Deep brine reserve calculation method based on Thiessen polygon and Monte Carlo simulation

Abstract

The invention discloses a deep brine reserve calculation method based on Thiessen polygon and Monte Carlo simulation, and the method comprises the steps: obtaining drilling position data and brine parameter data of a sulfate type salt lake, constructing a Thiessen polygon division graph, and generating a plurality of regional polygons; based on the region polygon, determining a brine parameter value range and parameter frequency distribution of each region, and generating a brine parameter probability distribution model; performing global sensitivity analysis based on the brine parameter probability distribution model, determining key sensitive parameters, and generating a sensitive parameter evaluation result; based on the sensitive parameter evaluation result and the regional polygon, calculating brine resource reserves of each region by adopting a Monte Carlo simulation method, and generating regional brine resource reserve distribution; and based on regional brine resource reserve distribution, summarizing and calculating the total brine resource reserve of the whole region, and generating a deep brine reserve calculation result. The limitation of a traditional method is overcome, and reliable technical support is provided for evaluation and development of deep brine resources.

Description

Technical Field

[0001] The present invention relates to the technical field of mineral resource reserve calculation, and in particular to a deep brine reserve calculation method based on Thiessen polygons and Monte Carlo simulation. Background Art

[0002] In recent years, with the rapid development of the new energy industry, shallow brine resources have been unable to meet the growing demand. Deep brine, rich in important elements such as lithium, potassium, magnesium, and bromine, has higher economic value and utilization potential. However, the occurrence conditions of deep brine resources are complex and are affected by multiple factors such as tectonic setting, sedimentary environment, and hydrogeological conditions, which poses significant challenges to resource investigation and assessment. Although high-precision geological, geophysical, and geochemical exploration methods can provide more comprehensive underground information, specific reserve assessments still face problems such as large uncertainty and limited data, which cannot provide sufficient and accurate support for resource development decisions.

[0003] Existing assessment methods primarily rely on geophysical exploration and drilling results, such as seismic, electrical, and gravity / magnetic surveys. These methods are supplemented by geostatistical methods such as interpolation and averaging to estimate reserves from drilling data. However, when drillhole density is insufficient or data distribution is uneven, interpolation often produces a significant smoothing effect and is easily affected by local outliers. Mean and median methods also struggle to reflect the true spatial variation of resources, resulting in inaccurate assessment results. Furthermore, geophysical exploration itself is affected by the complexity and noise of the subsurface medium, making prediction accuracy difficult to guarantee. This leads to consistently high uncertainty in reserve calculations.

[0004] Single geophysical or statistical analysis methods struggle to fully utilize existing drilling data while simultaneously assessing spatial variability and uncertainty. Existing technologies for processing deep brine resources urgently require a computational approach that can effectively partition the spatial domain, reduce smoothing effects, and quantify uncertainty through simulation. Summary of the Invention

[0005] The present invention provides a deep brine reserves calculation method based on Thiessen polygons and Monte Carlo simulation to solve the above problems existing in the prior art.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation, characterized by comprising:

[0008] S1: Obtain the sulfate salt lake drilling location data and brine parameter data, construct a Thiessen polygon partitioning map, and generate several regional polygons;

[0009] S2: Based on the regional polygons, determine the brine parameter value range and parameter frequency distribution of each region, and generate a brine parameter probability distribution model;

[0010] S3: Based on the brine parameter probability distribution model, conduct global sensitivity analysis, determine key sensitive parameters, and generate sensitive parameter evaluation results;

[0011] S4: Based on the evaluation results of sensitive parameters and regional polygons, the Monte Carlo simulation method is used to calculate the brine resource reserves in each region and generate the regional brine resource reserve distribution;

[0012] S5: Based on the regional brine resource reserves distribution, the total brine resource reserves of the entire region are calculated and the deep brine reserves calculation results are generated.

[0013] Among them, step S1 includes:

[0014] S11: Collect the location coordinate data of all drill holes in the study area;

[0015] S12: Based on the drilling location coordinate data, construct the Thiessen polygon Voronoi diagram to generate the initial area polygon;

[0016] S13: Optimize the initial region polygon to generate the final region polygon;

[0017] Specifically, the weight coefficient of each borehole is calculated based on the brine parameter data at the borehole, which includes the area of ​​the ore layer, the thickness of the ore layer, the brine density, the porosity and the brine grade, to reflect the spatial heterogeneity;

[0018] Generate a weighted Voronoi diagram based on the calculated weight coefficient to form a regional polygon;

[0019] Apply the convex hull algorithm or the minimum spanning tree algorithm to adjust the boundaries of the polygons to ensure that the area of ​​each polygon is within the preset range and the shape is a regular polygon.

[0020] Among them, step S12 includes:

[0021] S121: Calculate the Euclidean distance from each borehole to all other boreholes using the Euclidean distance formula;

[0022] S122: Find the nearest neighbor borehole of each borehole based on Euclidean distance;

[0023] S123: Connect adjacent drill holes and draw corresponding perpendicular bisectors;

[0024] S124: The area enclosed by the perpendicular bisectors forms a Voronoi diagram area.

[0025] The S2 step includes:

[0026] S21: Obtain brine parameter data of wells drilled within each regional polygon, including ore layer area, ore layer thickness, brine density, porosity and brine grade;

[0027] S22: Determine the value range of each parameter based on geological data and expert evaluation;

[0028] S23: Based on historical data, experience from similar areas and geostatistical methods, determine the frequency distribution type of each parameter and generate a probability distribution model for brine parameters.

[0029] The S3 step includes:

[0030] S31: Based on the brine parameter probability distribution model, a Sobol sequence is generated to uniformly sample the parameter space. The Sobol sequence represents a low-discrepancy sequence used to uniformly cover the parameter space.

[0031] S32: Calculate the variance contribution of each parameter to the reserve calculation result using the variance decomposition formula;

[0032] S33: Calculate the Sobol sensitivity index of each parameter based on the variance contribution;

[0033] Among them, the Sobol sensitivity index S i Defined as:

[0034]

[0035] Among them, Vi represents the contribution of the i-th parameter to the output variance, and V0 is the total variance of the model output;

[0036] S34: Select parameters whose sensitivity index is higher than the preset index as key sensitive parameters and generate sensitive parameter evaluation results.

[0037] Among them, step S4 includes:

[0038] S41: Randomly sample key sensitive parameters within the corresponding probability distribution range to generate multiple groups of parameter combinations;

[0039] S42: Substitute each set of parameter combinations into the resource reserve calculation model to calculate the resource reserves of a single area;

[0040] S43: Repeat steps S41 and S42 multiple times until the simulation results converge to the preset convergence conditions, and generate the regional brine resource reserves distribution.

[0041] The resource reserve calculation model in step S42 is:

[0042] P kL=S×M×D×K×C

[0043] Among them, P kL It represents the available resources of liquid potash ore; S represents the distribution area of ​​the brine ore layer to be calculated; M represents the average effective thickness of the ore layer; D represents the average density of the brine ore layer; K represents the average porosity; C represents the average grade.

[0044] The simulation results in step S43 converge to the preset convergence conditions, which include:

[0045] Calculate whether the standard deviation of the iteration result is less than the preset threshold and tends to zero;

[0046] Or calculate whether the relative difference between the result of the nth iteration and the average of the results of the previous n-1 iterations is less than a preset threshold;

[0047] Or observe whether the width of the 95% confidence interval of the simulation results narrows to the preset range as the number of iterations increases;

[0048] Or observe whether the change range of key sensitive parameters is less than the preset threshold as the number of iterations increases.

[0049] Among them, step S5 includes:

[0050] S51: Summarize the resource reserves within all regional polygons to obtain the total resource volume of the study area;

[0051] S52: performing statistical analysis on all calculated resource quantities to obtain a probability distribution of the resource quantities;

[0052] S53: 95% confidence interval for the determined resource;

[0053] S54: Calculate the standard deviation of the resource volume, evaluate the uncertainty of the resource volume, and generate the deep brine reserve calculation results.

[0054] Among them, also include:

[0055] According to the uncertainty assessment results, adjust the regional polygon division or parameter value range, and re-execute steps S1 to S5 until the uncertainty assessment results meet the preset conditions;

[0056] Based on the final calculation results of deep brine reserves, the development potential of deep brine resources is analyzed to provide a basis for resource development decisions.

[0057] Compared with the prior art, the present invention has the following advantages:

[0058] A deep brine reserve calculation method based on Thiessen polygons and Monte Carlo simulations includes the following steps: S1: Obtaining sulfate lake borehole location data and brine parameter data, constructing a Thiessen polygon partitioning map, and generating several regional polygons; S2: Based on the regional polygons, determining the brine parameter value range and parameter frequency distribution for each region, and generating a brine parameter probability distribution model; S3: Based on the brine parameter probability distribution model, performing a global sensitivity analysis, identifying key sensitive parameters, and generating sensitive parameter assessment results; S4: Based on the sensitive parameter assessment results and the regional polygons, using Monte Carlo simulation, calculating the brine resource reserves in each region and generating a regional brine resource distribution; S5: Based on the regional brine resource distribution, summarizing and calculating the total brine resource reserves for the entire region and generating a deep brine reserve calculation result. Calculations based on borehole data avoid the risk of using unverified values ​​in interpolation methods in areas without boreholes, ensuring the reliability of the results. By dividing the study area into multiple blocks, even if drilling errors or mistakes cause problems in one block, the impact on other blocks is minimal, thus reducing overall error. Integrating prior knowledge of the region into the analysis, rather than relying solely on drilling data, improves the accuracy of the assessment. Through random sampling and Monte Carlo simulation, the uncertainty of resource quantities can be estimated probabilistically, providing a scientific basis for decision-making in resource development.

[0059] Other features and advantages of the present invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the present invention.

[0060] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0062] Figure 1 This is a flow chart of a method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation in an embodiment of the present invention;

[0063] Figure 2 A flowchart of generating a plurality of area polygons in an embodiment of the present invention;

[0064] Figure 3 This is a flow chart of generating a brine parameter probability distribution model in an embodiment of the present invention. DETAILED DESCRIPTION

[0065] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0066] The embodiment of the present invention provides Figure 1 As shown, a deep brine reserve calculation method based on Thiessen polygons and Monte Carlo simulation includes:

[0067] S1: Obtain the sulfate salt lake drilling location data and brine parameter data, construct a Thiessen polygon partitioning map, and generate several regional polygons;

[0068] S2: Based on the regional polygons, determine the brine parameter value range and parameter frequency distribution of each region, and generate a brine parameter probability distribution model;

[0069] S3: Based on the brine parameter probability distribution model, conduct global sensitivity analysis, determine key sensitive parameters, and generate sensitive parameter evaluation results;

[0070] S4: Based on the evaluation results of sensitive parameters and regional polygons, the Monte Carlo simulation method is used to calculate the brine resource reserves in each region and generate the regional brine resource reserve distribution;

[0071] S5: Based on the regional brine resource reserves distribution, the total brine resource reserves of the entire region are calculated and the deep brine reserves calculation results are generated.

[0072] The working principle of the above technical solution is: collect the drilling location data and brine parameter data of sulfate salt lakes, where the brine parameter data includes the latitude and longitude coordinates of the drilling hole and the corresponding brine characteristics (such as concentration, ore layer thickness, etc.). Use the Voronoi diagram to construct Thiessen polygons. The Voronoi diagram divides the study area into multiple polygons, each polygon corresponds to a drilling location, reflecting the influence range of the drilling hole. In order to improve the accuracy of the model, spatial heterogeneity (such as geological structure and lithology differences), drilling density and polygon shape optimization need to be considered when constructing polygons.

[0073] After constructing the Thiessen polygons, the range of brine parameters in each area is determined. The borehole data within each polygon is analyzed and combined with geological data to determine the reasonable range of parameter values.

[0074] A global sensitivity analysis was conducted based on the probability distribution model of brine parameters. The Sobol method was used to assess the impact of each input parameter on the output. Key sensitive parameters were identified by calculating the variance contribution of each parameter.

[0075] After determining key sensitive parameters, the Monte Carlo simulation method was used to calculate resource reserves. For each sensitive parameter, random sampling was performed within its prior distribution to generate multiple parameter combinations. These parameter combinations were then substituted into the resource reserve calculation model to calculate the brine resource in each area. Through multiple iterations and gradual convergence, the probability distribution of the resource and the uncertainty assessment results were finally obtained.

[0076] The brine resources in all regions were aggregated to calculate the total reserves for the entire study area. Statistical analysis of all calculated resources was performed to obtain a probability distribution and determine confidence intervals to assess the uncertainty of the resources. This process provides a scientific basis for analyzing the development potential of deep brine resources.

[0077] The beneficial effect of this technical solution is that it uses Thiessen polygons to spatially partition borehole data and, combined with Monte Carlo simulation, generates multiple possible resource distribution models, thereby comprehensively assessing the uncertainty of deep brine resources. This method fully considers the spatial location and correlation of borehole data, effectively solving the prediction problem when data is insufficient or unevenly distributed.

[0078] In another embodiment, Figure 2 As shown, step S1 includes:

[0079] S11: Collect the location coordinate data of all drill holes in the study area;

[0080] S12: Based on the drilling location coordinate data, construct the Thiessen polygon Voronoi diagram to generate the initial area polygon;

[0081] S13: Optimize the initial region polygon to generate the final region polygon;

[0082] Specifically, the weight coefficient of each borehole is calculated based on the brine parameter data at the borehole, which includes the area of ​​the ore layer, the thickness of the ore layer, the brine density, the porosity and the brine grade, to reflect the spatial heterogeneity;

[0083] Generate a weighted Voronoi diagram based on the calculated weight coefficient to form a regional polygon;

[0084] Apply the convex hull algorithm or the minimum spanning tree algorithm to adjust the boundaries of the polygons to ensure that the area of ​​each polygon is within the preset range and the shape is a regular polygon.

[0085] The working principle of the above technical solution is: collect the location information of all boreholes in the study area, the location information includes the longitude and latitude of each borehole, or the X and Y coordinates in the plane coordinate system, and store the data in a data structure, such as an array or a list.

[0086] Using the collected borehole coordinate data, construct an initial Voronoi diagram. The Voronoi diagram divides a plane into multiple regions, each corresponding to a borehole. All points in the region are closer to that borehole than to other boreholes. The specific steps are as follows:

[0087] For each borehole, calculate the distance to all other boreholes;

[0088] Find the nearest neighbor of each borehole and connect two adjacent boreholes to form connecting lines, which are the perpendicular bisectors of the two boreholes;

[0089] Through these connecting lines, preliminary Voronoi polygons are formed;

[0090] Based on the brine parameter data at the borehole (such as ore area, ore thickness, brine density, porosity, and brine grade), a weight coefficient is calculated for each borehole. These weight coefficients will be used to weight the Voronoi diagram to better reflect the characteristics of different areas. After calculating the weight coefficients, we apply these weights to the construction of the Voronoi diagram. Specifically, a weighted distance metric is used to re-divide the Voronoi region so that the shape and size of each polygon better reflect the characteristics of its corresponding borehole.

[0091] To ensure that the area of ​​each Voronoi polygon is within the preset range and the shape is a regular polygon, we will apply the convex hull algorithm or the minimum spanning tree algorithm for optimization. The specific process is as follows:

[0092] Convex Hull Algorithm: Determines the smallest convex polygon that encompasses all drill holes, ensuring the polygon's boundary is minimal and regular. Minimum Spanning Tree Algorithm: Connects the minimum boundaries of all drill holes to form a tree-like structure, which then adjusts the polygon's boundary to ensure the reasonableness of the shape. Finally, the optimized Voronoi polygon is obtained, which is the final region polygon.

[0093] The beneficial effects of this technical solution include: by optimizing the regional division of Thiessen polygons and incorporating brine parameters such as ore seam area, ore seam thickness, and brine density, spatial heterogeneity is reflected, thereby improving the accuracy of regional polygon division. Adjusting the boundaries using weighted Voronoi diagrams and convex hull algorithms not only ensures that the area of ​​each region is within the preset range, but also avoids potential errors caused by relying solely on drill hole data.

[0094] In another embodiment, step S12 includes:

[0095] S121: Calculate the Euclidean distance from each borehole to all other boreholes using the Euclidean distance formula;

[0096] S122: Find the nearest neighbor borehole of each borehole based on Euclidean distance;

[0097] S123: Connect adjacent drill holes and draw corresponding perpendicular bisectors;

[0098] S124: The area enclosed by the perpendicular bisectors forms a Voronoi diagram area.

[0099] The working principle of the above technical solution is: for each drill hole, we need to obtain its coordinate information, assuming the coordinates of the drill holes are (x1, y1) and (x2, y2). Use the Euclidean distance formula to calculate the distance between two drill holes:

[0100]

[0101] The meaning of this formula is to determine the straight-line distance between two boreholes by calculating the coordinate difference between them on the plane. The purpose of this step is to provide basic data for subsequent nearest neighbor search.

[0102] After calculating the distances between all the holes, we compare each hole's distance to all other holes and find the hole with the smallest distance. The hole with the smallest distance is the nearest neighbor of the current hole. This step works by iterating through all the holes and recording the smallest distance and its corresponding hole index.

[0103] Once each pair of adjacent drill holes is determined, a line segment is drawn by connecting the coordinates of the two drill holes. Next, we need to calculate the perpendicular bisector of this line segment. The equation of the perpendicular bisector is:

[0104]

[0105] This equation represents the line that passes through the midpoints of the two holes and is perpendicular to the connecting line. The purpose of this step is to define the region boundaries in the subsequent Voronoi diagram.

[0106] By connecting the perpendicular bisectors of all adjacent boreholes, a region is formed. Each region corresponds to a borehole, and the distance from each point in the region to the borehole is smaller than the distance to all other boreholes.

[0107] To analyze and process these areas, Delaunay triangulation is performed. This connects the drill hole locations into triangles, ensuring that the circumcircle of each triangle contains no other drill holes. This process not only optimizes space utilization but also provides a more detailed structure for subsequent resource analysis.

[0108] The beneficial effects of this technical solution are: Euclidean distance calculation and the nearest-neighbor borehole partitioning method make regional division more scientific and reasonable. By drawing perpendicular bisectors to form Voronoi diagram regions, the spatial relationships between regions are more closely aligned and the problem of inaccurate data transfer between boreholes is avoided. This method can more accurately reflect the spatial distribution of boreholes and reduce the errors caused by traditional interpolation methods.

[0109] In another embodiment, Figure 3 As shown, step S2 includes:

[0110] S21: Obtain brine parameter data of wells drilled within each regional polygon, including ore layer area, ore layer thickness, brine density, porosity and brine grade;

[0111] S22: Determine the value range of each parameter based on geological data and expert evaluation;

[0112] S23: Based on historical data, experience from similar areas and geostatistical methods, determine the frequency distribution type of each parameter and generate a probability distribution model for brine parameters.

[0113] The working principle of the above technical solution is as follows: In step S21, the drilling data is divided into geographical regions. Using the Voronoi diagram or Delaunay triangulation method, the regions are divided into multiple polygonal or triangular blocks, each representing a geological region. Each vertex of these polygonal or triangular regions represents a drilling location, and each region contains brine parameter data for one or more wells. At each drilling point, multiple key brine-related parameters are collected: ore layer area, ore layer thickness, brine density, porosity, and brine grade. Specific values ​​for these parameters are assigned to each well. These parameters are then distributed within the corresponding Voronoi regions according to their spatial location.

[0114] In step S22, when determining the value range of each parameter, the geological data in the area (including stratigraphic profiles and lithologic analysis) must be combined. Through these data, the geological structure of the area, the thickness of the ore layer, the lithologic changes, etc. are determined to provide a basis for determining the range of each parameter. Geological experts conduct geological analysis, and based on their experience and the results of on-site surveys, they estimate the reasonable range of each parameter (such as ore layer thickness, brine grade, etc.). Expert evaluation and geological data analysis will provide a range for each parameter. For example, the thickness of the ore layer is between 10 and 50 meters, and the brine density is between 1.2 and 1.5. These parameter ranges will be used to limit subsequent resource assessment calculations.

[0115] When determining the frequency distribution of each parameter in step S23, historical drilling data in the area will be collected first, and the distribution of parameters will be analyzed with the help of experience from other similar areas. This analysis process usually involves statistical methods, such as mean, standard deviation, and coefficient of variation, to help understand the range and trend of parameter values ​​in various regions. Based on the statistical results of historical data, the corresponding probability distribution type can be selected to describe the changes in each parameter. For example, the thickness of the ore layer conforms to the normal distribution, and the brine grade conforms to the log-normal distribution. Selecting the corresponding distribution type can more accurately reflect the changing patterns of brine parameters in nature. Once the frequency distribution type of each parameter is determined, a probability distribution model of the brine parameters is constructed using geostatistical methods (such as Monte Carlo simulation, Bayesian analysis, etc.).

[0116] The beneficial effects of this technical solution include: by combining geological data and expert evaluation, the range of brine parameters is determined, and a probability distribution model for these parameters is generated using historical data and geostatistical methods. This method fully utilizes prior knowledge during the evaluation process, avoiding the limitations of single drilling data, and significantly improving the accuracy of brine parameter estimation.

[0117] In another embodiment, step S3 includes:

[0118] S31: Based on the brine parameter probability distribution model, a Sobol sequence is generated to uniformly sample the parameter space. The Sobol sequence represents a low-discrepancy sequence used to uniformly cover the parameter space.

[0119] S32: Calculate the variance contribution of each parameter to the reserve calculation result using the variance decomposition formula;

[0120] S33: Calculate the Sobol sensitivity index of each parameter based on the variance contribution;

[0121] Among them, the Sobol sensitivity index S i Defined as:

[0122]

[0123] Among them, Vi represents the contribution of the i-th parameter to the output variance, and V0 is the total variance of the model output;

[0124] S34: Select parameters whose sensitivity index is higher than the preset index as key sensitive parameters and generate sensitive parameter evaluation results.

[0125] The working principle of the above technical solution is as follows: Step S31 selects a suitable probability distribution (including normal distribution, lognormal distribution, etc.) for each parameter based on prior knowledge and geostatistical methods.

[0126] Generate Sobol sequence: Number of directions: Determine the number of directions for each input parameter, typically choosing a power of 2 to generate more uniform sampling. Generator matrix: Construct a generator matrix that contains a binary representation of the number of directions for each dimension. Recursive generation: Using the generator matrix and initial values, recursively generate the Sobol sequence to ensure that the sequence has low variance and can evenly cover the parameter space. Parameter space sampling: Use the generated Sobol sequence to uniformly sample the parameters to obtain the input dataset of the model.

[0127] Step S32 uses the sampled parameter input to calculate the output of the model Y=f(X1, X2, ..., X n ), where Y is the output result, X i is the i-th input parameter. According to the variance decomposition formula, the total variance V0 of the model output is calculated and decomposed into the variance contribution V of each input parameter i and the interaction variance V between parameters ij The formula is:

[0128]

[0129] Where: V0 is the total variance of the model output; V i is the contribution of the i-th parameter to the output variance; V ij is the parameter X i and X j The contribution of the interaction of V to the output variance; 12…n Is the contribution of all parameter interactions to the output variance. By calculating the variance contribution of each input parameter V i , and evaluate its impact on the model output.

[0130] Step S33 calculates the Sobol sensitivity index Si according to the variance contribution of each parameter, and the formula is:

[0131]

[0132] Among them, V i represents the contribution of the i-th parameter to the output variance, V0 is the total variance of the model output; S i Indicates the contribution ratio of the i-th parameter to the output variance, ranging from 0 to 1. i The larger the value, the greater the impact of the parameter on the output result. Using the Sobol sequence, the sensitivity index of each parameter can be calculated by the following formula:

[0133]

[0134] Among them, E[Y|X i ] means that at a given X iIn the case of , the model outputs the expected value of Y.

[0135] In step S34, according to a preset sensitivity index threshold, parameters with sensitivity indexes higher than the threshold are screened out, and the selected key sensitive parameters are recorded to form a sensitive parameter evaluation result for subsequent model optimization and decision support.

[0136] The number of simulations affects the results of a global sensitivity analysis. A greater number of simulations results in more accurate results, but the computational effort also increases. Choosing an appropriate number of simulations depends on the specific situation. In addition to the Sobol index, other sensitivity indices, such as the Morris index and the FAST index, can also be used. Selecting an appropriate sensitivity index is crucial. Interactions between parameters should be considered, such as the potential for synergistic or antagonistic effects between two parameters. Parameter combinations can be analyzed to assess these interactions.

[0137] The beneficial effect of this technical solution is that it uniformly samples the brine parameter space through the Sobol sequence, and combined with variance decomposition and sensitivity analysis, it can identify the key sensitive parameters that have the greatest impact on reserve calculations. By calculating the Sobol sensitivity index of each parameter, key sensitive parameters are selected, providing more accurate input data for resource reserve calculations.

[0138] In another embodiment, step S4 includes:

[0139] S41: Randomly sample key sensitive parameters within the corresponding probability distribution range to generate multiple groups of parameter combinations;

[0140] S42: Substitute each set of parameter combinations into the resource reserve calculation model to calculate the resource reserves of a single area;

[0141] S43: Repeat steps S41 and S42 multiple times until the simulation results converge to the preset convergence conditions, and generate the regional brine resource reserves distribution.

[0142] The working principle of the above technical solution is as follows: In step S41, the key sensitive parameters to be evaluated are first determined. These parameters typically have a certain a priori distribution range, such as normal distribution or uniform distribution. Using a random number generator, we randomly sample within this distribution range to generate multiple sets of possible parameter combinations. Each parameter combination represents a possible scenario, reflecting the changes in resource reserves under different conditions. The purpose of this process is to ensure that all possible scenarios are considered, thereby improving the accuracy of the model.

[0143] In step S42, each parameter combination obtained through random sampling is entered into a resource reserve calculation model. This model, based on geological, physical, and chemical properties, can estimate resource reserves within a specific area. Specifically, we can use Voronoi polygons or Delaunay triangulation to divide the study area into multiple small areas and calculate the resource reserves within each small area. This process involves inputting parameter combinations into the model to obtain resource estimates for each area.

[0144] In step S43, steps S41 and S42 are repeated multiple times, performing multiple iterations. After each iteration, the sum of all Voronoi or Delaunay results is calculated and combined with the previous results to form a dataset. Convergence of the simulation results can be determined by observing whether the standard deviation gradually approaches 0. Furthermore, the mean difference between the current iteration result and the previous result is calculated as the error. If the error decreases with increasing iterations, the reliability of the results is improving.

[0145] During this phase, you can also calculate the confidence interval for the simulation results. If the confidence interval narrows with increasing iterations, it indicates that the simulation results are stabilizing and have converged. Furthermore, observing whether the changes in highly sensitive parameters become relatively stable after a certain number of iterations is also an important indicator of convergence.

[0146] There are several key factors to consider when performing Monte Carlo simulations:

[0147] Number of simulations: The more simulations you run, the more accurate the results will be, but the computational effort will also increase. Therefore, it is important to choose an appropriate number of simulations based on your actual situation to balance computational efficiency and result accuracy.

[0148] Convergence judgment: In addition to standard deviation, error and confidence interval, other indicators such as Monte Carlo standard error can also be used to judge the convergence of the simulation.

[0149] Result analysis: Statistical analysis is performed on the simulation results to obtain the probability distribution, confidence interval and uncertainty assessment of resource reserves, so as to provide a basis for subsequent decision-making.

[0150] Model validation: Ensure the accuracy and reliability of the model by validating it using historical data or independent data.

[0151] The beneficial effect of the above technical solution is that it generates a probabilistic distribution of regional brine reserves by randomly sampling parameters through Monte Carlo simulation and calculating resource reserves through multiple iterations. This method fully considers the uncertainty of parameters, avoids the limitations of traditional static methods, and provides a more comprehensive and reliable resource reserve estimate.

[0152] In another embodiment, the resource reserve calculation model in step S42 is:

[0153] P kL =S×M×D×K×C

[0154] Among them, P kL It represents the available resources of liquid potash ore; S represents the distribution area of ​​the brine ore layer to be calculated; M represents the average effective thickness of the ore layer; D represents the average density of the brine ore layer; K represents the average porosity; C represents the average grade.

[0155] The working principle of the above technical solution is as follows: When evaluating the reserves of salt lake brine potash ore, the porosity and water supply resources are calculated according to the "volumetric method". After determining the distribution range of unit water yield and the effective thickness of the ore layer, the resource reserves of a single well can be calculated according to the following formula:

[0156] P kL =S×M×D×K×C

[0157] Among them, P kL It represents the available resources of liquid potash ore, in t; S represents the distribution area of ​​brine ore layer to be calculated, in m 2 ; M represents the average effective thickness of the ore layer, in m; D represents the average density of the brine ore layer, in kg / m 3 ; K represents the average porosity, in %; C represents the average grade, in %.

[0158] The beneficial effect of the above technical solution is that, through the resource reserve calculation model, this step can quantify the available amount of liquid potash ore and accurately calculate the brine resource reserves in each area. This model not only considers parameters such as the effective thickness of the ore layer and brine density, but also incorporates actual ore layer area and grade data, thus ensuring the accuracy and reliability of the reserve calculation.

[0159] In another embodiment, the simulation result in step S43 converges to a preset convergence condition including:

[0160] Calculate whether the standard deviation of the iteration result is less than the preset threshold and tends to zero;

[0161] Or calculate whether the relative difference between the result of the nth iteration and the average of the results of the previous n-1 iterations is less than a preset threshold;

[0162] Or observe whether the width of the 95% confidence interval of the simulation results narrows to the preset range as the number of iterations increases;

[0163] Or observe whether the change range of key sensitive parameters is less than the preset threshold as the number of iterations increases.

[0164] The working principle of the above technical solution is as follows: at each iteration, the calculation results are recorded and stored in an array. The standard deviation is a statistic used to measure the dispersion of these results. Specifically, the standard deviation of the current iteration result and the results of all previous iterations is calculated, and the standard deviation is determined to be less than a preset threshold. If it is less than the threshold, and the standard deviation approaches zero as the number of iterations increases, the results are considered to have converged.

[0165] Record the results of each iteration and calculate the mean of the results of the previous n iterations. Then, calculate the relative difference between the result of the nth iteration and the mean.

[0166] By statistically analyzing the iteration results, we calculate the probability distribution of the resource volume and determine a 95% confidence interval. Specifically, after each iteration, we calculate the confidence interval for the current result and observe whether its width gradually narrows to within the preset range as the number of iterations increases. If the width narrows, the reliability of the result is improving.

[0167] In each iteration, the values ​​of key sensitive parameters are recorded and their magnitude of change is calculated. Specifically, this involves comparing the parameter values ​​of the current iteration with those of the previous iteration, calculating the absolute difference, and determining whether the difference is less than a preset threshold. If the magnitude of change decreases, the simulation results are stabilizing.

[0168] The beneficial effect of this technical solution is that the convergence criteria for the simulation results, including the standard deviation, relative difference, and 95% confidence interval width of the calculated iterative results, ensure that the final simulation results tend to be stable. This convergence criterion effectively determines the reliability of the calculation results, ensuring that the error during the model iteration process is within the preset range, thus providing a solid foundation for the final resource assessment.

[0169] In another embodiment, step S5 includes:

[0170] S51: Summarize the resource reserves within all regional polygons to obtain the total resource volume of the study area;

[0171] S52: performing statistical analysis on all calculated resource quantities to obtain a probability distribution of the resource quantities;

[0172] S53: 95% confidence interval for the determined resource;

[0173] S54: Calculate the standard deviation of the resource volume, evaluate the uncertainty of the resource volume, and generate the deep brine reserve calculation results.

[0174] The working principle of the above technical solution is: in each iteration, the current calculation results are recorded and stored in a result array. The standard deviation is a statistic used to measure the degree of dispersion of these results. Specifically, the standard deviation is calculated using a formula. If the calculated standard deviation is less than a preset threshold and tends to zero over multiple iterations, the simulation results are considered to have converged.

[0175] After each iteration, the relative difference between the current iteration result and the mean of the previous n-1 iterations is calculated. Specifically, the mean of the previous n-1 iterations is calculated, and then the relative difference is determined to be less than a preset threshold. If the relative difference is less than the preset threshold, the results are stable and the simulation has converged.

[0176] At each iteration, a 95% confidence interval is calculated for the current result. A confidence interval is used to estimate the likely range of a parameter (such as resource quantity). This is done by using statistical methods to calculate the upper and lower bounds of the confidence interval and observing its width. If the width of the confidence interval gradually narrows to within the preset range as the number of iterations increases, the reliability of the simulation results is improving and the results are converging.

[0177] In each iteration, the values ​​of key sensitive parameters are recorded and their magnitude of change is calculated. Specifically, this is done by comparing the parameter values ​​of the current iteration with those of the previous iteration and calculating the absolute difference. If the magnitude of change gradually decreases and falls below a preset threshold, the simulation results are becoming more stable and convergent.

[0178] The beneficial effect of this technical solution is that by summarizing and analyzing the reserves within all regional polygons, combining statistical methods to calculate the probability distribution of the reserves, and further assessing uncertainty, the total reserves of deep brine resources are determined. This method combines multiple calculation results to provide more comprehensive resource reserves data, providing a scientific basis for subsequent resource development decisions.

[0179] In another embodiment, further comprising:

[0180] According to the uncertainty assessment results, adjust the regional polygon division or parameter value range, and re-execute steps S1 to S5 until the uncertainty assessment results meet the preset conditions;

[0181] Based on the final calculation results of deep brine reserves, the development potential of deep brine resources is analyzed to provide a basis for resource development decisions.

[0182] The working principle of the above technical solution is to analyze the rationality of the polygonal divisions of each area based on the uncertainty assessment results. If there is significant uncertainty in the resource volume of certain areas, these areas need to be re-divided to more accurately assess the resource volume. The parameter value ranges are adjusted to ensure that the model parameters used more accurately reflect the actual situation. This includes adjusting the input parameters of the geological model. After the adjustment is completed, steps S1 to S5 are repeated to obtain updated results.

[0183] After obtaining the final deep brine reserve calculation results, a resource development potential analysis is conducted. Specifically, based on the calculation results, the deep brine's mineability is assessed, taking into account factors such as economic feasibility, technical feasibility, and environmental impact. Combining market demand with resource reserves, the deep brine development potential is analyzed to provide a scientific basis for resource development decisions. This process involves the construction of an economic model to assess the benefits and risks of different development options.

[0184] The beneficial effect of this technical solution is that it optimizes the reserve calculation process by adjusting regional divisions or parameter ranges based on uncertainty assessment results. This feedback mechanism ensures the robustness of the results and enables dynamic adjustments based on different assessment conditions, resulting in more accurate final reserve assessments and more reliable decision support for resource development.

[0185] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the invention.

Claims

1. A method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation, characterized in that: include: S1: Obtain the sulfate salt lake drilling location data and brine parameter data, construct a Thiessen polygon partitioning map, and generate several regional polygons; S2: Based on the regional polygons, determine the brine parameter value range and parameter frequency distribution of each region, and generate a brine parameter probability distribution model; S3: Based on the brine parameter probability distribution model, conduct global sensitivity analysis, determine key sensitive parameters, and generate sensitive parameter evaluation results; S4: Based on the evaluation results of sensitive parameters and regional polygons, the Monte Carlo simulation method is used to calculate the brine resource reserves in each region and generate the regional brine resource reserve distribution; S5: Based on the regional brine resource reserves distribution, the total brine resource reserves of the entire region are calculated and the deep brine reserves calculation results are generated.

2. The method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation according to claim 1, characterized in that: Step S1 includes: S11: Collect the location coordinate data of all drill holes in the study area; S12: Based on the drilling location coordinate data, construct the Thiessen polygon Voronoi diagram to generate the initial area polygon; S13: Optimize the initial region polygon to generate the final region polygon; Specifically, the weight coefficient of each borehole is calculated based on the brine parameter data at the borehole, which includes the area of ​​the ore layer, the thickness of the ore layer, the brine density, the porosity and the brine grade, to reflect the spatial heterogeneity; Generate a weighted Voronoi diagram based on the calculated weight coefficient to form a regional polygon; Apply the convex hull algorithm or the minimum spanning tree algorithm to adjust the boundaries of the polygons to ensure that the area of ​​each polygon is within the preset range and the shape is a regular polygon.

3. The method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation according to claim 2, characterized in that: Step S12 includes: S121: Calculate the Euclidean distance from each borehole to all other boreholes using the Euclidean distance formula; S122: Find the nearest neighbor borehole of each borehole based on Euclidean distance; S123: Connect adjacent drill holes and draw corresponding perpendicular bisectors; S124: The area enclosed by the perpendicular bisectors forms a Voronoi diagram area.

4. The method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation according to claim 1, characterized in that: Step S2 includes: S21: Obtain brine parameter data of wells drilled within each regional polygon, including ore layer area, ore layer thickness, brine density, porosity and brine grade; S22: Determine the value range of each parameter based on geological data and expert evaluation; S23: Based on historical data, experience from similar areas and geostatistical methods, determine the frequency distribution type of each parameter and generate a probability distribution model for brine parameters.

5. The method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation according to claim 1, characterized in that: The S3 steps include: S31: Based on the brine parameter probability distribution model, a Sobol sequence is generated to uniformly sample the parameter space. The Sobol sequence represents a low-discrepancy sequence used to uniformly cover the parameter space. S32: Calculate the variance contribution of each parameter to the reserve calculation result using the variance decomposition formula; S33: Calculate the Sobol sensitivity index of each parameter based on the variance contribution; Among them, the Sobol sensitivity index S i Defined as: Among them, Vi represents the contribution of the i-th parameter to the output variance, and V0 is the total variance of the model output; S34: Select parameters whose sensitivity index is higher than the preset index as key sensitive parameters and generate sensitive parameter evaluation results.

6. The method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation according to claim 1, characterized in that: The S4 step includes: S41: Randomly sample key sensitive parameters within the corresponding probability distribution range to generate multiple groups of parameter combinations; S42: Substitute each set of parameter combinations into the resource reserve calculation model to calculate the resource reserves of a single area; S43: Repeat steps S41 and S42 multiple times until the simulation results converge to the preset convergence conditions, and generate the regional brine resource reserves distribution.

7. The method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation according to claim 6, characterized in that: The resource reserve calculation model in step S42 is: P kL =S×M×D×K×C Among them, P kL It represents the available resources of liquid potash ore; S represents the distribution area of ​​the brine ore layer to be calculated; M represents the average effective thickness of the ore layer; D represents the average density of the brine ore layer; K represents the average porosity; C represents the average grade.

8. The method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation according to claim 6, characterized in that: The simulation results in step S43 converge to the preset convergence conditions including: Calculate whether the standard deviation of the iteration result is less than the preset threshold and tends to zero; Or calculate whether the relative difference between the result of the nth iteration and the average of the results of the previous n-1 iterations is less than a preset threshold; Or observe whether the width of the 95% confidence interval of the simulation results narrows to the preset range as the number of iterations increases; Or observe whether the change range of key sensitive parameters is less than the preset threshold as the number of iterations increases.

9. The method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation according to claim 1, characterized in that: Step S5 includes: S51: Summarize the resource reserves within all regional polygons to obtain the total resource volume of the study area; S52: performing statistical analysis on all calculated resource quantities to obtain a probability distribution of the resource quantities; S53: 95% confidence interval for the determined resource; S54: Calculate the standard deviation of the resource volume, evaluate the uncertainty of the resource volume, and generate the deep brine reserve calculation results.

10. The method for calculating deep brine reserves based on Thiessen polygons and Monte Carlo simulation according to claim 1, characterized in that: Also includes: According to the uncertainty assessment results, adjust the regional polygon division or parameter value range, and re-execute steps S1 to S5 until the uncertainty assessment results meet the preset conditions; Based on the final calculation results of deep brine reserves, the development potential of deep brine resources is analyzed to provide a basis for resource development decisions.

Images