A method for evaluating the quality of coal-based rare metal minerals

Abstract

This invention relates to the field of resource valuation technology, specifically a method for evaluating the quality of coal-bearing rare metal minerals. It constructs multi-dimensional attribute data through spectral scanning of core samples, introduces sequence position weights to correct polymetallic concentrations, and combines grade interval determination and conflict identification mechanisms to determine the quality grade of coal-bearing rare metal minerals. This invention focuses on the stratigraphic continuity and multi-element synergistic characteristics of coal-bearing rare metal mineral quality formation. It acquires polymetallic information from core samples using spectral methods and constructs continuous attribute expressions based on the collected sequences. During the evaluation process, stratigraphic position weights are introduced to reflect the influence of different depths on the enrichment degree of rare metals, shifting the quality determination from discrete numerical values ​​to a comprehensive expression with stratigraphic constraints. In grade classification, interval matching and distribution ratio analysis are introduced to identify differences in multi-element orientations, and the degree of boundary deviation is used to screen for more representative quality grades, improving the consistency and reliability of coal-bearing rare metal mineral quality evaluation.

Description

Technical Field

[0001] This invention relates to the field of resource valuation technology, and in particular to a method for evaluating the quality of coal-based rare metal minerals. Background Technology

[0002] The field of resource valuation technology involves a set of technologies for systematically analyzing and evaluating the quantity, scale, quality characteristics, occurrence conditions, development and utilization potential, and economic attributes of various natural and mineral resources. This field typically includes resource geological surveys, sampling and testing, index system construction, quality grading standards formulation, resource grade and composition characteristic analysis, statistical calculation and comprehensive judgment methods, etc. By objectively describing the physical and chemical properties of the resources themselves and their distribution characteristics, a comparable and quantifiable evaluation basis is established, thereby providing technical support for resource management planning and related research.

[0003] Among them, the coal-series rare metal mineral quality evaluation method refers to the technical method for judging the quality of rare, dispersed, and rare earth metal minerals associated with coal-series strata. The technical issues it addresses are to clarify the content level, occurrence form, and quality grade of coal-series rare metals. Traditional methods usually take coal seam and surrounding rock samples as research objects, obtain elemental content data of rare, dispersed, and rare earth metals through field sampling and laboratory testing, classify the samples by combining coal and rock type mineral composition and stratigraphic characteristics, and judge and describe the quality of coal-series rare metal minerals according to pre-set evaluation indicators and grading thresholds.

[0004] Current quality assessment of coal-bearing rare earth metals relies on discrete sampling and experimental testing results. Sample acquisition is constrained by borehole spacing and stratigraphic segmentation, making it difficult to fully reflect the continuous characteristics of rare earth metal variations along depth in coal-bearing strata. The evaluation process directly classifies samples by comparing the content of a single element with a fixed grade threshold, failing to reflect the differences in the contribution of different stratigraphic layers to the formation of mineral quality. When multiple rare and dispersed rare earth metals point to different grade levels, there is a lack of effective coordination mechanisms, leading to grade fluctuations in samples near grade boundaries, which affects the stability and repeatability of quality assessment of coal-bearing rare earth metals. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides a method for evaluating the quality of coal-based rare metal minerals, comprising the following steps:

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for evaluating the quality of coal-based rare metal minerals, comprising the following steps:

[0007] S1: Use a spectrometer to scan coal-bearing core samples, extract sample spectral intensity data and sample sequence index, convert sample spectral intensity data into metal element concentration, and construct a multidimensional attribute dataset of the samples;

[0008] S2: Based on the sample multidimensional attribute dataset, calculate the position weight coefficient by statistically analyzing the proportion of the sample sequence index in the total number of samples obtained, correct the metal element concentration, and generate a weighted concentration feature value;

[0009] S3: Obtain the critical threshold for quality grading that distinguishes the quality grade of minerals, analyze the critical threshold for quality grading to construct a sequence of quality judgment numerical intervals, match the weighted concentration feature value to the sequence of quality judgment numerical intervals, and determine the single-parameter quality grade identifier.

[0010] S4: Calculate the percentage of the quality level distribution of the single parameter quality level identifier relative to the total number of parameters, compare the percentage of the quality level distribution for each sample, and mark the evaluation conflict status label;

[0011] S5: Obtain the sample multidimensional attribute dataset of the sample labeled as the evaluation conflict state label, determine the candidate judgment quality level, and generate the mineral quality assessment result based on the candidate judgment quality level.

[0012] As a further aspect of the present invention, the sample multidimensional attribute dataset includes metal element concentration and sample sequence index. The weighted concentration feature value is specifically a numerical result after correcting the metal element concentration based on the position weight coefficient. The single-parameter quality grade identifier specifically refers to the interval category to which a single metal element belongs in the quality judgment interval sequence. The evaluation conflict status label is specifically a marker generated when the distribution proportion of quality grades does not meet the consistency judgment threshold, indicating the divergence of multiple parameters. The mineral quality assessment result is specifically the minimum deviation quality grade with the smallest deviation degree selected based on the boundary deviation difference.

[0013] As a further aspect of the present invention, step S1 specifically comprises:

[0014] S101: Using a spectral analysis instrument, perform point-by-point scanning on coal-bearing core samples arranged in order of stratum depth, detect and extract the original spectral signals that characterize the internal components of the coal-bearing core samples, perform noise reduction and characteristic peak extraction processing on the original spectral signals, and obtain sample spectral intensity data.

[0015] S102: Read the physical number of the recorded sample collection sequence, parse the physical number according to the arrangement order in the geological profile, convert the physical number into a value that represents the relative position of the sample in the overall borehole sequence, and establish a sample sequence index.

[0016] S103: Based on the preset numerical mapping relationship between spectral intensity and material content, the sample spectral intensity data is converted into a value representing the content of each metal element inside the sample, generating the metal element concentration, and the metal element concentration is associated and integrated with the sample sequence index to construct a multidimensional attribute dataset of the sample.

[0017] As a further aspect of the present invention, step S2 specifically comprises:

[0018] S201: Obtain the statistical number of all samples in the sample set to be evaluated, extract the sample sequence index from the sample multidimensional attribute dataset, and calculate the layer depth ratio based on the sample sequence index and the statistical number of all samples.

[0019] S202: Based on the depth ratio of the stratigraphic layers, retrieve the corresponding rules for the differences in the contribution of various stratigraphic layers to the ore deposit quality, match and calculate the location weight coefficient representing the importance of each depth layer;

[0020] S203: Based on the location weight coefficient, the metal element concentration in the sample multidimensional attribute dataset is numerically corrected, and the influence reflecting the importance of the sample's stratum location is superimposed on the original metal element concentration to generate a weighted concentration feature value.

[0021] As a further aspect of the present invention, step S3 specifically comprises:

[0022] S301: Obtain the critical threshold for quality grading that distinguishes the quality grade of each mineral, analyze the numerical breakpoints in the critical threshold for quality grading, and construct a sequence of numerical intervals for quality determination.

[0023] S302: Match the weighted concentration feature values ​​to the sequence of quality judgment numerical intervals, determine the quality judgment numerical interval in which each weighted concentration feature value falls, and obtain the relationship parsing result;

[0024] S303: Based on the relation parsing results, determine the quality level corresponding to each metal element in the sample multidimensional attribute dataset, and map the quality level to an identifier to generate a single-parameter quality level identifier.

[0025] As a further aspect of the present invention, step S4 specifically comprises:

[0026] S401: Collect all the single-parameter quality level identifiers under the same sample multidimensional attribute dataset, count the number of times each identifier appears, summarize the frequency of each quality level, and generate quality level frequency statistics.

[0027] S402: Obtain the total number of all metal element concentrations in the multidimensional attribute dataset of the sample to determine the total number of parameters, and calculate the distribution ratio of the specified quality level based on the quality level frequency statistics and the total number of parameters;

[0028] S403: Compare the proportion of the quality grade distribution with a preset consistency judgment threshold, filter samples whose quality grade distribution proportion is lower than the consistency judgment threshold, mark the target samples, and generate evaluation conflict status labels.

[0029] As a further aspect of the present invention, step S5 specifically comprises:

[0030] S501: Identify the sample multidimensional attribute dataset that is labeled as the evaluation conflict state label, call the weighted concentration feature value of the associated sample multidimensional attribute dataset, retrieve the grade boundary point on the numerical axis adjacent to the weighted concentration feature value from the quality grading critical threshold, and determine the candidate judgment quality level.

[0031] S502: Calculate the absolute difference between the weighted concentration feature value and the reference critical boundary corresponding to the candidate quality level, quantify the degree of deviation of the parameter from the grading critical point, and generate the boundary deviation difference;

[0032] S503: Sort the boundary deviation differences in ascending order, filter the candidate judgment quality level corresponding to the minimum boundary deviation, define the target candidate judgment quality level as the minimum deviation quality level, and generate the mineral quality assessment result.

[0033] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0034] In this invention, focusing on the stratigraphic continuity and multi-element synergistic characteristics of coal-bearing rare metal mineral formation, multimetallic information from core samples is obtained through spectral methods and combined with the acquisition sequence to construct a continuous attribute expression. During the evaluation process, stratigraphic position weights are introduced to reflect the influence of different depths on the enrichment degree of rare metals, transforming quality judgment from discrete numerical values ​​to a comprehensive expression with stratigraphic constraints. When classifying grades, interval matching and distribution ratio analysis are introduced to identify differences in multi-element orientations, and more representative quality grades are selected by the degree of boundary deviation, thereby improving the consistency and reliability of coal-bearing rare metal mineral quality evaluation. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a schematic diagram of the steps of the present invention;

[0037] Figure 2 This is a detailed schematic diagram of S1 of the present invention;

[0038] Figure 3 This is a detailed schematic diagram of S2 of the present invention;

[0039] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0040] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0041] Figure 6 This is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0042] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0043] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0044] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0045] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0046] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0047] Please see Figure 1 This invention provides a method for evaluating the quality of coal-based rare metal minerals, comprising the following steps:

[0048] S1: Use a spectrometer to scan coal-bearing core samples, extract sample spectral intensity data and sample sequence index, convert sample spectral intensity data into metal element concentration, and construct a multidimensional attribute dataset of the samples;

[0049] S2: Based on the multidimensional attribute dataset of the samples, calculate the position weight coefficient by statistically analyzing the proportion of the sample sequence index in the total number of samples obtained, correct the metal element concentration, and generate a weighted concentration feature value.

[0050] S3: Obtain the critical threshold for quality grading that distinguishes the quality grade of minerals, analyze the critical threshold for quality grading to construct a sequence of quality judgment numerical intervals, match the weighted concentration feature value to the sequence of quality judgment numerical intervals, and determine the single-parameter quality grade identifier.

[0051] S4: Statistically analyze the distribution of quality grade labels for individual parameters out of the total number of parameters, compare the distribution of quality grade for each sample, and mark the evaluation conflict status label.

[0052] S5: Obtain the multidimensional attribute dataset of samples labeled as evaluation conflict state labels, determine the candidate judgment quality level, and generate mineral quality assessment results based on the candidate judgment quality level.

[0053] The sample multidimensional attribute dataset includes metal element concentration, sample sequence index, weighted concentration feature value (specifically, the numerical result after correcting the metal element concentration based on the position weight coefficient), single-parameter quality grade identifier (specifically, the interval category to which a single metal element belongs in the quality judgment interval sequence), evaluation conflict status label (specifically, the marker indicating the divergence of multiple parameters when the distribution proportion of quality grades does not meet the consistency judgment threshold), and mineral quality assessment result (specifically, the minimum deviation quality grade with the smallest deviation degree selected based on the boundary deviation difference).

[0054] Please see Figure 2 Step S1 is as follows:

[0055] S101: Using a spectral analysis instrument, the coal-bearing core samples arranged in order of strata depth are scanned point by point to detect and extract the original spectral signals that characterize the internal components of the coal-bearing core samples. The original spectral signals are then denoised and feature peaks are extracted to obtain the spectral intensity data of the samples.

[0056] Using a near-infrared spectrometer, 300 coal-bearing core samples from the same borehole, arranged from top to bottom according to stratum depth, were scanned point-by-point. During the scanning process, the spectrometer probe moved along the axis of each core sample in 1-cm steps, emitting a near-infrared beam with a wavelength range of 900 nm to 2500 nm at each scanning point and receiving the spectral signal reflected from the sample surface. This signal is the raw spectral signal characterizing the internal composition of the coal-bearing core sample. For each acquired raw spectral signal, a wavelet transform denoising program was first initiated. Specifically, the Symlet-8 wavelet basis was selected, and the decomposition level was set to 5 levels. The signal was decomposed and reconstructed. The coefficients of the high-frequency coefficients in each decomposed level that were less than the preset noise threshold were set to zero. Then, the signal was reconstructed to obtain the denoised spectral signal. The noise threshold was set based on the analysis of 100 sets of unloaded signals collected by the instrument on a standard white board. The standard deviation of these signal energies was calculated, and 1.5 times the standard deviation was used as the noise threshold. For example, if the standard deviation was 0.002, the noise threshold was 0.003. Next, characteristic peak extraction processing was performed on the denoised spectral signal. Specifically, the absorption valleys in the spectral curve were identified using a second-derivative peak-finding algorithm. The wavelength positions corresponding to the absorption valleys were taken as the positions of the characteristic peaks, and the reciprocal of the spectral reflectance at those positions was recorded as the spectral intensity. For example, characteristic absorption peaks related to specific rare earth elements (such as scandium Sc and gallium Ga) were detected near wavelengths of 1450 nm, 1780 nm, and 2210 nm, and their spectral intensity values ​​were recorded as 0.85 and 0.72, respectively. Finally, a set of sample spectral intensity data consisting of multiple characteristic peak spectral intensities was generated for each sample.

[0057] S102: Read the physical number of the recorded sample collection sequence, parse the physical number according to the arrangement order in the geological profile, convert the physical number into a value that represents the relative position of the sample in the overall borehole sequence, and establish a sample sequence index.

[0058] The physical numbers assigned and recorded at the sample collection site for each coal-bearing core sample are read, such as the labels "ZK01-001" to "ZK01-300" affixed to the sample cassette. These physical numbers are then parsed based on the arrangement information in the geological profile. The parsing process involves identifying the borehole identifier "ZK01" and the sample serial numbers "001" to "300" in the numbers, confirming that all samples originate from the same borehole sequence, and that the collection order is consistent with the increasing direction of the serial numbers; that is, the smaller the serial number, the shallower its burial depth in the formation. The numerical serial numbers in the physical numbers are directly converted into numerical values ​​representing the relative position of the sample within the overall borehole sequence. This conversion quantifies the non-numerical physical numbers into sequence indices that can be used for subsequent calculations. For example, the sample with physical number "ZK01-001" has a sequence number of 1, indicating that it is at the very top of the entire borehole sample sequence; the sample with physical number "ZK01-150" has a sequence number of 150, indicating that it is in the middle of the sequence; and the sample with physical number "ZK01-300" has a sequence number of 300, indicating that it is at the very end of the sequence. Through this conversion, a sample sequence index is established for each sample, continuously increasing from 1 to 300.

[0059] S103: Based on the preset numerical mapping relationship between spectral intensity and material content, the sample spectral intensity data is converted into a value that characterizes the content of each metal element in the sample, generating the metal element concentration, and the metal element concentration is associated with and integrated with the sample sequence index to construct a multidimensional attribute dataset of the sample.

[0060] The numerical mapping relationship was established based on the spectral detection of a large number of standard coal-bearing rock samples with known compositional contents. The correlation between the spectral intensity and the concentration of rare metal elements (in grams per ton) obtained from chemical analysis was statistically analyzed, and a conversion model was established through multiple linear regression analysis. For example, for the rare metal element gallium (Ga), the pre-defined mapping relationship determined a linear conversion coefficient between the characteristic peak spectral intensity at 1780 nm and the gallium content: gallium concentration = spectral intensity × 80 + 5. If a sample has a sample sequence index of 150 and its spectral intensity at 1780 nm is 0.72, then its corresponding gallium concentration is calculated as 0.72 × 80 + 5 = 62.6 g / ton. Similarly, the characteristic peak intensities related to other rare metal elements such as scandium (Sc) in the sample's spectral data were also calculated using their respective conversion coefficients to obtain the corresponding metal element concentrations. The calculated concentrations of all metal elements (e.g., gallium concentration 62.6 g / t, scandium concentration 35.5 g / t) are associated with the sample sequence index (150) of the sample to construct a data structure containing sample location information and multiple metal element concentration information. Finally, a multidimensional attribute dataset of samples is constructed for all 300 samples.

[0061] Please see Figure 3 Step S2 is as follows:

[0062] S201: Obtain the statistical count of all samples in the sample set to be evaluated, extract the sample sequence index from the multidimensional attribute dataset of the samples, and calculate the layer depth ratio based on the sample sequence index and the statistical count of all samples.

[0063] By counting the total number of records in the sample multidimensional attribute dataset, the total number of samples was determined to be 300. Subsequently, the sample sequence index for each sample was extracted from the dataset. Based on the extracted sample sequence index and the total number of samples (300), the stratigraphic depth percentage for each sample was calculated. The calculation process is as follows: divide the sample sequence index value of a single sample by the total number of samples. For example, for a sample with a sample sequence index of 1, its stratigraphic depth percentage is 1 ÷ 300 = 0.0033; for a sample with a sample sequence index of 150, its stratigraphic depth percentage is 150 ÷ ​​300 = 0.5; and for a sample with a sample sequence index of 300, its stratigraphic depth percentage is 300 ÷ 300 = 1.0. This stratigraphic depth percentage is a dimensionless value between 0 and 1, which normalizes the discrete sample number and represents the sample's relative position from top to bottom in the stratigraphic profile, where 0 represents the topmost point and 1 represents the bottommost point.

[0064] S202: Based on the depth ratio of the stratigraphic layers, retrieve the corresponding rules for the differences in the contribution of various stratigraphic layers to the ore deposit quality, match and calculate the location weight coefficient representing the importance of each depth layer;

[0065] Based on the calculated stratigraphic depth percentage of each sample, a pre-defined rule is used to determine the differences in the contribution of various stratigraphic layers to the quality of the deposit. This rule is determined based on the analysis of historical exploration data and geological metallogenic theories, reflecting the differences in mineralization enrichment at different depths within a specific coal-bearing rare metal deposit type. The rule divides the entire stratigraphic depth (represented by a stratigraphic depth percentage of 0.0 to 1.0) into three intervals, assigning a fixed positional weight coefficient to each interval. The interval division and coefficient setting are based on the following: statistical analysis of geological data from over 50 similar proven deposits revealed that the vertical distribution of their main ore bodies is concentrated primarily in the middle stratigraphic layers. Specifically, the region with a stratigraphic depth ratio between 0.4 and 0.7 is defined as the core metallogenic belt, whose samples contribute the most to the overall deposit quality; the regions with a stratigraphic depth ratio between 0.2 and 0.4 and between 0.7 and 0.9 are defined as secondary metallogenic belts, which contribute the least; and the regions with a stratigraphic depth ratio between 0.0 and 0.2 and between 0.9 and 1.0 are defined as the marginal belt, which contribute the least. Based on this, the corresponding rules are set as follows: when the percentage of layer depth falls within the range of [0.4, 0.7], the location weight coefficient is 1.2; when the percentage of layer depth falls within the range of [0.2, 0.4) or (0.7, 0.9], the location weight coefficient is 1.0; and when the percentage of layer depth falls within the range of [0.0, 0.2) or (0.9, 1.0], the location weight coefficient is 0.8. For example, for a sample with a layer depth percentage of 0.5, which falls within the range of [0.4, 0.7], the matched location weight coefficient is 1.2. For a sample with a layer depth percentage of 0.15, which falls within the range of [0.0, 0.2), the matched location weight coefficient is 0.8.

[0066] S203: Numerical correction of metal element concentration in sample multidimensional attribute dataset based on location weight coefficient, superimposing the influence reflecting the importance of the sample's stratum location onto the original metal element concentration to generate weighted concentration feature value;

[0067] The process of numerically correcting the metal element concentration in the multidimensional attribute dataset of the sample based on the location weighting coefficient is as follows:

[0068] Obtain the position weight coefficients corresponding to the specified sample and the concentration of each metal element recorded in the sample's multidimensional attribute dataset;

[0069] The position weighting coefficient is multiplied by the concentration of each metal element to obtain a set of corrected concentration values.

[0070] First, obtain the position weighting coefficient corresponding to the specified sample and the concentration of each metal element recorded in the sample's multidimensional attribute dataset. For example, for the sample with sequence index 150, its layer depth ratio is 0.5, the calculated position weighting coefficient is 1.2, and the metal element concentrations recorded in its multidimensional attribute dataset are: gallium concentration 62.6 g / t and scandium concentration 35.5 g / t. Next, multiply the position weighting coefficient 1.2 with each metal element concentration recorded for the sample to obtain a set of corrected concentration values. Specifically, the corrected gallium concentration = 62.6 × 1.2 = 75.12 g / t; the corrected scandium concentration = 35.5 × 1.2 = 42.6 g / t. This set of corrected concentration values ​​is the weighted concentration feature value of the sample, and this set of weighted concentration feature values ​​is updated in the sample's attribute data.

[0071] Please see Figure 4 Step S3 is as follows:

[0072] S301: Obtain the critical threshold for quality grading that distinguishes the quality grade of each mineral, analyze the numerical breakpoints in the critical threshold for quality grading, and construct a sequence of numerical intervals for quality determination.

[0073] The process of analyzing the numerical breakpoints in the critical threshold of quality grading is as follows:

[0074] Obtain the critical threshold for quality grading, including critical values ​​corresponding to multiple quality grades;

[0075] All critical values ​​in the quality grading threshold are sorted in ascending order to form an ordered numerical sequence.

[0076] Each critical value in the ordered numerical sequence is defined as a numerical breakpoint, and adjacent quality levels are divided based on the numerical breakpoints.

[0077] Select two adjacent numerical breakpoints in the ordered numerical sequence in sequence, and define the numerical range between the two adjacent numerical breakpoints as an independent interval, with one numerical breakpoint as the starting boundary of the interval and the other numerical breakpoint as the ending boundary of the interval.

[0078] A set of pre-defined critical thresholds for quality grading of coal-based gallium resources was obtained. These thresholds were formulated based on the national standard "Geological Exploration Specifications for Gallium, Germanium, Lithium, Niobium, Tantalum, Zirconium, Hafnium, and Rare Earth Minerals in Coal" (DZ / T0214-2020) regarding boundary grades and industrial grades, combined with the specific economic and technical conditions of the mining area. These thresholds include critical values ​​corresponding to three quality grades: "industrial grade," "boundary grade," and "below grade." Specifically, the critical value for gallium content in industrial grade is 60 g / t, and the critical value for gallium content in boundary grade is 40 g / t. The numerical breakpoints in this quality grading critical threshold were analyzed. First, these two critical values ​​were sorted in ascending order to form an ordered numerical sequence: 40, 60. Each critical value 40 and 60 in the ordered numerical sequence was defined as a numerical breakpoint, and adjacent quality grades were divided based on these breakpoints. Subsequently, two adjacent numerical breakpoints or a single breakpoint and an infinite boundary were selected sequentially in the ordered numerical sequence, and the numerical range between them was defined as an independent interval. Specifically, the range of values ​​greater than or equal to 60 is defined as the "industrial grade" interval; the range of values ​​between 40 (inclusive) and 60 (exclusive) is defined as the "borderline grade" interval; and the range of values ​​less than 40 is defined as the "below grade" interval. Thus, a sequence of quality judgment numerical intervals is constructed, consisting of three non-overlapping numerical intervals.

[0079] S302: Match the weighted concentration feature values ​​to the sequence of quality judgment numerical intervals, determine the quality judgment numerical interval in which each weighted concentration feature value falls, and obtain the relationship parsing result;

[0080] Taking sample sequence index 150 as an example, its weighted concentration characteristic value for gallium is 75.12 g / t. Comparing the value 75.12 with the quality assessment value interval sequence: 75.12 is greater than or equal to 60, therefore it falls into the "industrial grade" interval. Through this matching operation, it is determined that the weighted concentration characteristic value of gallium in this sample falls into the quality assessment value interval corresponding to "industrial grade". The same matching operation for the weighted concentration characteristic values ​​of other metal elements (such as scandium) in this sample is also performed for their respective quality grading standards. After completing the matching of the weighted concentration characteristic values ​​of all samples and all metal elements, the relationship resolution result is obtained, which records the quality assessment value interval to which each weighted concentration characteristic value belongs.

[0081] S303: Based on the relation parsing results, determine the quality level corresponding to each metal element in the sample multidimensional attribute dataset, and map the quality level to an identifier to generate a single-parameter quality level identifier;

[0082] For sample with sequence index 150, the weighted concentration characteristic value of gallium falls within the "industrial grade" range, thus determining the gallium quality grade of this sample as "industrial grade". Next, the determined quality grade is mapped to a preset identifier. The identifier setting rule is: industrial grade is mapped to "H", borderline grade to "M", and below grade to "L". Therefore, the gallium quality grade "industrial grade" of this sample is mapped to the identifier "H". The same grade determination and identifier mapping operation is performed on other metal elements (such as scandium) in the multidimensional attribute dataset of this sample. For example, if its weighted concentration of 42.6 g / t falls within the "borderline grade" range, its identifier is "M". Finally, an independent single-parameter quality grade identifier, such as "H" or "M", is generated for each metal element.

[0083] Please see Figure 5 Step S4 is as follows:

[0084] S401: Collect all single-parameter quality grade identifiers under the same sample multidimensional attribute dataset, count the number of times each identifier appears, summarize the frequency of each quality grade, and generate quality grade frequency statistics.

[0085] This dataset compiles all single-parameter quality grade identifiers from a multidimensional attribute dataset of the same sample (e.g., sample with a sequence index of 150). Assume the sample was analyzed for 10 rare metal elements, resulting in 10 single-parameter quality grade identifiers: H, M, M, H, M, L, M, L, M, H. The frequency of each identifier is counted: "H" appears 3 times, "M" appears 5 times, and "L" appears 2 times. These counts are then summed to generate the quality grade frequency statistics for the sample, resulting in a data structure recording the frequency of each quality grade: {H: 3, M: 5, L: 2}.

[0086] S402: Obtain the total number of all metal element concentrations in the multidimensional attribute dataset of the sample to determine the total number of parameters, and calculate the distribution ratio of the specified quality level based on the quality level frequency statistics and the total number of parameters;

[0087] Obtain the total number of metal element concentrations in the sample's multidimensional attribute dataset. By querying the dataset's structure or metadata, determine the total number of parameters to be 10. Based on the generated quality grade frequency statistics and the total number of parameters (10), calculate the distribution percentage of a specified quality grade. The calculation method is: divide the frequency statistics of a quality grade by the total number of parameters. For example, for a sample with a sample sequence index of 150, the distribution percentage of the "H" grade is 3 ÷ 10 = 0.3; the distribution percentage of the "M" grade is 5 ÷ 10 = 0.5; and the distribution percentage of the "L" grade is 2 ÷ 10 = 0.2.

[0088] S403: Compare the proportion of quality grade distribution with the preset consistency judgment threshold, screen samples whose quality grade distribution proportion is lower than the consistency judgment threshold, mark the target samples, and generate evaluation conflict status labels.

[0089] The calculated distribution percentage of each quality level is compared with a preset consistency threshold. This consistency threshold is determined based on statistical analysis of the grade distribution characteristics of "undisputed" samples in historical evaluation data. The experiment proceeds as follows: 1000 samples with clear final evaluation results are selected, and the distribution percentage of each quality level within each sample is calculated. It is found that in 95% of the samples, the proportion of the dominant quality level exceeds a certain value. The 5th percentile of the dominant grade proportion of these 1000 samples is taken as the threshold. If this value is calculated to be 0.68, the consistency threshold is set to 0.7. During the comparison, samples whose quality level distribution percentage is lower than the consistency threshold of 0.7 are selected. For the sample with sequence index 150, its highest quality level distribution percentage is 0.5 for the "M" grade, which is lower than 0.7. Therefore, this sample is selected. A "conflict" status label is added to all selected samples, generating a conflict status label.

[0090] Please see Figure 6 Step S5 is as follows:

[0091] S501: Identify the multidimensional attribute dataset of samples that are labeled as evaluation conflict status, call the weighted concentration feature value of the associated multidimensional attribute dataset of samples, retrieve the grade boundary point on the numerical axis adjacent to the weighted concentration feature value from the quality grading critical threshold, and determine the candidate judgment quality level.

[0092] The process of determining the candidate quality level is as follows:

[0093] The weighted concentration feature value is called, and the two grade boundary points adjacent to the weighted concentration feature value on the numerical axis are retrieved from the quality grading critical threshold. One grade boundary point is defined as the upper critical point, and the other grade boundary point is defined as the lower critical point.

[0094] Extract the attribution levels defined by the upper and lower critical points respectively, and set these two attribution levels together as the candidate quality levels for judgment;

[0095] Identify the multidimensional attribute dataset of samples labeled with the "evaluation conflict" status tag, such as sample sequence index 150. Retrieve the weighted concentration feature value associated with this sample's multidimensional attribute dataset that triggered the conflict. Assume the conflict is primarily caused by gallium, with a weighted concentration feature value of 58.5 g / t (a new conflict example is used here to better illustrate the logic). From the defined quality grading critical thresholds (numerical breakpoints 40 and 60), retrieve the two grade boundary points adjacent to 58.5 on the numerical axis: 40 and 60. Define the grade boundary point 60, with the larger value, as the upper critical point, and the grade boundary point 40, with the smaller value, as the lower critical point. Next, extract the classification grades defined by the upper critical point 60 and the lower critical point 40, respectively. The upper critical point 60 is the starting boundary for "industrial grade," defining the classification grade as "industrial grade" (H); the lower critical point 40 is the starting boundary for "border grade," defining the classification grade as "border grade" (M). The two classification levels, "industrial grade" and "borderline grade", are jointly set as the candidate quality levels for this parameter.

[0096] S502: Calculate the absolute difference between the weighted concentration characteristic value and the reference critical boundary corresponding to the candidate quality level, quantify the degree of deviation of the parameter from the grading critical point, and generate the boundary deviation difference;

[0097] The absolute difference between the weighted concentration characteristic value of the conflict parameter (58.5 g / t) and the reference critical boundary corresponding to each candidate quality grade is calculated. For the candidate grade "Industrial Grade", the reference critical boundary is the upper critical point 60, and the calculated boundary deviation is |58.5 - 60| = 1.5. For the candidate grade "Boundary Grade", the reference critical boundary is the lower critical point 40, and the calculated boundary deviation is |58.5 - 40| = 18.5. Through this calculation, the degree of deviation of the parameter from the grading critical point is quantified, generating a set of boundary deviations: {Industrial Grade: 1.5, Boundary Grade: 18.5}.

[0098] S503: Sort the boundary deviations in ascending order, filter the candidate judgment quality levels corresponding to the minimum boundary deviation, define the target candidate judgment quality level as the minimum deviation quality level, and generate the mineral quality assessment results;

[0099] The generated boundary deviations {1.5, 18.5} are sorted in ascending order to obtain the sorted sequence: 1.5, 18.5. The candidate quality grade corresponding to the minimum boundary deviation of 1.5 is selected, i.e., "industrial grade". This target candidate quality grade "industrial grade" is defined as the final evaluation result for this conflict parameter (gallium element), and this result is updated in the sample dataset as the final mineral quality evaluation result for this parameter.

[0100] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for evaluating the quality of coal-bearing rare metal minerals, characterized in that, Includes the following steps: S1: Use a spectrometer to scan coal-bearing core samples, extract sample spectral intensity data and sample sequence index, convert sample spectral intensity data into metal element concentration, and construct a multidimensional attribute dataset of the samples; S2: Based on the sample multidimensional attribute dataset, calculate the position weight coefficient by statistically analyzing the proportion of the sample sequence index in the total number of samples obtained, correct the metal element concentration, and generate a weighted concentration feature value; S3: Obtain the critical threshold for quality grading that distinguishes the quality grade of minerals, analyze the critical threshold for quality grading to construct a sequence of quality judgment numerical intervals, match the weighted concentration feature value to the sequence of quality judgment numerical intervals, and determine the single-parameter quality grade identifier. S4: Calculate the percentage of the quality level distribution of the single parameter quality level identifier relative to the total number of parameters, compare the percentage of the quality level distribution for each sample, and mark the evaluation conflict status label; S5: Obtain the multidimensional attribute dataset of the sample labeled as the evaluation conflict state label sample, determine the candidate judgment quality level, and generate the mineral quality assessment result based on the candidate judgment quality level. Step S2 is as follows: S201: Obtain the statistical number of all samples in the sample set to be evaluated, extract the sample sequence index from the sample multidimensional attribute dataset, and calculate the layer depth ratio based on the sample sequence index and the statistical number of all samples. S202: Based on the depth ratio of the stratigraphic layers, retrieve the corresponding rules for the differences in the contribution of various stratigraphic layers to the ore deposit quality, match and calculate the location weight coefficient representing the importance of each depth layer; S203: Based on the location weighting coefficient, the metal element concentration in the sample multidimensional attribute dataset is numerically corrected, and the influence reflecting the importance of the sample's stratum location is superimposed on the original metal element concentration to generate a weighted concentration feature value. Step S4 is as follows: S401: Collect all the single-parameter quality level identifiers under the same sample multidimensional attribute dataset, count the number of times each identifier appears, summarize the frequency of each quality level, and generate quality level frequency statistics. S402: Obtain the total number of types of metal element concentrations in the multidimensional attribute dataset of the sample to determine the total number of parameters, and calculate the distribution ratio of the specified quality level based on the quality level frequency statistics and the total number of parameters; S403: Compare the proportion of the quality grade distribution with a preset consistency judgment threshold, filter samples whose quality grade distribution proportion is lower than the consistency judgment threshold, mark the target samples, and generate evaluation conflict status labels.

2. The method for evaluating the quality of coal-bearing rare metal minerals according to claim 1, characterized in that, The sample multidimensional attribute dataset includes metal element concentration and sample sequence index. The weighted concentration feature value is specifically a numerical result after correcting the metal element concentration based on the position weight coefficient. The single-parameter quality grade identifier specifically refers to the interval category to which a single metal element belongs in the quality judgment interval sequence. The evaluation conflict status label is specifically a marker generated when the distribution proportion of quality grades does not meet the consistency judgment threshold, indicating that multiple parameters point to divergence. The mineral quality assessment result is specifically the minimum deviation quality grade with the smallest deviation degree selected based on the boundary deviation difference.

3. The method for evaluating the quality of coal-bearing rare metal minerals according to claim 1, characterized in that, Step S1 is as follows: S101: Using a spectral analysis instrument, perform point-by-point scanning on coal-bearing core samples arranged in order of stratum depth, detect and extract the original spectral signals that characterize the internal components of the coal-bearing core samples, perform noise reduction and characteristic peak extraction processing on the original spectral signals, and obtain sample spectral intensity data. S102: Read the physical number of the recorded sample collection sequence, parse the physical number according to the arrangement order in the geological profile, convert the physical number into a value that represents the relative position of the sample in the overall borehole sequence, and establish a sample sequence index. S103: Based on the preset numerical mapping relationship between spectral intensity and material content, the sample spectral intensity data is converted into a value representing the content of each metal element inside the sample, generating the metal element concentration, and the metal element concentration is associated and integrated with the sample sequence index to construct a multidimensional attribute dataset of the sample.

4. The method for evaluating the quality of coal-bearing rare metal minerals according to claim 1, characterized in that, The process of numerically correcting the metal element concentration in the multidimensional attribute dataset of the sample based on the location weighting coefficient is as follows: Obtain the position weight coefficients corresponding to the specified sample and the concentration of each metal element recorded in the sample's multidimensional attribute dataset; The position weighting coefficient is multiplied by the concentration of each metal element to obtain a set of corrected concentration values.

5. The method for evaluating the quality of coal-bearing rare metal minerals according to claim 1, characterized in that, Step S3 is as follows: S301: Obtain the critical threshold for quality grading that distinguishes the quality grade of each mineral, analyze the numerical breakpoints in the critical threshold for quality grading, and construct a sequence of numerical intervals for quality determination. S302: Match the weighted concentration feature values ​​to the sequence of quality judgment numerical intervals, determine the quality judgment numerical interval in which each weighted concentration feature value falls, and obtain the relationship parsing result; S303: Based on the relation parsing results, determine the quality level corresponding to each metal element in the sample multidimensional attribute dataset, and map the quality level to an identifier to generate a single-parameter quality level identifier.

6. The method for evaluating the quality of coal-bearing rare metal minerals according to claim 5, characterized in that, The process of analyzing the numerical breakpoints in the critical threshold of quality grading is as follows: Obtain the critical threshold for quality grading, including critical values ​​corresponding to multiple quality grades; All critical values ​​in the quality grading threshold are sorted in ascending order to form an ordered numerical sequence. Each critical value in the ordered numerical sequence is defined as a numerical breakpoint, and adjacent quality levels are divided based on the numerical breakpoints. Select two adjacent numerical breakpoints in the ordered numerical sequence in sequence, and define the numerical range between the two adjacent numerical breakpoints as an independent interval, with one numerical breakpoint as the starting boundary of the interval and the other numerical breakpoint as the ending boundary of the interval.

7. The method for evaluating the quality of coal-bearing rare metal minerals according to claim 1, characterized in that, Step S5 is as follows: S501: Identify the sample multidimensional attribute dataset that is labeled as the evaluation conflict state label, call the weighted concentration feature value of the associated sample multidimensional attribute dataset, retrieve the grade boundary point on the numerical axis adjacent to the weighted concentration feature value from the quality grading critical threshold, and determine the candidate judgment quality level. S502: Calculate the absolute difference between the weighted concentration feature value and the reference critical boundary corresponding to the candidate quality level, quantify the degree of deviation of the parameter from the grading critical point, and generate the boundary deviation difference; S503: Sort the boundary deviation differences in ascending order, filter the candidate judgment quality level corresponding to the minimum boundary deviation, define the target candidate judgment quality level as the minimum deviation quality level, and generate the mineral quality assessment result.

8. The method for evaluating the quality of coal-bearing rare metal minerals according to claim 7, characterized in that, The process of determining the candidate quality level is as follows: The weighted concentration feature value is called, and the two grade boundary points adjacent to the weighted concentration feature value on the numerical axis are retrieved from the quality grading critical threshold. One grade boundary point is defined as the upper critical point, and the other grade boundary point is defined as the lower critical point. Extract the attribution levels defined by the upper and lower critical points respectively, and set these two attribution levels together as the candidate quality levels for judgment.

Images