mining areas, detailed as follows: like Figure 1 As shown in the figure, this embodiment provides a multivariate dynamic monitoring platform for groundwater in mining areas, comprising: Data acquisition module 01 is used to acquire sample environmental hydrological disturbance data sequences and sample hydrological and water quality data sequences.

[0011] Because different hydrological and water quality types exhibit structural differences in their responses to various environmental hydrological disturbances—for example, pH values ​​are easily affected by short-term disturbances from meteorological and hydrological conditions such as heavy rainfall infiltration and surface runoff erosion, but are not significantly affected by aquifer structure—while the concentrations of some variable-valence metal ions (such as iron and manganese) are both controlled by aquifer structure (i.e., redox conditions) and easily affected by rainfall infiltration, making them sensitive variables with complex control mechanisms. This difference can lead to problems in the recursive feature selection of decision tree models built based on information gain calculated using traditional methods. Relying on variables with strong resistance to interference but limited discriminative information (hydrological and water quality) while ignoring key variables with high sensitivity, high noise, or high response complexity, or variables with weak resistance to environmental hydrological disturbances, can weaken the feature selection of hydrological and water quality data sensitive to environmental hydrological disturbances and with complex responses, based on the information gain calculated using traditional information gain calculation methods. Consequently, decision tree models built using information gain calculated in traditional ways have poor pollution identification and monitoring capabilities, as well as poor robustness and generalization performance, making them prone to errors in pollution identification. The accuracy of monitoring is relatively low. To improve the accuracy of the constructed model in pollution identification and monitoring, this embodiment will subsequently combine the response specificity of hydrological and water quality data to environmental hydrological disturbances under long-term hydrological cycles, or the driving characteristics of environmental hydrological disturbances on hydrological and water quality data, as well as the synergistic characteristics between environmental hydrological disturbances, to compensate for the information gain of traditional calculations or to reconstruct the determination logic of information gain. This aims to improve the decision tree model's ability to identify and monitor pollution, as well as its robustness and generalization performance. Specifically, this embodiment compensates for the gain of variables with high environmental sensitivity, high noise, and high response complexity, or variables with weak resistance to environmental hydrological disturbances, to restore their discriminative weights in the decision tree model, or their competitiveness in splitting feature selection. This ensures that variables with low environmental sensitivity, low noise, or high stability, or variables with strong resistance to environmental hydrological disturbances, maintain the information gain of traditional calculations as much as possible. This allows the constructed model to more robustly balance the discriminative power and anti-interference ability of features, thereby improving the decision tree model's ability to identify and monitor pollution, as well as its robustness and generalization performance.

[0012] Typically, multiple pollution monitoring points are deployed in a mining area to monitor and identify groundwater pollution. In this embodiment, a binary classification decision tree model is trained and constructed for each pollution monitoring point in the same mining area. The pollution monitoring and identification of the pollution monitoring point is achieved based on the binary classification decision tree model corresponding to each pollution monitoring point. Furthermore, the method for training and constructing the corresponding binary classification decision tree model for each pollution monitoring point in this embodiment is the same, and the process of monitoring and identifying pollution at the corresponding pollution monitoring point based on the constructed binary classification decision tree model is also the same. Therefore, for ease of understanding and description, this embodiment will subsequently describe the construction process of the binary classification decision tree model corresponding to any pollution monitoring point in any mining area as an example, and will be referred to as the target mining area and the target pollution monitoring point, respectively. That is, the hydrological and water quality data used in this embodiment are collected by the sensors or acquisition equipment at the location of the target pollution monitoring point, and the environmental hydrological disturbance data used are those corresponding to the target mining area. Furthermore, in this embodiment, the pollution monitoring points are located in key monitoring wells along the main flow path of groundwater in the mining area. The key monitoring wells are selected based on the hydrogeological conditions of the mining area, the distribution of pollution sources, and the location of the main flow channel to represent the trend of groundwater quality changes in the mining area. For example, monitoring wells located in the main flow direction of groundwater in the mining area (i.e., the main flow channel from the recharge area to the discharge area) can be used as key monitoring wells.

[0013] Then, sensors or acquisition devices deployed at the target pollution monitoring point location are used to collect various types of hydrological and water quality data at different monitoring times. All hydrological and water quality data collected at the same monitoring time at the target pollution monitoring point location are recorded as the hydrological and water quality data for the target pollution monitoring point location at the corresponding monitoring time. The hydrological and water quality data can reflect the groundwater pollution status at the corresponding monitoring point. The types of hydrological and water quality collected at the same monitoring time include, but are not limited to, pH value, conductivity, and heavy metal concentration. The heavy metal concentration types include, but are not limited to, Fe²⁺ concentration, Mn²⁺ concentration, and Cu²⁺ concentration. The sensors or acquisition devices deployed at the target pollution monitoring point location include, but are not limited to, pH sensors, conductivity sensors, and heavy metal ion selective electrode arrays. Furthermore, the acquisition of data such as pH value, conductivity, and heavy metal concentration is a known technology.

[0014] Since training and generating the decision tree model requires sample data, this embodiment first needs to determine the sample hydrological and water quality data used to construct the decision tree model corresponding to the target pollution monitoring point. Furthermore, to ensure the long-term recognition performance of the constructed model, this embodiment uses the result of data preprocessing of all hydrological and water quality data collected from the target pollution monitoring point within the historical long-term hydrological cycle as the sample hydrological and water quality data corresponding to the target pollution monitoring point at the corresponding monitoring time. In specific applications, the historical long-term hydrological cycle needs to be set by the implementer according to the actual situation, but it is required that the historical long-term hydrological cycle must be a complete hydrological cycle including one or more regular and extreme hydrological and meteorological events. A segment, a complete hydrological cycle period or cycle, refers to the time period corresponding to the recharge, runoff, and discharge of the groundwater system in the mining area. It is also the time period corresponding to the complete path of groundwater in the aquifer system from input, migration to output. For example, in this embodiment, 45 days can be selected as the historical long-term hydrological cycle. Conventional hydro-meteorological events refer to natural meteorological-hydrological processes that are periodic, predictable, and whose intensity is within the historical normal range in the hydrological cycle of the mining area. Extreme hydro-meteorological events refer to abnormal, sudden, and high-intensity hydro-meteorological processes that deviate significantly from the historical statistical threshold (usually the 95th or 99th percentile). They are highly destructive, have low probability, and are highly driven. The distinction between conventional and extreme hydro-meteorological events is a well-known technique.

[0015] This embodiment subsequently needs to adjust the information gain when constructing the decision tree based on the specificity of the response of hydrological and water quality data to environmental hydrological disturbances, the driving characteristics of environmental hydrological disturbances on hydrological and water quality data, and the synergistic characteristics between environmental hydrological disturbances. This adjustment is based on the correlation between hydrological and water quality data and environmental hydrological disturbance data. Therefore, after obtaining the sample hydrological and water quality data corresponding to the target pollution monitoring point, this embodiment monitors and collects multiple types of environmental hydrological disturbance data of the target mining area at different monitoring times within the historical long-term hydrological cycle. The results of data preprocessing of all environmental hydrological disturbance data of the target mining area collected at the same monitoring time are recorded as the sample environmental hydrological disturbance data of the target mining area at the corresponding monitoring time. The historical long-term hydrological cycle corresponding to the collected sample environmental hydrological disturbance data is consistent with the historical long-term hydrological cycle corresponding to the collected sample hydrological and water quality data. The types of environmental hydrological disturbances collected in this embodiment include, but are not limited to, rainfall, runoff intensity, soil moisture content in the vadose zone of the mining area, and leaching water volume from the slag heap. Rainfall at any given monitoring time refers to the cumulative depth of rainwater falling into a representative area of ​​the target mining area's surface during the period from the previous monitoring time to the current monitoring time. Rainfall is collected using rain gauges, which are deployed in any representative area of ​​the target mining area, such as along mining roads. Runoff intensity refers to surface runoff volume, which can be collected using runoff sensors deployed along main runoff channels or paths. Soil moisture content in the vadose zone of the mining area refers to the soil water content in the unsaturated soil layer between the surface and the water table. The ratio of soil mass to dry soil mass; soil moisture content in the vadose zone of the mining area can be collected using soil moisture sensors, which are deployed in typical profiles of the vadose zone; leaching water volume in slag heaps refers to the total amount of liquid that dissolves from open-air slag heaps and migrates with water flow under the influence of precipitation or groundwater; leaching water volume in slag heaps can be collected using leaching water volume monitoring instruments, which are deployed downstream of the heaps; and the collection of data such as rainfall, runoff intensity, soil moisture content in the vadose zone of the mining area, and leaching water volume in slag heaps are known technologies.

[0016] In this embodiment, both hydrological and water quality data and environmental hydrological disturbance data adopt a unified spatial reference and acquisition frequency, meaning that hydrological and water quality data and environmental hydrological disturbance data are acquired synchronously. In specific applications, implementers need to set the data acquisition frequency or the time interval between adjacent monitoring times according to the actual situation, but it should be much smaller than the historical long-term hydrological cycle. For example, in this embodiment, the time interval between adjacent monitoring times can be set to 6 hours. The data preprocessing process in this embodiment includes, but is not limited to, missing value imputation, data filtering, etc., and data preprocessing is a well-known technology.

[0017] Then, all hydrological and water quality data of the same type are arranged in chronological order of collection time to obtain the hydrological and water quality data sequence corresponding to each hydrological and water quality type of the target pollution monitoring point. Similarly, all environmental hydrological disturbance data of the same type are arranged in chronological order of collection time to obtain the environmental hydrological disturbance data sequence corresponding to each environmental hydrological disturbance type of the target pollution monitoring point. Each hydrological and water quality type corresponds to one sample hydrological and water quality data sequence, and the data type of all data in the sample hydrological and water quality data sequence corresponding to one hydrological and water quality type is that hydrological and water quality type. The same applies to environmental hydrological disturbances.

[0018] Therefore, this embodiment can obtain the hydrological and water quality data sequences and environmental hydrological disturbance data sequences of each sample corresponding to the target pollution monitoring point through the above process, as well as the sample hydrological and water quality data and sample environmental hydrological disturbance data at different monitoring times in the historical long-term hydrological cycle.

[0019] The evaluation value acquisition module 02 is used to obtain the collaborative characteristic value between environmental hydrological disturbance types and the driving characteristic value of environmental hydrological disturbance types on hydrological and water quality types based on the correlation between sample environmental hydrological disturbance data sequences and the correlation between sample hydrological and water quality data sequences and sample environmental hydrological disturbance data sequences. Based on the collaborative characteristic value, the driving characteristic value is corrected to obtain the target driving characteristic value of environmental hydrological disturbance type on hydrological and water quality type. Based on the target driving characteristic value, the complex sensitivity evaluation value of environmental hydrological disturbance response corresponding to each hydrological and water quality type is obtained.

[0020] Traditional decision tree models are constructed based on traditional information gain as an indicator for feature selection. However, in the context of complex hydrology in mining areas, the response of hydrological and water quality data to environmental hydrological disturbances has structural differences, which weakens the hydrological and water quality data that are sensitive to pollution in feature selection based on traditional information gain. In this embodiment, traditional information gain refers to the information gain calculated by traditional calculation methods when generating decisions. Therefore, after obtaining the sample hydrological and water quality data sequence and the sample environmental hydrological disturbance data sequence, this embodiment constructs a hydrological and water quality-environmental hydrological disturbance response relationship network, an environmental hydrological disturbance-hydrological and water quality driving relationship network, and an environmental hydrological disturbance-environmental hydrological disturbance collaborative relationship network. This assesses the complexity sensitivity of each hydrological and water quality type to environmental hydrological disturbance responses from the perspectives of correlation dimensions and correlation strength. Based on this, the traditional information gain is dynamically adjusted or compensated to restore the discrimination weight of hydrological and water quality with high complexity sensitivity to environmental hydrological disturbance responses in the decision tree model, or in other words, to restore the competitiveness of hydrological and water quality with high complexity sensitivity to environmental hydrological disturbance responses in splitting feature selection. This improves the decision tree model's ability to identify and monitor pollution, as well as its robustness and generalization.

[0021] Therefore, as described above, this embodiment will first analyze and calculate the response characteristics of hydrological and water quality to environmental hydrological disturbances, or analyze and calculate the driving characteristics of environmental hydrological disturbances on hydrological and water quality, as well as the synergistic characteristics between environmental hydrological disturbances. Subsequently, based on the driving characteristics and synergistic characteristics, the complex sensitivity of hydrological and water quality types to environmental hydrological disturbances will be assessed. Furthermore, since the correlation between the sample environmental hydrological disturbance data sequences and the sample hydrological and water quality data sequences obtained above can characterize the driving characteristics of different types of environmental hydrological disturbances on different hydrological and water quality types... The correlation between sample environmental hydrological disturbance data sequences can characterize the synergistic characteristics between different types of environmental hydrological disturbances. Therefore, this embodiment will first obtain the synergistic characteristic values ​​between environmental hydrological disturbance types and the driving characteristic values ​​of environmental hydrological disturbance types on hydrological and water quality types based on the correlation between sample environmental hydrological disturbance data sequences and the correlation between sample hydrological and water quality data sequences and sample environmental hydrological disturbance data sequences. The specific process for obtaining the synergistic characteristic values ​​between environmental hydrological disturbance types and the driving characteristic values ​​of environmental hydrological disturbance types on hydrological and water quality types is as follows: For the i-th hydrological and water quality type and the j-th environmental hydrological disturbance type, the absolute values ​​of the Spearman correlation coefficients between the sample hydrological and water quality data sequence corresponding to the i-th hydrological and water quality type and the sample environmental hydrological disturbance data sequence corresponding to the j-th environmental hydrological disturbance type are calculated at each lag value within a preset lag range. The maximum absolute value among the Spearman correlation coefficients is selected as the driving characteristic value of the j-th environmental hydrological disturbance type on the i-th hydrological and water quality type. Furthermore, the larger the maximum value among the absolute values ​​of the Spearman correlation coefficients or the larger the driving characteristic value of the j-th environmental hydrological disturbance type on the i-th hydrological and water quality type, the stronger the monotonic correlation between the two time series, i.e., the stronger the hydrological driving effect of the j-th environmental hydrological disturbance type on the i-th hydrological and water quality type. The more significant the response of the i-th hydrological and water quality type to the j-th environmental hydrological disturbance type, the smaller the maximum absolute value of the Spearman correlation coefficient, or the smaller the driving characteristic value of the j-th environmental hydrological disturbance type to the i-th hydrological and water quality type, the weaker the monotonic correlation between the two time series. That is, the weaker the hydrological driving effect of the j-th environmental hydrological disturbance type to the i-th hydrological and water quality type, or the weaker the response of the i-th hydrological and water quality type to the j-th environmental hydrological disturbance type. The Spearman correlation coefficient can effectively measure the monotonic relationship between two time series, adapt to the nonlinear and non-normal distribution characteristics of time series data between hydrological and water quality variables and environmental hydrological disturbance factors in the water environment system, and can more realistically reflect the inherent correlation between the two.

[0022] Similarly, for the q-th and j-th types of environmental hydrological disturbances, the maximum absolute value of the Spearman correlation coefficient between the environmental hydrological disturbance data sequence corresponding to the q-th type and the sample environmental hydrological disturbance data sequence corresponding to the j-th type within the preset lag range is the synergistic characteristic value between the q-th and j-th types of environmental hydrological disturbances. Furthermore, the larger the synergistic characteristic value or the larger the maximum absolute value of the Spearman correlation coefficient between the q-th and j-th types of environmental hydrological disturbances, the more significant the synergistic effect or collinearity between them. Conversely, the smaller the synergistic characteristic value or the smaller the maximum absolute value of the Spearman correlation coefficient, the less obvious the synergistic effect or collinearity between the q-th and j-th types of environmental hydrological disturbances.

[0023] In practical applications, implementers can set a preset lag range according to the actual situation, but the lag value is required not to exceed the total number of data in the sample environmental hydrological disturbance data sequence. For example, in this embodiment, -20 to 20 can be used as the preset lag range, with positive numbers indicating lag and negative numbers indicating lead. In addition, due to the natural hydrological transmission lag characteristics of the groundwater system in the mining area, environmental disturbances such as surface rainfall and runoff leaching need to be infiltrated through the vadose zone and transported by the aquifer before they can affect the groundwater quality. The response of water quality variables to environmental disturbances is not instantaneous and synchronous. Therefore, the above-mentioned calculation characteristic value in this embodiment introduces the concept of lead or lag to accurately match the physical laws of groundwater hydrological transmission, truly capture the asynchronous intrinsic driving correlation between hydrological water quality and environmental hydrological disturbances or the asynchronous synergistic relationship between environmental hydrological disturbances, and thus avoid the deviation of the correlation strength calculation from the actual hydrological process due to transmission delay.

[0024] In this embodiment, the Spearman correlation coefficient between the sample hydrological and water quality data sequence corresponding to the i-th hydrological and water quality type and the sample environmental hydrological disturbance data sequence corresponding to the j-th environmental hydrological disturbance type at any lag value within a preset lag range is illustrated. Let the sample hydrological and water quality data sequence corresponding to the i-th hydrological and water quality type be denoted as sequence i, and the sample environmental hydrological disturbance data sequence corresponding to the j-th environmental hydrological disturbance type be denoted as sequence j. If the lag value is 2, then the Spearman correlation coefficient between sequence i and sequence j at a lag value of 2 refers to the Spearman correlation coefficient between sequence j and sequence i when sequence j lags by 2 steps. If the lag value is -2, then the Spearman correlation coefficient between sequence i and sequence j at a lag value of -2 refers to the Spearman correlation coefficient between sequence j and sequence i when sequence j leads by 2 steps. The total number of data points in both sequence i and sequence j is 25. Therefore, the Spearman correlation coefficient between sequence j and sequence i when sequence j lags by 2 steps is the same as the Spearman correlation coefficient between the sequence formed by the 3rd to 25th data points in sequence j and the sequence formed by the 1st to 23rd data points in sequence i. Similarly, the Spearman correlation coefficient between sequence j and sequence i when sequence j leads by 2 steps is the same as the Spearman correlation coefficient between the sequence formed by the 1st to 23rd data points in sequence j and the sequence formed by the 3rd to 25th data points in sequence i. The calculation process for the Spearman correlation coefficients between the environmental hydrological disturbance data sequence corresponding to the qth type of environmental hydrological disturbance and the sample environmental hydrological disturbance data sequence corresponding to the jth type of environmental hydrological disturbance at various lag values ​​within the preset lag range is the same.

[0025] Furthermore, in mining groundwater systems, multiple environmental hydrological disturbance factors often exhibit significant temporal synergy or collinearity, which can lead to the quantified driving characteristic values ​​deviating from the actual situation or becoming overestimated. If the synergy among all environmental hydrological disturbance types is significant, then the calculated driving characteristic values ​​of the environmental hydrological disturbance type on the hydrological and water quality type will contain the driving force of other environmental hydrological disturbance types on the hydrological and water quality type. In other words, there may be signal aliasing in the calculated driving characteristic values ​​of the environmental hydrological disturbance type on the hydrological and water quality type, which will cause the quantified driving characteristic values ​​to deviate from the actual situation or become overestimated. This will affect the subsequent assessment of the hydrological and water quality type based on the driving characteristic values. The complex sensitivity of the environmental hydrological disturbance response obtained is not very reliable. In order to ensure the credibility of the assessment of the complex sensitivity of the environmental hydrological disturbance response to the hydrological water quality type, this embodiment will next modify the driving characteristic value of the environmental hydrological disturbance type on the hydrological water quality type based on the cooperative characteristic value between environmental hydrological disturbance types, so as to obtain the target driving characteristic value of the environmental hydrological disturbance type on the hydrological water quality type. That is, the purpose of the modification is to more accurately quantify the independent driving or true driving force of each environmental hydrological disturbance type on the hydrological water quality type, or in other words, to make the final target driving characteristic value able to truly reflect the independent driving characteristics of a single environmental hydrological disturbance factor on the hydrological water quality variable.

[0026] In this embodiment, the specific process of correcting the driving characteristic value of environmental hydrological disturbance type on hydrological and water quality type based on the cooperative characteristic value between environmental hydrological disturbance types to obtain the target driving characteristic value of environmental hydrological disturbance type on hydrological and water quality type is as follows: First, based on the synergistic characteristic values ​​among different types of environmental hydrological disturbances, correction factors are obtained for the driving characteristic values ​​of each environmental hydrological disturbance type on each hydrological and water quality type. The specific process for obtaining the correction factors is as follows: For the i-th hydrological and water quality type and the j-th environmental hydrological disturbance type, the set of the remaining environmental hydrological disturbance types other than the j-th environmental hydrological disturbance type is denoted as the residual set of the j-th environmental hydrological disturbance type. The maximum synergistic characteristic value among the synergistic characteristic values ​​between the j-th environmental hydrological disturbance type and each environmental hydrological disturbance type in the residual set is obtained and denoted as the driving characteristic value of the j-th environmental hydrological disturbance type on the i-th hydrological and water quality type. The correction factor is used to determine the maximum synergistic characteristic value. This is to capture the synergistic effect strength most closely associated with the j-th type of environmental hydrological disturbance, or to focus on other environmental hydrological disturbance types with the strongest synergy with the j-th type of environmental hydrological disturbance, so as to achieve accurate stripping of signal aliasing and realize the true characterization of the independent driving characteristics of environmental hydrological disturbance factors on hydrological and water quality variables. The larger the correction factor, the stronger the synergy or collinearity between the j-th type of environmental hydrological disturbance and other types of environmental hydrological disturbance, and the greater the degree of correction or adjustment of the driving characteristic value of the j-th type of environmental hydrological disturbance on the i-th type of hydrological and water quality.

[0027] Then, based on the correction factors for the driving characteristic values ​​of each environmental hydrological disturbance type on each hydrological and water quality type, the driving characteristic values ​​of the corresponding environmental hydrological disturbance type on the corresponding hydrological and water quality type are corrected to obtain the target driving characteristic values ​​of each environmental hydrological disturbance type on each hydrological and water quality type. The specific process of obtaining the target driving characteristic value is as follows: For the i-th hydrological and water quality type and the j-th environmental hydrological disturbance type, the product of the result of the inverse mapping of the correction factor of the driving characteristic value of the j-th environmental hydrological disturbance type on the i-th hydrological and water quality type and the driving characteristic value of the j-th environmental hydrological disturbance type on the i-th hydrological and water quality type is denoted as the target driving characteristic value of the j-th environmental hydrological disturbance type on the i-th hydrological and water quality type. Here, the inverse mapping is achieved by subtracting the mapped data from the constant 1. Therefore, the expression for the target driving characteristic value of the j-th environmental hydrological disturbance type on the i-th hydrological and water quality type is:

[0028] in, Let j be the target driving characteristic value of the j-th type of environmental hydrological disturbance on the i-th type of hydrological and water quality. Let be the driving characteristic value of the j-th type of environmental hydrological disturbance on the i-th type of hydrological and water quality. This is a correction factor for the driving characteristic value of the j-th type of environmental hydrological disturbance on the i-th type of hydrological and water quality. And the correction factor... The larger the value, the stronger the synergy or collinearity between the j-th type of environmental hydrological disturbance and other types of environmental hydrological disturbance. The greater the degree of correction or adjustment, that is... The larger, Compared to The smaller, or rather When it is larger, The smaller; conversely When the value is larger or closer to 0, it indicates that there is no significant synergy or collinearity between the j-th type of environmental hydrological disturbance and other types of environmental hydrological disturbance. and The closer the correlation is, the more accurate the correlation strength quantification or driving characteristic quantification is guaranteed in the absence of cooperative interference. Furthermore, the correction mechanism in this embodiment ensures the interpretability and physical rationality of the driving characteristics under the synergistic background of multiple environmental hydrological disturbance factors, providing more accurate data support for subsequent assessment of the complex sensitivity of hydrological and water quality types to environmental hydrological disturbance responses.

[0029] In this embodiment, after obtaining the target driving characteristic values ​​of environmental hydrological disturbance types on hydrological and water quality types, the target driving characteristic values ​​of each environmental hydrological disturbance type on the same hydrological and water quality type are fused and analyzed. Based on the analysis results, the complexity sensitivity of the hydrological and water quality type to environmental hydrological disturbance responses is evaluated. That is, this embodiment will next obtain the environmental hydrological disturbance response complexity sensitivity assessment value corresponding to each hydrological and water quality type based on the target driving characteristic values ​​of each environmental hydrological disturbance type on the same hydrological and water quality type. The environmental hydrological disturbance response complexity sensitivity assessment value is the key basis for adjusting or compensating the traditional information gain when constructing the decision tree. That is, in the information gain calculation of this embodiment, the environmental hydrological disturbance response complexity sensitivity assessment value will be introduced to modulate the traditional information gain. The specific process of obtaining the environmental hydrological disturbance response complexity sensitivity assessment value corresponding to each hydrological and water quality type based on the target driving characteristic values ​​of each environmental hydrological disturbance type on the same hydrological and water quality type is as follows: For the i-th hydrological and water quality type: the set of target driving characteristic values ​​for the i-th hydrological and water quality type for each type of environmental hydrological disturbance is denoted as the target driving characteristic value set corresponding to the i-th hydrological and water quality type; in the target driving characteristic value set corresponding to the i-th hydrological and water quality type, the set of target driving characteristic values ​​greater than the preset driving discrimination threshold is denoted as the screening set; the proportion of the number of target driving characteristic values ​​in the screening set relative to the total amount of data in the target driving characteristic value set corresponding to the i-th hydrological and water quality type is calculated and denoted as the response complexity characterization value; the square of the mean of the screening set is calculated and denoted as the sensitivity characterization value; the product of the response complexity characterization value and the sensitivity characterization value is calculated and denoted as the environmental hydrological disturbance response complexity sensitivity assessment value corresponding to the i-th hydrological and water quality type.

[0030] The specific expression for the complex sensitivity assessment value of the environmental hydrological disturbance response corresponding to the i-th hydrological and water quality type is:

[0031] in, This represents the complex sensitivity assessment value for the environmental hydrological disturbance response corresponding to the i-th hydrological and water quality type. To filter the number of target driving characteristic values ​​in the set, M represents the total number of target driving characteristic values ​​in the set corresponding to the i-th hydrological and water quality type. In response to the complexity representation value, The percentage of target driving characteristic values ​​in the set of target driving characteristic values ​​corresponding to the i-th hydrological and water quality type that are greater than a preset driving discrimination threshold. To filter the k-th target driving characteristic value in the set.

[0032] The larger or The larger the value, the more independent driving forces the i-th hydrological and water quality type is from environmental hydrological disturbances, the more complex the response pattern of the i-th hydrological and water quality type to environmental hydrological disturbances, or the higher the complexity of the response of the i-th hydrological and water quality type to environmental hydrological disturbances, and vice versa. smaller or The smaller the value, the fewer the number of independent environmental hydrological disturbances that drive the i-th hydrological and water quality type, the simpler the response pattern of the i-th hydrological and water quality type to environmental hydrological disturbances, or the lower the response complexity of the i-th hydrological and water quality type to environmental hydrological disturbances. The larger or The larger the value, the stronger the independent driving force of each environmental hydrological disturbance on the i-th hydrological and water quality type; the higher the sensitivity of the i-th hydrological and water quality type to environmental hydrological disturbances; the worse the data stability of the i-th hydrological and water quality type; and the more susceptible it is to noise interference or the weaker its anti-interference ability. smaller or The smaller the value, the weaker the independent driving force of each environmental hydrological disturbance on the i-th hydrological and water quality type, the lower the sensitivity of the i-th hydrological and water quality type to environmental hydrological disturbances, the better the data stability of the i-th hydrological and water quality type, and the less susceptible it is to noise interference or the better its anti-interference ability. Squaring this value is done to non-linearly amplify the contribution of high average correlation strength or high average driving characteristic value, ensuring that subsequent information gain weight adjustments can more accurately identify and compensate for sensitive key features easily masked by noise or more accurately identify and compensate for key variables with high sensitivity and high response complexity. And because... The greater the sum When it is larger, The larger, therefore The larger the value, the higher the response complexity of the i-th hydrological and water quality type to environmental hydrological disturbances, the higher the sensitivity of the i-th hydrological and water quality type to environmental hydrological disturbances, the greater the noise level or the weaker the anti-interference ability of the i-th hydrological and water quality type data, the easier it is to weaken the information gain of the i-th hydrological and water quality type calculated based on the traditional information gain calculation method in feature selection, or the easier it is to underestimate the information gain of the i-th hydrological and water quality type calculated based on the traditional information gain calculation method, and the more necessary it is to perform gain compensation in the future; conversely, the smaller the value, the more likely it is to be underestimated. The smaller the value, the lower the response complexity of the i-th hydrological and water quality type to environmental hydrological disturbances, the lower the sensitivity of the i-th hydrological and water quality type to environmental hydrological disturbances, and the better the stability or anti-interference ability of the i-th hydrological and water quality type data. In this case, the information gain of the i-th hydrological and water quality type calculated based on the traditional information gain calculation method can truly reflect its discrimination ability. Therefore, it is advisable to maintain the information gain calculated based on the traditional information gain calculation method as much as possible. The smaller the value, the less gain compensation will be applied to the i-th hydrological and water quality type. The larger the value, the greater the degree of gain compensation for the i-th hydrological and water quality type. Traditional information gain calculation methods are usually based on entropy and conditional entropy in Shannon's information theory or on the amount of information entropy reduction, which are well-known techniques.

[0033] Since the target driving characteristic value of the j-th type of environmental hydrological disturbance on the i-th type of hydrological and water quality is greater than the preset driving discrimination threshold, it indicates that the j-th type of environmental hydrological disturbance has a substantial independent driving effect on the i-th type of hydrological and water quality. This also indicates that the j-th type of environmental hydrological disturbance is the core environmental hydrological disturbance factor that causes significant noise in the i-th type of hydrological and water quality data, characterizes whether the i-th type of hydrological and water quality belongs to a highly sensitive environment, and has a complex response. Therefore, it can be determined that the j-th type of environmental hydrological disturbance and the i-th type of hydrological and water quality are significantly correlated, or that the j-th type of environmental hydrological disturbance has a substantial independent driving effect on the i-th type of hydrological and water quality. Conversely, when the target driving characteristic value of the j-th type of environmental hydrological disturbance on the i-th type of hydrological and water quality is not greater than the preset driving discrimination threshold, it indicates that the substantial independent driving effect of the j-th type of environmental hydrological disturbance on the i-th type of hydrological and water quality is weak and negligible. This also indicates that the j-th type of environmental hydrological disturbance is not the core environmental hydrological disturbance factor causing significant noise in the i-th type of hydrological and water quality data, and characterizes whether the i-th type of hydrological and water quality belongs to a highly sensitive environment. The data for hydrological and water quality type i has relatively high noise, which indicates whether type i is a core environmental hydrological disturbance factor that is highly sensitive to the environment and has a complex response. Therefore, it can be determined that type j is weakly correlated with type i or that type j does not have a substantial independent driving effect on type i. As can be seen from the above, in this embodiment, when evaluating the complex sensitivity of the environmental hydrological disturbance response corresponding to type i, only environmental hydrological disturbance types that have a substantial independent driving effect on type i should be considered. This is to ensure the accuracy and reference value of the evaluated complex sensitivity of the environmental hydrological disturbance response corresponding to type i. That is, the purpose of screening based on the preset driving discrimination threshold is to accurately screen out environmental hydrological disturbance factors that have a substantial independent disturbance or driving effect on hydrological and water quality variables, and to exclude invalid interference from weakly correlated or weakly driving environmental hydrological disturbance factors. In practical applications, implementers need to set a preset driving discrimination threshold based on the actual situation such as the range of target driving characteristic values. For example, in this embodiment, the preset driving discrimination threshold can be set to 0.6.

[0034] Therefore, this embodiment obtains the environmental hydrological disturbance response complexity and sensitivity assessment value corresponding to the hydrological and water quality type through the above process. Furthermore, in this embodiment, only specific numerical values ​​are involved in all data participating in the target information gain calculation; dimensions are not considered.

[0035] The gain compensation and decision tree model construction module 03 is used to compensate the traditional information gain of the corresponding hydrological and water quality type based on the complex and sensitive evaluation value of the environmental hydrological disturbance response corresponding to each hydrological and water quality type, to obtain the target information gain, and to use the target information gain as the node splitting criterion to conduct supervised learning training on the combination of sample hydrological and water quality data to generate a binary classification decision tree model for the identification and monitoring of groundwater pollution in mining areas.

[0036] After obtaining the environmental hydrological disturbance response complexity sensitivity assessment value corresponding to each hydrological and water quality type, this embodiment compensates for the traditional information gain of the corresponding hydrological and water quality type based on the environmental hydrological disturbance response complexity sensitivity assessment value to obtain the target information gain. Subsequently, the target information gain is used as the node splitting criterion to conduct supervised learning training on the sample hydrological and water quality data combination to generate a binary classification decision tree model for identifying groundwater pollution in mining areas. At this time, the generated model is the binary classification decision tree model corresponding to the target pollution monitoring point. Furthermore, in the process of constructing the binary classification decision tree model for water pollution identification, if the traditional information gain of a certain hydrological and water quality type calculated before node splitting is... If so, then the target information gain after compensating for the traditional information gain of this hydrological and water quality type based on the complex sensitivity assessment value of the environmental hydrological disturbance response corresponding to this hydrological and water quality type is: F represents the complexity and sensitivity assessment value of the environmental hydrological disturbance response corresponding to this hydrological and water quality type. The larger F is, the greater the weakening of the traditional information gain based on this hydrological and water quality type when selecting features. Therefore, more gain compensation is required, or the target information gain is larger. The closer F is to 0, the more accurately the traditional information gain based on this hydrological and water quality type can reflect its ability to distinguish pollution. In this case, the traditional information gain can accurately reflect its ability to distinguish pollution. Therefore, the smaller F is, the closer the target information gain of this hydrological and water quality type should be to its traditional information gain, so as to avoid model misjudgment due to overcompensation and ensure the normal operation of the model in a stable environment. The constant 1 in the target information gain calculation formula is the preset benchmark compensation constant.

[0037] In this embodiment, when constructing the decision tree, only the process of determining or calculating the information gain used for node splitting is changed. In other words, this embodiment only adds a process of compensating for the traditional information gain, without changing other processes. It should be noted that the node splitting in this embodiment is based on the compensated information gain. In constructing the decision tree model, the input sample is a combination of hydrological and water quality data, which is a set of all types of hydrological and water quality data acquired at the same monitoring time. Similarly, the input after model training is a set formed by preprocessing all types of hydrological and water quality data from the target pollution monitoring points collected at the real-time monitoring time. The output of the constructed binary classification decision tree model has only two results: polluted and unpolluted. That is, the input structure during model training is the same as the input structure used after training. In other words, if the target pollution monitoring points collected at the current monitoring time are used... The set of all types of hydrological and water quality data, after preprocessing, is directly input into the trained binary classification decision tree model. The model can quickly complete the binary classification of the target pollution monitoring point at the current monitoring time based on the constructed discrimination path. If the output is "pollution", it indicates that there is groundwater pollution in the mining area at the current target pollution monitoring point; if the output is "no pollution", it indicates that there is no groundwater pollution in the mining area at the current target pollution monitoring point. Since the training and construction process of the model is a well-known technology after knowing the sample data and the method of determining information gain, it will not be described in detail. In this embodiment, for the result that the model judges that there is pollution, the validity of the pollution signal can be verified by combining the real-time mining operation conditions and groundwater runoff patterns in the mining area to eliminate misjudgments caused by equipment errors and accidental disturbances. When it is verified that it is a real pollution signal, the monitoring platform immediately triggers a pollution warning and pushes core information such as the location of the monitoring well where the pollution occurred and the monitoring time to the environmental monitoring management department of the mining area, providing real-time data support for emergency response to pollution in the mining area. In this embodiment, the types and number of hydrological and water quality data input into the model during pollution identification are consistent with the types and number of hydrological and water quality data input during model training and construction.

[0038] Thus, this embodiment completes the identification and monitoring of groundwater pollution in the mining area. Furthermore, the above-mentioned compensation of the traditional information gain of hydrological and water quality types based on the environmental hydrological disturbance response complexity and sensitivity assessment value corresponding to each hydrological and water quality type can restore the discrimination weight of hydrological and water quality variables with high environmental sensitivity and high response complexity in the decision tree model, or restore the competitiveness of hydrological and water quality variables with high environmental sensitivity and high response complexity in the selection of splitting features. This can improve the decision tree model's ability to identify and monitor pollution, as well as the model's robustness and generalization performance.

[0039] In summary, this embodiment includes a data acquisition module for acquiring sample environmental hydrological disturbance data sequences and sample hydrological and water quality data sequences; an evaluation value acquisition module for obtaining collaborative characteristic values ​​between environmental hydrological disturbance types and driving characteristic values ​​of environmental hydrological disturbance types on hydrological and water quality types based on the correlation between sample environmental hydrological disturbance data sequences and the correlation between sample hydrological and water quality data sequences and sample environmental hydrological disturbance data sequences; and a correction of the driving characteristic values ​​based on the collaborative characteristic values ​​to obtain the target driving characteristic values ​​of environmental hydrological disturbance types on hydrological and water quality types, and obtaining the environmental hydrological disturbance response complexity sensitivity evaluation values ​​corresponding to each hydrological and water quality type based on the target driving characteristic values; and a gain compensation and decision tree model construction module for compensating the traditional information gain of the corresponding hydrological and water quality type based on the environmental hydrological disturbance response complexity sensitivity evaluation values ​​corresponding to each hydrological and water quality type to obtain the target information gain, and using the target information gain as a node splitting criterion to conduct supervised learning training on the sample hydrological and water quality data combination to generate a binary classification decision tree model for groundwater pollution identification and monitoring in mining areas. Furthermore, this embodiment compensates for the traditional information gain of hydrological and water quality types based on the environmental hydrological disturbance response complexity and sensitivity assessment value corresponding to each hydrological and water quality type. This can restore the discrimination weight of hydrological and water quality variables with high environmental sensitivity and high response complexity in the decision tree model, or restore the competitiveness of hydrological and water quality with weak resistance to environmental hydrological disturbances in the selection of split features. In this way, it can improve the accuracy of the decision tree model for pollution identification and monitoring, as well as improve the robustness and generalization performance of the model.

[0040] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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. Such 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 this application, and should all be included within the protection scope of this application.

Claims

1. A multivariate dynamic monitoring platform for groundwater in mining areas, characterized in that, The multivariate dynamic monitoring platform for groundwater in the mining area includes: The data acquisition module is used to acquire sample environmental hydrological disturbance data sequences and sample hydrological and water quality data sequences; The evaluation value acquisition module is used to obtain the collaborative characteristic value between environmental hydrological disturbance types and the driving characteristic value of environmental hydrological disturbance types on hydrological and water quality types based on the correlation between sample environmental hydrological disturbance data sequences and the correlation between sample hydrological and water quality data sequences and sample environmental hydrological disturbance data sequences. Based on the collaborative characteristic value, the driving characteristic value is corrected to obtain the target driving characteristic value of environmental hydrological disturbance type on hydrological and water quality type. Based on the target driving characteristic value, the complex sensitivity evaluation value of environmental hydrological disturbance response corresponding to each hydrological and water quality type is obtained. The gain compensation and decision tree model construction module is used to compensate the traditional information gain of the corresponding hydrological and water quality type based on the complex and sensitive evaluation value of the environmental hydrological disturbance response corresponding to each hydrological and water quality type, to obtain the target information gain, and to use the target information gain as the node splitting criterion to conduct supervised learning training on the combination of sample hydrological and water quality data to generate a binary classification decision tree model for the identification and monitoring of groundwater pollution in mining areas.

2. The multivariate dynamic monitoring platform for groundwater in mining areas as described in claim 1, characterized in that, Methods for obtaining driver characteristic values ​​include: For the i-th hydrological and water quality type and the j-th environmental hydrological disturbance type, the maximum value among all absolute values ​​of the Spearman correlation coefficients between the sample hydrological and water quality data sequence corresponding to the i-th hydrological and water quality type and the sample environmental hydrological disturbance data sequence corresponding to the j-th environmental hydrological disturbance type within a preset lag range is taken as the driving characteristic value of the j-th environmental hydrological disturbance type on the i-th hydrological and water quality type.

3. The multivariate dynamic monitoring platform for groundwater in mining areas as described in claim 1, characterized in that, Methods for obtaining target-driven characteristic values ​​include: Based on the synergistic characteristic values ​​among environmental hydrological disturbance types, the correction factors for the driving characteristic values ​​of each environmental hydrological disturbance type on each hydrological and water quality type are obtained. The product of the inverse mapping of the correction factor of each environmental hydrological disturbance type to the driving characteristic value of each hydrological and water quality type and the driving characteristic value of the corresponding environmental hydrological disturbance type to the corresponding hydrological and water quality type is denoted as the target driving characteristic value of the corresponding environmental hydrological disturbance type to the corresponding hydrological and water quality type.

4. The multivariate dynamic monitoring platform for groundwater in mining areas as described in claim 3, characterized in that, The maximum value among the collaborative characteristic values ​​between any environmental hydrological disturbance type and other environmental hydrological disturbance types is the correction factor for the driving characteristic values ​​of the environmental hydrological disturbance type on each hydrological and water quality type.

5. The multivariate dynamic monitoring platform for groundwater in mining areas as described in claim 1, characterized in that, Methods for obtaining complex and sensitive assessment values ​​of environmental hydrological disturbance response include: For any hydrological and water quality type, the set of target driving characteristic values ​​of each environmental hydrological disturbance type to the hydrological and water quality type is denoted as the target driving characteristic value set. The proportion of target driving characteristic values ​​in the target driving characteristic value set that are greater than a preset driving discrimination threshold is denoted as the response complexity characterization value. The square of the mean of the target driving characteristic values ​​in the target driving characteristic value set that are greater than the preset driving discrimination threshold is denoted as the sensitivity characterization value. The product of the response complexity characterization value and the sensitivity characterization value is denoted as the environmental hydrological disturbance response complexity sensitivity assessment value corresponding to the hydrological and water quality type.

6. The multivariate dynamic monitoring platform for groundwater in mining areas as described in claim 1, characterized in that, Methods for obtaining target information gain include: The product of the complex and sensitive assessment value of the environmental hydrological disturbance response corresponding to each hydrological and water quality type, plus the preset benchmark compensation constant, and the traditional information gain of the corresponding hydrological and water quality type is the target information gain of the corresponding hydrological and water quality type.

7. The multivariate dynamic monitoring platform for groundwater in mining areas as described in claim 1, characterized in that, Environmental hydrological disturbance types include rainfall, runoff intensity, soil moisture content in the vadose zone of the mining area, and leaching water volume from the slag heap. Hydrological and water quality types include pH, conductivity, and heavy metal ion concentration.

8. The multivariate dynamic monitoring platform for groundwater in mining areas as described in claim 1, characterized in that, The sample hydrological and water quality data combination is formed by all types of sample hydrological and water quality data acquired at the same time.

9. A multivariate dynamic monitoring platform for groundwater in mining areas as described in claim 1, characterized in that, The method for obtaining the collaborative characteristic value is the same as the method for obtaining the driving characteristic value.

Images