CN104899358A - Prediction method for lateral distribution of ordovician limestone karst crack water network - Google Patents

Abstract

The invention discloses a method for predicting the lateral distribution of the karst fissure water network of Austrian limestone, which comprises: firstly determining the indicators closely related to the distribution of the fissure water network of the Austrian limestone karst; Raw index data, establish a nuclear principal component model for the collected original index data, extract new principal components, and then perform fuzzy normalization, and use the fuzzy normalized principal component data and geophysical detection of Orsanite karst anomaly types to form a sample set, and establish a genetic Algorithm optimizes the prediction model of SVM; use the established model to predict the type of Austrian lime anomaly area in the borehole without geophysical detection; finally draw the distribution map of the Austrian lime anomaly area type to judge the Austrian limestone karst The distribution range of abnormal area types, and the direction of the seepage field of the Austrian limestone fissure water network is analyzed. The design principle of the invention is reliable, the prediction method is simple, the prediction accuracy is high, and the prediction environment is friendly.

Description

Prediction method of lateral distribution of fissure water network in Austrian limestone karst

technical field

The invention relates to a method for predicting the lateral distribution of the karst fissure water network of Austrian limestone, in particular to a method for predicting the lateral distribution of the fissure water network of the Austrian limestone karst coalfield in North China.

Background technique

Water inrush from floor of mine stope is a common problem in coal mine production. nature and cannot be observed directly, so it is difficult to study. However, as far as the occurrence conditions of groundwater are concerned, they have their own laws, which can be studied qualitatively and quantitatively. The direct cause of the water inrush from the stope floor is the groundwater network below the floor, without the existence of the groundwater network, it is impossible for a large water inrush accident to occur. After nearly half a century of mining in North China-type coalfields in my country, most of the mines have entered deep mining and are generally threatened by the outburst of Ordos limestone solution water. Therefore, determining the spatial distribution of the groundwater network is a key issue in the prevention and control of Ordos limestone water inrush. The primary task. Relevant scholars at home and abroad have done many studies on the vertical development of Orsanite karst, and have achieved some important results. However, there are few studies on the lateral distribution of Orszanolite karst. In the prior art, the position of the main vein of the groundwater network is roughly determined mainly through technical means such as water discharge test, location of floor water inrush points, distribution of karst collapse columns, development of faults, tracer tests, and drilling cores. There is no comprehensive analysis or quantitative research, and the various test methods are expensive, the test points are few, and the location of water inrush points and the development points of karst collapse columns are also extremely limited. For the prediction of the lateral distribution of Ordolime Karst fissure water network, it is now difficult to In the prior art, there is no report on establishing a quantitative model by using quantitative parameters. Therefore, it is necessary to find a method that can not only save money but also comprehensively reflect the development factors of karst fissure water network in most areas to predict the lateral distribution of Austrian limestone karst fissure water network. And provide a basis for forecasting and forecasting of water inrush.

The karst fissure water network is continuously evolved on the basis of the structural fissure water network under the action of the regional groundwater vector seepage field. The network system has karst fissure channels and abundant groundwater. The spatial distribution of karst fractures is mainly controlled by the degree of development of various structures formed by structural damage. Therefore, the spatial distribution of the fractured water network in Ordos lime karst can be clarified by studying the interaction of three factors: the development degree of structural fractures, the karst channel and the water-rich degree of Ordos lime karst. If the three factors can be quantitatively evaluated with easy-to-obtain and abundant indicators, and a reasonable and reliable prediction model can be constructed, the distribution of the fissure water network in Ordolime Karst can be determined. Tectonic movements form large-scale structural fault zones, folds, and numerous small fissures in underground hard rocks. The development of groundwater networks depends on structural fissures. The spatial combination of these fissures forms the initial fissure water network system. The impact of comprehensive faults Factor, fault fractal dimension value, and fold fractal dimension value can quantitatively evaluate the development degree of structural fractures. The abnormality of the geothermal field is obviously controlled by the regional structure and large faults. If the groundwater circulation channel leads the low-temperature groundwater near the surface and shallow to the deep, the water temperature will decrease. If the deep groundwater rises along the fault, the water temperature will rise. Therefore, the groundwater temperature Anomalies can be used as an important indicator to judge whether structural fissures are karst channels. The division of the water-rich degree of Austrian ash is mainly based on the "Regulations on Water Prevention and Control in Coal Mine" and the water inflow per unit of drilling (q) value. Theoretically, this division standard is scientific, but objectively only uses the q value to divide the water content The operability of the water-rich layer is poor, mainly because the q value is usually obtained in the well field exploration stage, and the quantity is extremely limited, followed by the long investment and time-consuming to obtain the q value; cores, but not every borehole will take cores and count the maximum leakage of flushing fluid; and with the expansion of mine mining area, the data of downhole Austrian ash hydrological drilling is becoming more and more abundant, and the data obtained from downhole hydrological drilling is the water inflow of the borehole, which can reflect the water-richness of the aquifer to a certain extent. The larger the water inflow value, the better the water-richness of the aquifer and the better the connectivity. On the other hand, geophysical detection has achieved good detection results in the detection of water-rich anomalies and water-bearing structures in aquifers, but not every austrian ash hydrological borehole has geophysical detection. Therefore, it is necessary to find a This is an accurate method, using the index values and detection results obtained from hydrological boreholes with geophysical detection to predict the karst anomalies in other areas without geophysical detection, and provide a strong basis for the prevention and control of coal mine floor water inrush.

Contents of the invention

The purpose of the present invention is to overcome the deficiencies in the prior art above, and provide a method for predicting the lateral distribution of austrian limestone fissure water network. The factors that are closely related to the distribution of Austrian limestone fissure water network can be collected in most areas, and the results of geophysical detection of Austrian lime anomaly area can be comprehensively used to avoid blindness and subjectivity in judging the distribution of groundwater network. The design principle is reliable , the prediction method is simple, the prediction accuracy is high, and the prediction environment is friendly.

To achieve the above object, the present invention adopts the following technical solutions:

A method for predicting the lateral distribution of the karst fissure water network in Austrian limestone, comprising the following steps:

(1) Determine the indicators closely related to the distribution of the Austrian lime karst fracture water network, and then collect the original data of the indicators at the underground Austrian limestone hydrological borehole with geophysical detection;

(2) Establish KPCA-Fuzzy-GA-SVM forecasting model for the austrian limestone anomaly area: establish a kernel principal component model (KPCA) for the collected index raw data, extract new principal components, and then carry out fuzzy standardization (Fuzzy), The sample set is composed of the principal component data after fuzzy normalization and the type of geophysical detection australian lime karst anomaly area, and the prediction model of SVM optimized by genetic algorithm (GA) is established;

(3) Collect the original data of the indicators at the downhole Ordos ash hydrological borehole without geophysical detection, and use the established KPCA–Fuzzy-GA-SVM model to predict the type of Aoshi ash anomaly area;

(4) Draw the distribution map of the type of Austrian lime anomaly area, and judge the distribution range of the type of Austrian lime karst anomaly area;

(5) Analyze the direction of the seepage field of the Austrian limestone fissure water network.

The indicators closely related to the network distribution of Austrian lime karst fissure water in the step (1) refer to indicators that can reflect the development degree of karst fractures, karst channels and water-rich degree of Austrian limestone, specifically including fault influence factors and fault fractal values , fold fractal dimension value, the abnormal change value of Austrian ash water temperature and the water inflow of the underground Austenitic ash hydrological borehole.

The formula for calculating the abnormal change value of Austrian ash water temperature is as follows:

ΔT=|T-t|,

In the formula: ΔT is the abnormal change value of water temperature, in °C; T is the measured water temperature at this point, in °C; t is the normal temperature calculated according to the geothermal gradient, in °C; where t is calculated by the following formula:

$\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; t</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; +</span>\frac{\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; h</span>}}{\mathrm{<span\; class="notranslate">\; \&Delta;t</span>}}<span\; class="notranslate">\; ,</span>$

In the formula: t′ is the temperature of the constant temperature zone in the study area, in °C; H is the elevation of the Austrian ash roof, in m; h is the elevation of the constant temperature zone, in m; Δt is the geothermal gradient in the study area, in °C/100m.

The establishment of nuclear principal component model of described step (2), comprises the following steps:

① Record the raw data of 5 indicators collected from one underground austrian ash hydrological borehole with geophysical detection as a (l×5) dimensional raw data matrix A;

② Through nonlinear mapping Map the original data matrix A to a high-dimensional feature space, and calculate the kernel matrix K, K=(k _ij ) _l×l , _kij =K( _xi ,x _j ), (i,j=1,2, ...,l), l is the number of indicators; where the non-mapping function The kernel function of is Gaussian radial basis function;

③According to the equation lλα=Kα, obtain the eigenvalues λ ₁ ≤λ ₂ ≤...≤λ _l and the corresponding eigenvectors α ₁ , α ₂ ,...,α _l of the kernel matrix K, and through orthogonalization The method unit orthogonalizes the eigenvectors to obtain the normalized eigenvectors α′ ₁ ,α′ ₂ ,...,α′ _l ;

④According to the formula Select m largest eigenvalues λ ₁ , λ ₂ ,...,λ _m and corresponding eigenvectors α′ ₁ , α′ ₂ ,...,α′ _m ; where, 0<m<l;

⑤ Calculate the eigenvector Y=Kα′ obtained from the original data after dimensionality reduction by KPCA, where α′=[α′ ₁ ,α′ ₂ ,...,α′ _m ], Y is the sample data matrix after dimensionality reduction ;

The fuzzy standardization of described step (2), standardization formula is:

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; min</span>}\mathrm{<span\; class="notranslate">\; x</span>}}{\mathrm{<span\; class="notranslate">\; max</span>}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; min</span>}\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; .</span>$

The geophysical detection of the step (2) includes the types of austrian limestone anomaly areas, including strong anomaly areas, weak anomaly areas and no anomaly areas. The sample label of the strong anomaly area is set to 1, and the sample label of the weak anomaly area is set to 0. The sample label of the abnormal area is set to -1.

The judgment method of the type distribution range of the Austrian lime karst anomaly area in the step (4) refers to the type distribution map of the Austrian lime anomaly area drawn, wherein the strong anomaly area, weak anomaly area, and no anomaly area are respectively represented by 1, 0, - 1 means that using the Crick intermediate interpolation method to draw the boundary line I (0.5 line) between the strong anomaly area and the weak anomaly area, and the boundary line II (-0.5 line) between the weak anomaly area and the non-anomaly area, it is located at >0.5 The area of the line is the distribution area of fissure water network in Austrian lime karst, the area between 0.5 line and -0.5 line is the distribution area of fissure water network, and the area located at <-0.5 line is the area where karst and fractures are not developed.

The method for analyzing the basic direction of the seepage field of the Austrian lime karst fissure water network in the step (5) is: draw the contour line of the Austrian lime water level, and the direction from the high water level to the low water level is the basic direction of the seepage field.

Compared with the prior art, the present invention has the following advantages:

(1) Select five indicators including fault influence factor, fault fractal dimension value, fold fractal dimension value, Austrian ash water temperature anomaly change value and downhole Austrian ash hydrological drilling water inflow. The selected parameters are easy to obtain, extensive, and do When it comes to quantification, it is possible to comprehensively evaluate the three factors of karst fissure development, karst channels, and water-richness of Austrian ash water; comprehensively use the results of geophysical detection of Austrian ash anomalies to avoid blindness and subjectivity in judging the distribution of groundwater networks.

(2) Kernel principal component analysis combines kernel function with principal component analysis, uses nonlinear method to extract principal components, improves data quality, effectively reduces the influence of redundant information, and has a more significant effect than traditional principal component analysis; Fuzzy standardization is carried out on the component analysis results to eliminate the influence of inconsistency in data scales; finally, the SVC classification model is used to predict the distribution of Ordolime Karst fissure water network. The design principle is reliable, the prediction method is simple, the prediction accuracy is high, and the prediction environment is friendly.

Description of drawings

Fig. 1 is the specific flowchart of the inventive method;

Fig. 2 is a flow chart of establishing a prediction model of genetic algorithm (GA) optimized SVM.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

As shown in Figure 1, a method for predicting the lateral distribution of karst fissure water networks in Austrian limestone includes the following steps:

(1) Determine the indicators that are closely related to the distribution of the Austrian lime karst fracture water network, and then collect the original data of the indicators at the underground Austrian limestone hydrological boreholes with geophysical exploration.

Among them, the indicators closely related to the distribution of Austrian lime karst fracture water network refer to indicators that can reflect the development degree of karst fractures, karst channels and water-rich degree of Austrian limestone, including fault influence factors, fault fractal dimension values, and fold fractal dimension values. , the abnormal change value of the Austrian ash water temperature and the water inflow of the underground Aoshan ash hydrological drilling hole;

The formula for calculating the abnormal change value of Austrian ash water temperature is as follows:

ΔT=|T-t|,

In the formula: ΔT is the abnormal change value of water temperature, °C; T is the measured water temperature at this point, °C; t is the normal temperature calculated according to the geothermal gradient, °C; where t is calculated by the following formula:

$\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; t</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; +</span>\frac{\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; h</span>}}{\mathrm{<span\; class="notranslate">\; \&Delta;</span>}}<span\; class="notranslate">\; ,</span>$

In the formula: t′ is the temperature of the constant temperature zone in the study area, ℃; H is the elevation of the Austrian ash roof, m; h is the elevation of the constant temperature zone, in m; Δt is the geothermal gradient in the study area, ℃/100m.

(2) Establish KPCA-Fuzzy-GA-SVM forecasting model for the austrian limestone anomaly area: establish a kernel principal component model (KPCA) for the collected index raw data, extract new principal components, and then carry out fuzzy standardization (Fuzzy), The sample set is composed of the principal component data after fuzzy normalization and the types of geophysically detected austrian lime karst anomalies, and a genetic algorithm (GA) is used to optimize the prediction model of SVM.

The described establishment of kernel principal component model comprises the following steps:

① Record the 5-indicator raw data collected from one downhole austenitic hydrological borehole with geophysical detection as a (1×5) dimensional raw data matrix A;

② Through nonlinear mapping Map the original data matrix A to a high-dimensional feature space, and calculate the kernel matrix K, K=(k _ij ) _l×l , _kij =K( _xi ,x _j ), (i,j=1,2, ...,l), l is the number of indicators; where the non-mapping function The kernel function of is Gaussian radial basis function;

③According to the equation lλα=Kα, obtain the eigenvalues λ ₁ ≤λ ₂ ≤...≤λ _l and the corresponding eigenvectors α ₁ , α ₂ ,...,α _l of the kernel matrix K, and through orthogonalization The method unit orthogonalizes the eigenvectors to obtain the normalized eigenvectors α′ ₁ ,α′ ₂ ,...,α′ _l ;

④According to the formula Select m largest eigenvalues λ ₁ , λ ₂ ,...,λ _m and corresponding eigenvectors α′ ₁ , α′ ₂ ,...,α′ _m ; where, 0<m<l;

⑤ Calculate the eigenvector Y=Kα′ obtained from the original data after dimensionality reduction by KPCA, where α′=[α′ ₁ ,α′ ₂ ,...,α′ _m ], Y is the sample data matrix after dimensionality reduction .

Described fuzzy standardization, standardization formula is:

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; min</span>}\mathrm{<span\; class="notranslate">\; x</span>}}{\mathrm{<span\; class="notranslate">\; max</span>}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; min</span>}\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; .</span>$

The geophysical detection austrian lime karst anomaly type includes strong anomaly area, weak anomaly area and no anomaly area, the sample label of strong anomaly area is set to 1, the sample label of weak anomaly area is set to 0, and the sample label of no anomaly area is set to -1.

Described establishment genetic algorithm (GA) optimizes the predictive model of SVM, namely first utilizes genetic algorithm to optimize the penalty parameter C of SVM model and kernel function parameter σ (kernel parameter of SVM model selects RBF kernel function) to optimize, then utilize optimal parameter Perform SVM modeling. Fig. 2 is the flowchart of establishing the predictive model of Genetic Algorithm (GA) optimization SVM, comprises the following steps:

① Sample set setting: For the sample set described in step (2), 20% of the samples are randomly selected as test samples, and the remaining samples are used as training samples; Limekarst anomaly type as target vector;

② Genetic algorithm optimization: use genetic algorithm to determine the penalty parameter C and kernel function parameter σ;

③SVM training: input training samples, use the found optimal parameters for SVM training, and establish an SVM model;

④ Model inspection: Use test samples to test the prediction model. If the prediction model accuracy reaches more than 85%, the prediction model is qualified and can be applied; if the prediction model accuracy is less than 85%, the core principal component modeling should be performed again.

(3) Collect the raw index data of the downhole Ordos ash hydrological borehole without geophysical detection, and use the established KPCA-Fuzzy-GA-SVM model to predict the type of Oro ash anomaly area.

The specific implementation method is, firstly, for the index raw data at the downhole Austrian ash hydrological borehole without geophysical detection, use the core principal component model established in step (2) to calculate the principal component data, and then carry out fuzzy normalization, and the fuzzy normalized The principal component data is used as an input parameter, and the genetic algorithm (GA) established in step (2) is input to optimize the prediction model of SVM to predict the type of ash anomaly area.

(4) Draw the distribution map of the types of Austrian ash anomalous areas, and judge the distribution range of the types of Austrian limestone anomaly areas.

The specific judging method is as follows: draw the type distribution map of the Austrian ash anomaly area, in which the strong anomaly area, weak anomaly area, and no anomaly area are represented by 1, 0, and -1 respectively, and use the Crick intermediate interpolation method to draw the strong anomaly area and The boundary line Ⅰ (0.5 line) of the weak anomaly area, and the boundary line II (-0.5 line) between the weak anomaly area and no anomaly area, the area above the 0.5 line is the distribution area of the Austrian limestone fracture water network, and the 0.5 line ~ The area between the -0.5 line is the distribution area of the fissure water network, and the area below the -0.5 line is the area where karst and fractures are not developed.

(5) Analyze the direction of the seepage field of the Austrian limestone fissure water network.

The specific method is: draw the contour line of Austrian ash water level, and the direction from the high water level to the low water level is the basic direction of the seepage field.

Although the specific implementation of the present invention has been described above in conjunction with the accompanying drawings, it does not limit the protection scope of the present invention. Those skilled in the art should understand that on the basis of the technical solution of the present invention, those skilled in the art do not need to pay creative work Various modifications or variations that can be made are still within the protection scope of the present invention.

Claims (8)

1. A method for predicting lateral distribution of karst fissure water networks in Austrian limestone, characterized in that it comprises the following steps: (1) Determine the indicators closely related to the distribution of the Austrian lime karst fracture water network, and then collect the original data of the indicators at the underground Austrian limestone hydrological borehole with geophysical detection; (2) Establish KPCA-Fuzzy-GA-SVM forecasting model of austrian limestone anomaly area: establish a nuclear principal component model (KPCA) for the collected index raw data, extract new principal components, and then carry out fuzzy standardization (Fuzzy) to fuzzy The standardized principal component data and the type of geophysical detection austrian limestone anomaly area constitute a sample set, and a genetic algorithm, namely GA optimized SVM prediction model is established; (3) Collect the original data of the indicators at the downhole Ordos ash hydrological borehole without geophysical detection, and use the established KPCA–Fuzzy-GA-SVM model to predict the type of Aoshi ash anomaly area; (4) Draw the distribution map of the type of Austrian lime anomaly area, and judge the distribution range of the type of Austrian lime karst anomaly area; (5) Analyze the direction of the seepage field of the Austrian limestone fissure water network. 2. the method for predicting the lateral distribution of the Austrian limestone fissure water network as claimed in claim 1, is characterized in that, the index closely related with the Austrian limestone fracture water network distribution of described step (1), refers to the index that can reflect karst Indexes of fracture development degree, karst channel and water-rich degree of Austrian ash, including five indicators of fault influence factor, fault fractal dimension value, fold fractal dimension value, abnormal change value of Austrian ash water temperature and water inflow of underground Ordos ash hydrological drilling data. 3. the prediction method of lateral distribution of Austrian limestone fissure water network as claimed in claim 2, it is characterized in that, wherein the calculation formula of the abnormal change value of Austrian limestone water temperature is as follows: ΔT=|T-t|, In the formula: ΔT is the abnormal change value of water temperature, in °C; T is the measured water temperature at this point, in °C; t is the normal temperature calculated according to the geothermal gradient, in °C; where t is calculated by the following formula: $\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; t</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; +</span>\frac{\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; h</span>}}{\mathrm{<span\; class="notranslate">\; \&Delta;t</span>}}<span\; class="notranslate">\; ,</span>$ In the formula: t′ is the temperature of the constant temperature zone in the study area, in °C; H is the elevation of the Austrian ash roof, in m; h is the elevation of the constant temperature zone, in m; Δt is the geothermal gradient in the study area, in °C/100m. 4. the method for predicting lateral distribution of the Austrian limestone karst fissure water network as claimed in claim 1, is characterized in that, the establishment core principal component model of described step (2), comprises the following steps: ① Record the raw data of 5 indicators collected from one underground austrian ash hydrological borehole with geophysical detection as a (l×5) dimensional raw data matrix A; ② Through nonlinear mapping Map the original data matrix A to a high-dimensional feature space, and calculate the kernel matrix K, K=(k _ij ) _l×l , _kij =K( _xi ,x _j ), (i,j=1,2, ...,l), l is the number of indicators; where the non-mapping function The kernel function of is Gaussian radial basis function; ③According to the equation lλα=Kα, obtain the eigenvalues λ ₁ ≤λ ₂ ≤...≤λ _l and the corresponding eigenvectors α ₁ , α ₂ ,...,α _l of the kernel matrix K, and through orthogonalization The method unit orthogonalizes the eigenvectors to obtain the normalized eigenvectors α′ ₁ ,α′ ₂ ,...,α′ _l ; ④According to the formula Select m largest eigenvalues λ ₁ , λ ₂ ,...,λ _m and corresponding eigenvectors α′ ₁ , α′ ₂ ,...,α′ _m ; where, 0<m<l; ⑤ Calculate the eigenvector Y=Kα′ obtained from the original data after dimensionality reduction by KPCA, where α′=[α′ ₁ ,α′ ₂ ,...,α′ _m ], Y is the sample data matrix after dimensionality reduction . 5. the method for predicting lateral distribution of Austrian limestone fissure water network as claimed in claim 1, is characterized in that, the fuzzy standardization of described step (2), standardization formula is: ${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; min</span>}\mathrm{<span\; class="notranslate">\; x</span>}}{\mathrm{<span\; class="notranslate">\; max</span>}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; min</span>}\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; .</span>$ 6. the method for predicting lateral distribution of the Austrian lime karst fissure water network as claimed in claim 1, is characterized in that, the geophysics detection Austrian lime karst anomaly area type of described step (2), comprises strong anomaly area, weak anomaly area and no anomaly area, set the sample label of strong anomaly area to 1, the sample label of weak anomaly area to 0, and the sample label of no anomaly area to -1. 7. the method for predicting lateral distribution of the Austrian limestone fissure water network as claimed in claim 1, is characterized in that, the judgment method of the type distribution scope of the Austrian limestone karst anomaly area of described step (4) refers to, the Austrian limestone of drawing Anomaly area type distribution map, where strong anomaly area, weak anomaly area, and no anomaly area are represented by 1, 0, and -1 respectively. Using the Crick intermediate interpolation method, the boundary line I between strong anomaly area and weak anomaly area is drawn, which is 0.5 line, and the boundary line II between the weak anomaly area and the no anomaly area, i.e. the -0.5 line, the area above the 0.5 line is the distribution area of the Austrian limestone fissure water network, and the area between the 0.5 line and the -0.5 line is the fissure water network The distribution area, the area located at the <-0.5 line is the area where karst and fractures are not developed. 8. the prediction method of lateral distribution of Austrian limestone fissure water network as claimed in claim 1, it is characterized in that, the method for the analysis of Austrian limestone fracture water network seepage field basic direction of described step (5) is: draw Austrian limestone The water level isoline, the direction from the high water level to the low water level is the basic direction of the seepage field.