Intelligent discrimination method and apparatus for mine water inrush source, device, and storage medium

Abstract

An intelligent discrimination method and apparatus for a mine water inrush source, a device, and a storage medium. The method comprises: acquiring a water chemical index of a water source sample to be discriminated (S1), wherein the water chemical index comprises a plurality of inorganic indexes and a plurality of organic indexes; and inputting the water chemical index into a preset water source discrimination model to obtain a water source type of said water source sample (S2). In addition to inorganic indexes for discrimination, organic indexes are added, so that the problem of being difficult to accurately discriminate, by means of inorganic water chemical indexes, water samples with similar component is solved, thereby improving the accuracy of water source type discrimination.

Description

Mine water source intelligent discrimination method, device, equipment and storage medium TECHNICAL FIELD

[0001] [According to the rules 26 correction 19.12.2025]The invention belongs to the technical field of mines, and relates to a mine water source intelligent discrimination method, device, equipment and storage medium. BACKGROUND

[0002] [According to the rules 26 correction 19.12.2025]Mine water disaster is a major hidden danger affecting mine safety production. In China, almost all mines are troubled by water disaster accidents to varying degrees. The most important basis for water disaster prevention and control is to determine the source of the water, so as to take reasonable prevention and control measures and countermeasures. The indexes used to discriminate the water source at present mainly include inorganic indexes such as main anion and cation in water and total dissolved solids (TDS). However, for water sources with similar components, it is difficult to make accurate judgment only by inorganic water chemical indexes. SUMMARY

[0003] [According to the rules 26 correction 19.12.2025]In view of the problems in the prior art, the purpose of the present application is to provide a mine water source intelligent discrimination method, device, equipment and storage medium, which solves the problem that inorganic water chemical indexes are difficult to accurately judge the mine water source with similar components in the prior art.

[0004] [According to the rules 26 correction 19.12.2025]In order to solve the above technical problems, the present application adopts the following technical scheme to achieve it:

[0005] [According to the rules 26 correction 19.12.2025]A mine water source intelligent discrimination method, the method comprising the following steps:

[0006] [According to the rules 26 correction 19.12.2025]S1, obtaining the water chemical index of the water source sample to be discriminated; the water chemical index includes inorganic index and organic index; the inorganic index includes anion, cation and TDS value in the water sample; the organic index includes total organic carbon, ultraviolet absorbance and fluorescence intensity of dissolved organic matter fluorescence component in the water sample;

[0007] [According to the rules 26 correction 19.12.2025]S2, inputting the water chemical index into a preset water source discrimination model to obtain the water source type of the water source sample to be discriminated; the construction of the water source discrimination model comprises:

[0008] [According to the rules 26 correction 19.12.2025]S21, constructing a data set of water chemical indexes of a plurality of water source samples according to the water chemical indexes of the plurality of water source samples;

[0009] [According to the rules 26 corrections 19.12.2025] S22, dimensionality reduction is performed on the data set to obtain a reduced data set;

[0010] [According to the rules 26 corrections 19.12.2025] S23, the preset model is trained according to the reduced data set to obtain a water source discrimination model.

[0011] [According to the rules 26 corrections 19.12.2025] The present application also includes the following technical features:

[0012] [According to the rules 26 corrections 19.12.2025] Specifically, the data set of water source sample water chemical indicators in S21 is represented by the following formula (1):

[0013] [According to the rules 26 corrections 19.12.2025] Wherein, X is the data set, x is one of the water chemical indicators, n is the number of water source samples, and p is the number of water chemical indicators.

[0014] [According to the rules 26 corrections 19.12.2025] Specifically, S22 includes:

[0015] [According to the rules 26 corrections 19.12.2025] The data set is subjected to zero mean normalization processing to obtain a standardized data matrix;

[0016] [According to the rules 26 corrections 19.12.2025] The principal component score matrix is obtained by principal component extraction on the standardized data matrix;

[0017] [According to the rules 26 corrections 19.12.2025] The reduced data set is obtained by dimensionality reduction of the standardized data matrix through the principal component score matrix.

[0018] [According to the rules 26 corrections 19.12.2025] Specifically, the zero mean normalization processing of the data set is realized by the following formula (3):

[0019] [According to the rules 26 corrections 19.12.2025] Wherein, μ A is the mean of a water chemical indicator A, σ A is the standard deviation of a water chemical indicator A, and v is the normalized value of a certain value x of a water chemical indicator A;

[0020] [According to the rules 26 corrections 19.12.2025] The reduced data set is represented by formula (4): Y=XP (4)

[0021] [Rule 26 correction 19.12.2025] Wherein Y is the dimensionality-reduced data set, X is the data set of water chemical indicators of the plurality of water source samples after standardization, and P is the principal component score matrix.

[0022] [Rule 26 correction 19.12.2025] Specifically, in S23, the dimensionality-reduced data set is input into a preset model for training; the number and depth of the random forest model sub-tree are optimized through an intelligent swarm optimization algorithm to obtain a water source discrimination model.

[0023] [Rule 26 correction 19.12.2025] A mine water inrush water source intelligent discrimination device, which can realize the mine water inrush water source intelligent discrimination method, comprises:

[0024] [Rule 26 correction 19.12.2025] An acquisition module is configured to acquire water chemical indicators of a water source sample to be discriminated; the water chemical indicators include inorganic indicators and organic indicators; the inorganic indicators include anions and cations in water and a TDS value; the organic indicators include total organic carbon, ultraviolet absorbance, and fluorescence intensity of dissolved organic matter fluorescence components in the water sample.

[0025] [Rule 26 correction 19.12.2025] A discrimination module is configured to input the water chemical indicators into a preset water source discrimination model to obtain a water source type of the water source sample to be discriminated; the construction of the water source discrimination model comprises: constructing a data set of water chemical indicators of a plurality of water source samples according to the water chemical indicators of the plurality of water source samples; dimensionality-reducing the data set to obtain a dimensionality-reduced data set; and training a preset model according to the dimensionality-reduced data set to obtain the water source discrimination model.

[0026] [Rule 26 correction 19.12.2025] Specifically, in the discrimination module, the data set of water chemical indicators of the water source sample is represented by the following formula (1):

[0027] [Rule 26 correction 19.12.2025] Wherein X is the data set, x is one of the water chemical indicators, n is the number of water source samples, and p is the number of water chemical indicators.

[0028] [Rule 26 correction 19.12.2025] Specifically, in the discrimination module, the dimensionality-reduced data set is obtained by dimensionality-reducing the data set, comprising: performing zero-mean normalization processing on the data set to obtain a standardized data matrix; extracting principal components from the standardized data matrix to obtain a principal component score matrix; and dimensionality-reducing the standardized data matrix through the principal component score matrix to obtain the dimensionality-reduced data set; the zero-mean normalization processing on the data set is realized by the following formula (3):

[0029] [According to the rules 26 correction 19.12.2025] wherein, μ A is the mean of a water chemistry index A, σ A is the standard deviation of a water chemistry index A, v is the normalized value of a certain value x of a water chemistry index A;

[0030] [According to the rules 26 correction 19.12.2025] The reduced dimension data set is represented by formula (4): Y=XP (4)

[0031] [According to the rules 26 correction 19.12.2025] Wherein, Y is the reduced dimension data set, X is the data set obtained by standardizing the water chemistry index of the plurality of water source samples, and P is the principal component score matrix;

[0032] [According to the rules 26 correction 19.12.2025] The reduced dimension data set is input into a preset model for training; the number and depth of the random forest model sub-tree are optimized through the intelligent group optimization algorithm, and a water source discrimination model is obtained.

[0033] [According to the rules 26 correction 19.12.2025] A computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the intelligent discrimination method of the mine water inrush water source when executing the computer program.

[0034] [According to the rules 26 correction 19.12.2025] A computer readable storage medium for storing program instructions, the program instructions can be executed by a processor to implement the steps of the intelligent discrimination method of the mine water inrush water source.

[0035] [According to the rules 26 correction 19.12.2025] Compared with the prior art, the present application has the following technical effects:

[0036] [According to the rules 26 correction 19.12.2025] In the present application, according to the plurality of inorganic indexes and the plurality of organic indexes contained in the water chemistry index of the to-be-discriminated water source sample, the water source type of the to-be-discriminated water source sample is identified through the water source discrimination model. In this way, on the basis of inorganic index discrimination, organic indexes are added, solving the problem that inorganic water chemistry indexes are difficult to accurately discriminate similar water samples, and improving the accuracy of water source category judgment. BRIEF DESCRIPTION OF DRAWINGS

[0037] [According to the rules 26 correction 19.12.2025] Fig. 1 is an embodiment flowchart of a mine water inrush water source discrimination method provided by the present application;

[0038] [Rule 26 Correction 19.12.2025] FIG. 2 is a fluorescence spectrum and load diagram of an organic component;

[0039] [Rule 26 Correction 19.12.2025] FIG. 3 is a fluorescence spectrum and load diagram of another organic component;

[0040] [Rule 26 Correction 19.12.2025] FIG. 4 is a fluorescence spectrum and load diagram of another organic component;

[0041] [Rule 26 Correction 19.12.2025] FIG. 5 is an embodiment flow diagram for obtaining a water source discrimination model in an embodiment of the present application;

[0042] [Rule 26 Correction 19.12.2025] FIG. 6 is a performance diagram of a conventional RF model under different discrimination index data sets;

[0043] [Rule 26 Correction 19.12.2025] FIG. 7 is a performance diagram of an RF model improved based on AFSA in an embodiment of the present application under different discrimination index data sets;

[0044] [Rule 26 Correction 19.12.2025] FIG. 8 is a result diagram of water source type identification by a conventional RF model;

[0045] [Rule 26 Correction 19.12.2025] FIG. 9 is a result diagram of water source type identification by an RF model improved based on AFSA in an embodiment of the present application;

[0046] [Rule 26 Correction 19.12.2025] FIG. 10 is a structural diagram of a mine water inrush water source discrimination device in an embodiment of the present application;

[0047] [Rule 26 Correction 19.12.2025] FIG. 11 is a structural diagram of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0048] [Rule 26 Correction 19.12.2025] Currently, the indicators used for water source discrimination are mainly inorganic indicators such as major anions and cations in water and total dissolved solids (TDS). However, in practice, total organic carbon (TOC), ultraviolet absorbance (UV 254) and Dissolved Organic Matter (DOM) also exist great differences. At present, the methods of using hydrochemistry to distinguish water sources mainly focus on three categories: multivariate statistical method, nonlinear analysis method and machine learning. Compared with traditional methods, machine learning algorithms have more advantages in processing nonlinear and high-dimensional data, and have strong adaptability. Among commonly used machine algorithms, Support Vector Machine (SVM), Extreme Learning Machine (ELM), Random Forest (RF) and Back Propagation Neural Network (BP neural network) have been applied in mine water source discrimination; for water sources with similar composition, it is difficult to make accurate judgment only by inorganic water chemical indicators.

[0049] [According to Rule 26, correct 19.12.2025] The present application provides a kind of mine gushing water source intelligent discrimination method, device, equipment and storage medium, the method comprises: obtaining the water chemical index of sample to be discriminated water source;Wherein, water chemical index includes multiple inorganic indicators and multiple organic indicators;Water chemical index is input into preset water source discrimination model, obtains the water source type of sample to be discriminated water source.Such, on the basis of inorganic index discrimination, join organic index, solve the problem that inorganic water chemical index is difficult to accurately discriminate similar water sample, improve the accuracy of water source category judgment.

[0050] [According to Rule 26, correct 19.12.2025] The specific embodiments of the present application are given below, it should be noted that the present application is not limited to the following specific embodiments, any equivalent transformation made on the basis of the technical scheme of the present application falls within the protection scope of the present application.

[0051] [According to Rule 26, correct 19.12.2025] Embodiment:

[0052] [According to Rule 26, correct 19.12.2025] The present embodiment provides a kind of mine gushing water source intelligent discrimination method, Fig. 1 is the flowchart of mine gushing water source intelligent discrimination method of the present application, see Fig. 1, the method comprises the following steps:

[0053] [According to Rule 26, correct 19.12.2025] S1, obtaining the water chemical index of sample to be discriminated water source;Wherein, water chemical index includes multiple inorganic indicators and organic indicators;

[0054] [According to Rule 26, correct 19.12.2025] Inorganic indicators include main anion and cation in water and TDS value etc.Main ions in water include K + , Na+ Ca 2+ Mg 2+ Cl - SO4 2- HCO3 - and so on. GB / T 14848—2017 Groundwater Quality Standard can be referred to for testing the inorganic ions and TDS in the water sample. K + Na + are all common inorganic ions in water, but the content of K + is usually low. Therefore, K + Na + can be considered as a whole, represented in the form of K + + Na + .

[0055] [Corrected according to Rule 26 19.12.2025] Organic indicators include total organic carbon, ultraviolet absorbance, and dissolved organic matter in the water layer. The differences between different organic components in dissolved organic matter can be intuitively reflected by fluorescence spectrum. The multi N / C 2100 expert total organic carbon / total nitrogen analyzer can be used to detect TOC; the Evolution 60 ultraviolet-visible spectrophotometer can be used to detect UV 254 ; and the fluorescence spectrophotometer (HITACHI F-7000) can be used to extract DOM fluorescence data. In order to quantitatively analyze the differences of DOM in different water samples, the DOM fluorescence data of each water source sample is quantified by mathematical method. Specifically, the main components of fluorescence spectrum can be extracted and quantified by parallel factor method (Parallel Factors Analysis, PARAFAC). Specifically, first, arrange the three-dimensional fluorescence data of all water samples DOM, then use the fluorescence microscopy digital object identifier (Digital Object Identifier for Fluorescence Microscopy, DOMFluor) toolbox provided by matrix laboratory (Matrix Laboratory, MATLAB) to analyze and process the three-dimensional fluorescence data. The specific processing steps include:

[0056] [Corrected according to Rule 26 19.12.2025] (a1) Load data and draw excitation-emission matrix (Excitation-Emission Matrixs, EEMs);

[0057] [Corrected according to Rule 26 19.12.2025] (a2) Cut the spectral region affected by scattering peak;

[0058] [Rule 26 correction 19.12.2025] (a3) Outlier identification and load analysis of different component number models;

[0059] [Rule 26 correction 19.12.2025] (a4) Split-half test on the selected component number model;

[0060] [Rule 26 correction 19.12.2025] (a5) Derive experimental data, including: the fluorescence spectrum of each component, the fluorescence intensity of each component in each sample and its corresponding emission load and excitation load.

[0061] [Rule 26 correction 19.12.2025] FIGS. 2-4 are schematic diagrams of the fluorescence spectrum and load of the organic components in the embodiment, wherein FIGS. 2-4 are the fluorescence spectrum and load of three organic components, respectively. Referring to FIG. 2, Component 1 has two excitation peaks (250 nm / 330 nm) and one emission peak (405 nm), which indicates that Component 1 contains hydrophobic organic acid and humic acid-like and the latter is more. Referring to FIG. 3, Component 2 has three excitation peaks (230 nm / 250 nm / 280 nm) and one emission peak (308 nm), which indicates that Component 2 contains tryptophan-like and tryptophan-containing protein-like. Referring to FIG. 4, Component 3 has two excitation peaks (225 nm / 280 nm / 305 nm) and one emission peak (340 nm), which indicates that Component 3 contains tyrosine and tryptophan-containing protein-like and the former is more.

[0062] [Rule 26 correction 19.12.2025] S2, input the water chemical index into the preset water source discrimination model to obtain the water source type of the water source sample to be discriminated.

[0063] [Rule 26 correction 19.12.2025] The water source type includes Quaternary water, Cretaceous water, Zhuluo Formation water or Yan'an Formation water and the like aquifer water sample.

[0064] [Rule 26 correction 19.12.2025] The water source type of the water source is discriminated to realize the prevention of mine water disaster, therefore, in the embodiment, before the water source type is discriminated, the geological and hydrogeological conditions of the area to be studied can be studied to determine the target aquifer to be studied. That is, the water chemical index of the water source sample to be discriminated can be obtained from the target aquifer.

[0065] [According to the rules 26 correction 19.12.2025] Figure 5 is a flowchart of obtaining a water source discrimination model in an embodiment of the present application, referring to Figure 5, the process of obtaining a water source discrimination model includes the following steps:

[0066] [According to the rules 26 correction 19.12.2025] S21, according to the water chemical indicators of a plurality of water source samples, a data set of water chemical indicators of a plurality of water source samples is constructed;

[0067] [According to the rules 26 correction 19.12.2025] The water chemical indicators include a plurality of inorganic indicators and organic indicators as described above, so that for each sample water source, a set of data containing inorganic indicators and organic indicators as described above can be obtained. By obtaining the water chemical indicators of a plurality of sample water sources, a data set of water chemical indicators of a plurality of sample water sources can be constructed. For example, the inorganic indicators include K + +Na + , Ca 2+ , Mg 2+ , Cl - , SO4 2- , HCO3 - and TDS. The organic indicators include total organic carbon, ultraviolet absorbance and dissolved organic matter, and the dissolved organic matter can obtain specific organic components and contents by analyzing the fluorescence spectrum, for example, the organic components in the embodiment include three components as shown in Figures 2 to 4. Thus, the data set of water chemical indicators of each sample water source can be obtained, and a data set of water chemical indicators of a plurality of sample water sources can be constructed.

[0068] [According to the rules 26 correction 19.12.2025] The number of water source samples is n, and each water source sample includes p water chemical indicators, which are represented by x1, x2, …, x p ; the data set of n water source samples is represented by the following formula (1):

[0069] [According to the rules 26 correction 19.12.2025] Wherein X is the data set, x is one of the water chemical indicators, n is the number of water source samples, and p is the number of water chemical indicators.

[0070] [According to the rules 26 correction 19.12.2025] S22, the data set is reduced by a predetermined data method, and a reduced data set is obtained;

[0071] [Corrected according to Rule 26 19.12.2025] The discrimination of mine water sources is affected by the correlation between various water chemical indicators, such as the above-mentioned inorganic ions and organic ions, and there is a certain correlation between the water chemical indicators, which greatly increases the complexity of the analysis, so it is necessary to reduce the dimensionality of the original data set and process the high-dimensional, correlated variables into new synthetic variables that are low-dimensional, mutually independent, and can represent most of the information of the original variables. The correlation between each water chemical indicator is determined to reduce unnecessary calculations. The correlation between each water chemical indicator is determined by the Pearson correlation coefficient (Pearson), which is represented by formula (2):

[0072] [Corrected according to Rule 26 19.12.2025] Wherein, r xy is the Pearson correlation coefficient of index x and index y, n is the number of samples, x i is the i-th actual value of index x, y i is the i-th actual value of index y.

[0073] [Corrected according to Rule 26 19.12.2025] Table 1 is a correlation coefficient matrix composed of the correlation coefficients between each water chemical indicator in the actual measured sample water source.

[0074] [Corrected according to Rule 26 19.12.2025] Table 1 is a correlation coefficient matrix composed of the correlation coefficients between each water chemical indicator in the actual measured sample water source

[0075] [Corrected according to Rule 26 19.12.2025] Wherein, the water chemical indicators include inorganic indicators K + +Na + , Ca 2+ , Cl - , SO4 2- and other inorganic ions, and TDS, and organic indicators include UV 254 , TOC and component 2 as shown in Figure 2.

[0076] [Corrected according to Rule 26 19.12.2025] As can be seen from Table 1, there is a clear correlation between each indicator, among which the correlation of SO4 2- with K + +Na + , Ca 2+ , TDS reaches 0.916, 0.949 and 0.982 respectively; the correlation coefficients of K + +Na + with Ca 2+ , TDS reach 0.825 and 0.969 respectively; the correlation coefficient of Ca 2+The correlation coefficients with TDS reached 0.908 respectively. C2 showed a significant negative correlation with K + +Na + , Ca 2+ , SO4 2- and TDS, which indicated that there was information redundancy in the data set. If the eight indicators were directly used to identify the water source type, the complexity and calculation amount of the model would be increased, which would affect the discrimination efficiency and accuracy. Therefore, the data set needs to be reduced.

[0077] [According to Rule 26 Correction 19.12.2025]In the embodiments of the present application, the dimensionality reduction of the data set of the water chemical indicators of the above-mentioned multiple sample water sources is realized by principal component analysis (PCA).

[0078] [According to Rule 26 Correction 19.12.2025]The PCA is a classical statistical method, which extracts new variables reflecting the nature of things by linear transformation of the attribute characteristics of the original variables, removes redundancy and reduces noise, and achieves the purpose of dimensionality reduction.

[0079] [According to Rule 26 Correction 19.12.2025]The above step S22 includes the following steps:

[0080] [According to Rule 26 Correction 19.12.2025]S221, the data set is processed by zero-mean normalization to obtain a standardized data matrix;

[0081] [According to Rule 26 Correction 19.12.2025]S222, the principal component extraction is performed on the standardized data matrix to obtain a principal component score matrix;

[0082] [According to Rule 26 Correction 19.12.2025]S223, the dimensionality reduction of the standardized data matrix is performed by the principal component score matrix to obtain a reduced data set.

[0083] [According to Rule 26 Correction 19.12.2025]The zero-mean normalization processing of the data set can make the mean value of the transformed attribute of each data in the data set zero. It can be obtained by subtracting the mean value of each attribute in the data set from the value of the attribute.

[0084] [According to Rule 26 Correction 19.12.2025]The zero-mean normalization processing of the data set can be realized by the following formula (3):

[0085] [According to Rule 26 Correction 19.12.2025]Wherein, μ A is the mean value of a water chemical indicator A, and σA is a standard deviation of a water chemistry index A, and v is a normalized value of a certain value x of a water chemistry index A.

[0086] [Corrected according to Rule 26 19.12.2025] After the data set is zero-mean normalized by the above formula (3), a standardized data matrix can be obtained, and then a principal component score matrix is obtained by performing principal component extraction on the standardized data matrix.

[0087] [Corrected according to Rule 26 19.12.2025] Specifically, the above step S222 includes the following steps:

[0088] [Corrected according to Rule 26 19.12.2025] Step b1, determining the covariance matrix of the standardized data matrix;

[0089] [Corrected according to Rule 26 19.12.2025] Step b2, determining the eigenvalues of the covariance matrix and the eigenvectors corresponding to the eigenvalues;

[0090] [Corrected according to Rule 26 19.12.2025] Step b3, selecting N eigenvalues that meet the preset condition according to the size of the eigenvalues, to obtain an eigenvalue set;

[0091] [Corrected according to Rule 26 19.12.2025] Step b4, constructing a principal component score matrix according to the eigenvectors corresponding to each eigenvalue in the eigenvalue set.

[0092] [Corrected according to Rule 26 19.12.2025] When the water chemistry index of each sample water source contains p indexes, the covariance matrix can be solved to obtain p eigenvalues. The corresponding feature information contained in the data is gradually decreasing, that is, the contribution rate of each eigenvalue is gradually decreasing. Therefore, in the calculation process, in order to reduce redundant calculation, only the eigenvectors corresponding to part of the eigenvalues can be used as the composition of the principal components, so as to construct the principal component score matrix.

[0093] [Corrected according to Rule 26 19.12.2025] Wherein, the contribution rate of the eigenvalue is the proportion of the eigenvalue in all eigenvalues of the covariance matrix, that is, the proportion of the variance corresponding to the eigenvalue in the sum of all variances in the principal component. The greater the contribution rate of the eigenvalue, the more information the original water chemistry index contains in the corresponding principal component, and therefore the value of N can be determined by the contribution rate of the eigenvalue corresponding to the covariance matrix.

[0094] [According to Rule 26, the correction is made on 19.12.2025] The preset condition can be that the feature value is greater than a first preset threshold. Wherein, the value of the first preset threshold can be a pre-set fixed threshold, or can also be specifically determined based on the application demand in the actual calculation process.

[0095] [According to Rule 26, the correction is made on 19.12.2025] When the feature value is less than a certain threshold, the proportion of its corresponding contribution rate is small, that is, it contains less information. Therefore, this part of the feature value can not be considered. By selecting the feature value greater than the first preset threshold to exclude the part of the feature value containing less information, a large amount of redundant calculation can be reduced, and the calculation efficiency can be improved.

[0096] [According to Rule 26, the correction is made on 19.12.2025] According to the size of the feature value, N feature values satisfying the preset condition are selected to obtain the feature value set, which can be respectively compared each feature value with the first preset threshold, and the feature value greater than the first preset threshold is selected to construct the feature value set.

[0097] [According to Rule 26, the correction is made on 19.12.2025] In other embodiments, the preset condition can be that the cumulative contribution rate of each feature value is greater than a second preset threshold.

[0098] [According to Rule 26, the correction is made on 19.12.2025] Similarly, the value of the second preset threshold can also be a pre-set fixed threshold, or can also be specifically determined based on the application demand in the actual calculation process.

[0099] [According to Rule 26, the correction is made on 19.12.2025] In some embodiments, taking 85% as an example of the second threshold, according to the size of the feature value, the feature values are sorted from large to small, and the feature values whose cumulative contribution rate exceeds 85% can actually represent most of the information. Therefore, the value of N can be the first N feature values whose contribution rate exceeds 85%. The specific contribution rate can also be selected based on the demand in the actual application process, and the embodiments of the application do not make specific limitations.

[0100] [According to Rule 26, the correction is made on 19.12.2025] After the above steps b1 to b4, the principal component score matrix can be obtained, and the dimensionality reduction of the standardized data matrix is realized through the principal component score matrix, so as to obtain the reduced data set.

[0101] [According to Rule 26, the correction is made on 19.12.2025] The reduced data set is represented by formula (4): Y=XP (4)

[0102] [According to Rule 26, corrected on 19.12.2025] Wherein Y is the reduced dimension data set, X is the data set obtained by standardizing the water chemical indicators of the plurality of water samples, and P is the principal component score matrix.

[0103] [According to Rule 26, corrected on 19.12.2025] Wherein each principal component in the principal component score matrix is represented by formula (5): F m = p m1 X1+p m2 X2+…+p mp X p (5)

[0104] [According to Rule 26, corrected on 19.12.2025] Wherein F m is the mth principal component, p m1 to p mp is the principal component score corresponding to each indicator of the mth principal component (i.e. the mth row of the principal component matrix), X1 to X p is the standardized value of the 1st to pth indicators of the n samples, i.e. the 1st to pth columns of the matrix X.

[0105] [According to Rule 26, corrected on 19.12.2025] S23, training the preset model according to the reduced dimension data set to obtain a water source discrimination model.

[0106] [According to Rule 26, corrected on 19.12.2025] Wherein the preset model can be a random forest model;

[0107] [According to Rule 26, corrected on 19.12.2025] In some embodiments, the above step S23 can further include the following steps:

[0108] [According to Rule 26, corrected on 19.12.2025] S231, inputting the reduced dimension data set into the preset model for training;

[0109] [According to Rule 26, corrected on 19.12.2025] S232, optimizing the number and depth of the random forest model sub-tree through the intelligent swarm optimization algorithm to obtain the water source discrimination model.

[0110] [According to Rule 26, correct 19.12.2025] Specifically, after inputting the dimensionality-reduced dataset into the random forest model, K-fold cross-validation can be used to divide the dataset, with the Kth fold as the test set, where K can be any positive integer less than the number of data in the dataset. After taking the Kth fold as the test set, the remaining K-1 folds can be used as the training set, and this process is repeated K times. In the random forest model classifier, bootstrap sampling is performed on the training set, the Gini index is used to select the best split feature during feature selection, the Classification And Regression Tree (CART) algorithm is used to construct the decision tree, and the test set is predicted based on the voting results of the decision tree.

[0111] [According to Rule 26, correct 19.12.2025] At the same time, the number of random forest sub-trees n_estimators and depth can be optimized by intelligent swarm optimization algorithm, here the Artificial Fish Swarms Algorithm (AFSA) is used. In AFSA, the state (current position) of each artificial fish is X i = (n_estimators, depth), and the food concentration (fitness) of the current position of the artificial fish is the average accuracy of the RF classifier after K-fold cross-validation. First, the positions of all artificial fish in AFSA are initialized, and then the four behaviors of foraging, grouping, following and random are executed according to the current fitness, and the maximum iteration number is taken as the termination condition of the algorithm.

[0112] [According to Rule 26, correct 19.12.2025] Figure 4 is a schematic diagram of the performance comparison of the RF model actually measured under different index systems. The performance test is carried out under four different index datasets, including accuracy, precision, recall and f1 score. The performance of different models for water source type identification is compared based on the inorganic index discrimination provided in the embodiment of the application, and the performance of different models for water source type identification is compared based on the inorganic index discrimination provided in the embodiment of the application. Figure 6 is a schematic diagram of the performance of the traditional RF model under different index datasets. Figure 7 is a schematic diagram of the performance of the RF model improved based on AFSA in the embodiment of the application under different index datasets.

[0113] [According to Rule 26, correct 19.12.2025] As shown in Figure 4, the combination of inorganic indicators and organic indicators can significantly improve the accuracy of water source discrimination. At the same time, AFSA can further improve the accuracy of the RF model.

[0114] [According to Rule 26 Correction 19.12.2025] Figure 5 is a schematic diagram of the results of actual testing of different model discrimination of water source types in embodiments of the application. Among them, the sample water sources include 74, Figure 8 is a schematic diagram of the results of water source type identification by a traditional RF model, and Figure 9 is a schematic diagram of the results of water source type identification by an RF model improved based on AFSA in embodiments of the application.

[0115] [According to Rule 26 Correction 19.12.2025] As can be seen from Figures 8 to 9, the traditional RF model has 2 misjudgments in identifying the water source types of 74 samples. However, the RF model improved based on AFSA provided in embodiments of the application has only 1 misjudgment. This fully reflects the performance advantages of the RF model improved based on AFSA.

[0116] [According to Rule 26 Correction 19.12.2025] In embodiments of the application, the water source type of the water source sample to be discriminated is identified by the water source discrimination model according to the multiple inorganic indexes and multiple organic indexes contained in the water chemical indexes of the water source sample to be discriminated. In this way, on the basis of inorganic index discrimination, the addition of organic indexes solves the problem that inorganic water chemical indexes are difficult to accurately discriminate similar water samples, and improves the accuracy of water source category judgment. At the same time, the RF is optimized using AFSA, and the improved algorithm has good global search ability and convergence, improving the efficiency in the process of model parameter optimization.

[0117] [According to Rule 26 Correction 19.12.2025] In some embodiments, after obtaining the water source discrimination model through the above step S23, the above method can further include the following steps:

[0118] [According to Rule 26 Correction 19.12.2025](c1) obtaining water chemical indexes of a plurality of test sample water sources;

[0119] [According to Rule 26 Correction 19.12.2025](c2) inputting the water chemical indexes of the plurality of test sample water sources into the water source discrimination model to obtain water source type discrimination results of the plurality of test sample water sources;

[0120] [According to Rule 26 Correction 19.12.2025](c3) determining whether the accuracy of the judgment result is greater than a preset threshold;

[0121] [According to Rule 26 Correction 19.12.2025](c4) when the accuracy is greater than or equal to the preset threshold, outputting the water source discrimination model;

[0122] [Amended according to Rule 26 on 19.12.2025](c5) retraining the water source discrimination model iteratively when the accuracy is less than a preset threshold.

[0123] [Amended according to Rule 26 on 19.12.2025] Wherein the test sample water source can be a part of samples selected from a plurality of sample water sources, or can also be a water sample of a known water source type obtained separately. The embodiment of the present application does not make specific limitation on this.

[0124] [Amended according to Rule 26 on 19.12.2025] Due to the influence of the number of samples, the number of iterations in the model training process, and other reasons affecting the performance of the model, the performance of the water source discrimination model obtained by one model training may not meet the actual application requirements. Therefore, after obtaining the water source discrimination model, the performance of the water source discrimination model can be detected by the above method, and the accuracy of the water source discrimination result output by the water source discrimination model is used to detect the performance of the model. Only when the performance of the water source discrimination model meets the requirements, the model is output for actual use.

[0125] [Amended according to Rule 26 on 19.12.2025] In the embodiment of the present application, the water source type of the to-be-discriminated water sample is identified by the water source discrimination model according to the plurality of inorganic indexes and the plurality of organic indexes contained in the water chemical indexes of the to-be-discriminated water sample. In this way, on the basis of inorganic index discrimination, organic indexes are added to solve the problem that inorganic water chemical indexes are difficult to accurately discriminate similar water samples, thereby improving the accuracy of water source category judgment. At the same time, the AFSA is used to optimize the RF, and the improved algorithm has better global search ability and convergence, thereby improving the efficiency in the process of model parameter optimization.

[0126] [Amended according to Rule 26 on 19.12.2025] The embodiment of the present application also provides a mine gushing water source discrimination device. FIG. 10 is a structural schematic diagram of a mine gushing water source discrimination device in the embodiment of the present application, referring to FIG. 10, the discrimination device comprises: an acquisition module, configured to acquire water chemical indexes of a to-be-discriminated water sample; wherein the water chemical indexes comprise a plurality of inorganic indexes and a plurality of organic indexes; a discrimination module, configured to input the water chemical indexes into a preset water source discrimination model to obtain a water source type of the to-be-discriminated water sample.

[0127] [Amended according to Rule 26 on 19.12.2025] The preset water source discrimination model is obtained by the following processing: obtaining a data set of water chemical indexes of a plurality of water samples according to the water chemical indexes of the plurality of water samples; performing dimension reduction on the data set by a preset data processing method to obtain a dimension-reduced data set; training a preset model according to the dimension-reduced data set to obtain the water source discrimination model.

[0128] [According to the rules 26 correction 19.12.2025] by preset data processing method to data set dimension reduction, obtain the data set after dimension reduction, including: the data set is carried out zero mean normalization processing, obtains the standardized data matrix;The principal component extraction is carried out to the standardized data matrix, and the principal component score matrix is obtained;The standardized data matrix is reduced by the principal component score matrix, and the data set after dimension reduction is obtained.

[0129] [According to the rules 26 correction 19.12.2025] the principal component extraction is carried out to the standardized data matrix, and the principal component score matrix is obtained, comprising: determining the covariance matrix of the standardized data matrix;The eigenvalue and the eigenvector corresponding to the eigenvalue of the covariance matrix are determined;According to the size of the eigenvalue, the N eigenvalues meeting the preset condition are selected, and the eigenvalue set is obtained;The principal component score matrix is constructed according to the eigenvector corresponding to each eigenvalue in the eigenvalue set.

[0130] [According to the rules 26 correction 19.12.2025] the preset model is a random forest model;The water source discrimination model is obtained by training the preset model according to the data set after dimension reduction, comprising: the data set after dimension reduction is input into the preset model for training;The number and depth of random forest model sub-tree are optimized by intelligent group optimization algorithm, and the water source discrimination model is obtained.

[0131] [According to the rules 26 correction 19.12.2025] the discrimination device further comprises a test module, which is used for obtaining the water chemical index of a plurality of test sample water sources;The water chemical index of the plurality of test sample water sources is input into the water source discrimination model, and the water source type judgment result of the plurality of test sample water sources is obtained;Determine whether the accuracy of the judgment result is greater than the preset threshold;When the accuracy is greater than or equal to the preset threshold, the water source discrimination model is output;When the accuracy is less than the preset threshold, the water source discrimination model is retrained iteratively.

[0132] [According to the rules 26 correction 19.12.2025] the electronic device provided by the embodiment of the application can be consistent with the mine water inflow source discrimination method in one or more of the above embodiments. Figure 11 is a schematic structural diagram of an electronic device in an embodiment of the application, as shown in Figure 11, the electronic device can adopt general computer hardware, including processor, memory.

[0133] [Amended according to Rule 26 19.12.2025] The at least one processor can constitute any physical device having circuitry that performs logical operations on one or more inputs. For example, the at least one processor can include one or more integrated circuits (ICs), including application-specific integrated circuits (ASICs), microchips, microcontrollers, microprocessors, all or part of a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processor (DSP), a field-programmable gate array (FPGA), or other circuitry suitable for the execution of instructions or the performance of logical operations. Instructions executed by the at least one processor can be, for example, pre-loaded into a memory integrated with or embedded in the controller, or can be stored in a separate memory. The memory can include random access memory (RAM), read-only memory (ROM), a hard disk, an optical disk, a magnetic medium, a flash memory, other permanent, fixed, or volatile memory, or any other mechanism capable of storing instructions. In some embodiments, the at least one processor can include more than one processor. Each processor can have a similar structure, or the processors can have different structures that are electrically connected or disconnected from each other. For example, the processors can be separate circuits or integrated in a single circuit. When more than one processor is used, the processors can be configured to operate independently or cooperatively. The processors can be coupled electrically, magnetically, optically, acoustically, mechanically, or by other means that allow them to interact.

[0134] [Amended according to Rule 26 19.12.2025] The present application provides a computer storage medium, the computer storage medium stores computer executable instructions, the computer executable instructions are executed by a processor, and a mine water inrush water source discrimination method as one or more of the above embodiments can be realized.

[0135] [Amended according to Rule 26 19.12.2025] The various embodiments in the specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments.

[0136] [Amended according to Rule 26 19.12.2025] The above embodiments are only used to illustrate the technical solutions of the present application, and are not limited to the present application; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the present application.

Claims

1. [Amended according to Rule 26 19.12.2025] A method for intelligently identifying a mine water gushing source, characterized in that, The method comprises the following steps: S1, obtaining water chemical indexes of a water source sample to be identified; the water chemical indexes comprise inorganic indexes and organic indexes; the inorganic indexes comprise anions, cations and TDS values in the water sample; the organic indexes comprise total organic carbon, ultraviolet absorbance and fluorescence intensity of dissolved organic matter fluorescence components in the water sample; S2, inputting the water chemical indexes into a preset water source identification model to obtain a water source type of the water source sample to be identified; the construction of the water source identification model comprises: S21, constructing a data set of water chemical indexes of multiple water source samples according to the water chemical indexes of the multiple water source samples; S22, reducing the dimension of the data set to obtain a reduced dimension data set; S23, training a preset model according to the reduced dimension data set to obtain the water source identification model.

2. [Amended according to Rule 26 19.12.2025] The intelligent identification method of mine water inrush source according to claim 1, characterized in that, The data set of the water chemical index of the S21 reclaimed water source sample is represented by the following formula (1): Wherein, X is the data set, x is one of the water chemical indexes, n is the number of water source samples, and p is the number of water chemical indexes.

3. The method of claim 2, wherein the method further comprises: The S22 comprises: performing zero-mean normalization processing on the data set to obtain a standardized data matrix; extracting principal components from the standardized data matrix to obtain a principal component score matrix; reducing the dimension of the standardized data matrix through the principal component score matrix to obtain the reduced dimension data set.

4. [Amended according to Rule 26 19.12.2025] The intelligent identification method of mine water inrush source according to claim 3, characterized in that, The zero-mean normalization of the data set is implemented by the following formula (3): where μ A is the mean of a water chemistry index A, σ A is the standard deviation of a water chemistry index A, and v is the normalized value of a certain value x of a water chemistry index A. The reduced dimension data set is represented by formula (4): Y=XP (4) Wherein, Y is the reduced dimension data set, X is the data set obtained by standardizing the water chemical indexes of the multiple water source samples, and P is the principal component score matrix.

5. [Amended according to Rule 26 19.12.2025] The intelligent identification method of mine water inrush source according to claim 1, characterized in that, In the S23, the reduced dimension data set is input into the preset model for training; the number and depth of the random forest model sub-tree are optimized through an intelligent group optimization algorithm to obtain the water source identification model.

6. [Amended according to Rule 26 19.12.2025] A mine water inrush source intelligent identification device, characterized in that, The device can realize the mine water inflow water source intelligent identification method of claim 1, comprising: an acquisition module configured to acquire water chemical indexes of a water source sample to be identified; the water chemical indexes comprise inorganic indexes and organic indexes; the inorganic indexes comprise anions, cations and TDS values in the water sample; the organic indexes comprise total organic carbon, ultraviolet absorbance and fluorescence intensity of dissolved organic matter fluorescence components in the water sample; an identification module configured to input the water chemical indexes into a preset water source identification model to obtain a water source type of the water source sample to be identified; the construction of the water source identification model comprises: constructing a data set of water chemical indexes of multiple water source samples according to the water chemical indexes of the multiple water source samples; reducing the dimension of the data set to obtain a reduced dimension data set; training a preset model according to the reduced dimension data set to obtain the water source identification model.

7. The mine water inflow source intelligent discrimination device of claim 6, characterized in that, The data set of the water source sample water chemical index is expressed by the following formula (1): Wherein, X is the data set, x is one of the water chemical indexes, n is the number of water source samples, and p is the number of water chemical indexes.

8. The mine water inflow source intelligent discrimination device of claim 6, wherein, The discriminant module includes: performing zero-mean normalization processing on the data set to obtain a standardized data matrix; performing principal component extraction on the standardized data matrix to obtain a principal component score matrix; and performing dimension reduction on the standardized data matrix through the principal component score matrix to obtain the dimension-reduced data set; and the zero-mean normalization processing on the data set is implemented through the following formula (3): where μ A is the mean of a water chemistry index A, σ A is the standard deviation of a water chemistry index A, and v is the normalized value of a certain value x of a water chemistry index A. The reduced dimension data set is represented by formula (4): Y=XP (4) Wherein, Y is the reduced dimension data set, X is the data set obtained by standardizing the water chemical indexes of the multiple water source samples, and P is the principal component score matrix; The reduced dimension data set is input into the preset model for training; the number and depth of the random forest model sub-tree are optimized through an intelligent group optimization algorithm to obtain the water source identification model.

9. [Amended according to Rule 26 19.12.2025] A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor implements the steps of the mine water inrush water source intelligent identification method in any one of claims 1 to 5 when executing the computer program.

10. [Amended according to Rule 26 19.12.2025] A computer-readable storage medium, characterized in that, The computer readable storage medium is used to store program instructions executable by the processor to implement the steps of the mine water inrush water source intelligent identification method in any one of claims 1 to 5.

Images