BE1030200B1 - Method for distinguishing different calc-silicate rock deposits based on variations in garnet composition - Google Patents

Abstract

The present invention provides a method for distinguishing different calc-silicate rock deposits based on variations in garnet composition, comprising: acquiring training sample data, sifting through trace element content using systematic cluster analysis, and using as model factors for determining ore species type; determining canonical discriminant functions for ore species type and trace element content using canonical discriminant analysis, screening the optimal combination of canonical discriminant functions, creating a supervised classification model; Determining the ore species type of the deposit to be examined using the supervised classification model. The technical solution provided by the present invention has the following advantages: the method for determining potential ore species based on the composition of garnet in calc-silicate rock deposits proposes quantitative indicators and formulas of garnet trace element for rapid identification of ore species types, can increase the accuracy of rapid identification of ore species types and the prediction of the target area, and overcomes the problematic points of the traditional method of identifying ore species types in the initial exploration phase with poor efficiency, long period and high cost.

Description

Method for distinguishing different calc-silicate ice deposits based on variations in garnet composition

TECHNICAL AREA

The present invention relates to the technical field of ore exploration, in particular a method for distinguishing between different lime silicate rock

Deposits based on variations in garnet composition.

STATE OF THE ART

As the scale of ore exploration increases, the discovery becomes newer

Deposits are becoming increasingly difficult and there is an urgent need for new ones

to develop mining techniques and methods to guide the search for breakthroughs; one of the most important tasks is to quickly identify potential ore styles. With the development of techniques for analyzing mineral microzones, the use of mineral microzone compositions to identify potential mineral styles is an important direction for mineralization prediction.

The calc-silicate rock deposits include various types such as lead-zinc ores,

Woitram molybdenum ores, copper-molybdenum ores and calcareous silicate rock iron ores, for the

A key scientific and engineering challenge for geologists is to quickly determine the type of ore to be found in the early stages of an exploration.

The traditional method of determining the type of ore to be found is through a combination of large-scale mapping and systematic sample analysis, i.e. through the process of preliminary investigation - general investigation -

Detailed investigation, the resources of the different ore types are finally calculated and the main ore types of the deposits are then precisely determined,

However, the above procedures are on a lower scale with larger ones

Faced with uncertainties, leading to an inaccurate determination of the key

Ore types In polymetallic calc-silicate rock deposits, the further can lead

Exploration targets and directions severely impacted, resulting in long period and high cost, and unable to meet the needs of rapid green ore exploration.

CONTENT OF THE PRESENT INVENTION

In view of this, an embodiment of the present invention sets forth

Methods for distinguishing different calc-silicate rock deposits based on the variations in garnet composition are available to the above

To solve problems,

An embodiment of the present invention provides a method

Distinguishing different calc-silicate rock strata based on variations in garnet composition, comprising the following steps:

S1: Collect training sample data, where the training sample data is the

ore species type and the corresponding trace element content of each sample,

Carrying out a systematic cluster analysis for the trace element gshait,

Setting a threshold for the relative distance of the class, screening the

trace element content, using the screened trace element content ais

model factor to determine ore species type; 82: Determining the canonical discriminant functions for the ore species type and the

Trace element content using canonical discriminant analysis,

Sifting through the optimal combination of the canonical discriminant functions in

Conformity with the Wilk lambda value, the typical correlation of the functions, the significance of the differences between the groups, and the classification effect of each combination of the canonical discriminant functions, creating a

discriminant classification diagram to construct a supervised classification model for garnet species type and trace element content;

S3: Collecting garnet samples from the deposit to be examined, grinding and manufacturing laser disks from the samples, analyzing the

trace element content in each sample using laser exfoliation-inductively-coupled-

Plasma mass spectrometry in situ analysis technique, screening the

Trace element content of the deposits to be examined according to the type of

model factor and insertion into the çesieved discriminant function combination,

Inserting the calculated function variables into the identification diagram and

Determination of the ore species type of the deposits to be examined.

Further, in step S1, the squared Euclidean distance as

Distance measurement method selected, after defining the distance for the systematic cluster analysis, the intergroup linkage method is selected to perform a systematic cluster analysis for the trace element gshal,

Furthermore, the trace element content includes the rare earth element content and the

content of elements with high field strength,

Also in step 82, the equation of the discriminant function combination is:

Function 1: yl = -0.9log La +0.007logPr+ 0.76310g Nd —1.127log Sm +0.99l0g Eu + 0.348102 Gd — 0.609150 75 — 0.6871og Dy +1.089log Ho — 0.07 llog Er -0.686log7m +0.628logVb - 1.022log Lu + 0.91810g Hf +1.088logU — 0.7

Function 2: »p2=8.2711og La +1.655log Pr - 0.556 log Nd + 0.609 log Sm - 2.379 log Eu - 0.988 10g Gd + 0.416 10g Tb +1.594 10g Dy - 0.632 log Ho + 1.253 10g Er - 8.81 110g Tm + 0.671log Yb — 1.117log Lu + 0.796log Hf +0.086log U — 1.525

Further, in step ST, when the types of trace elements in the training sample data are not complete, the training sample is determined by obtaining

Garnet trace element data for a known calc-silicate ice deposit

ore type added,

Furthermore, a data pre-processing for the trace element content before step 51 is carried out before the systematic cluster analysis of the trace element content,

Furthermore, calibration information,

Standard samples and calibration elements were set and element integration intervals determined to filter out anomalous data, with the trace element content being compared to the

Calibration procedure is corrected.

Furthermore, the silicate trace element 2351 is selected as a calibration element.

Furthermore, the trace element content is corrected using a correction method with several external standards - without internal standard - or a simple correction method with external standard in combination with a correction for sensitivity drift.

Furthermore, the acquisition of training sample data in step S1 consists in particular in

Collecting garnet samples, grinding and crafting laser discs from the

samples, analyzing the trace element ash in each sample using the laser

Exfollation inductively coupled plasma mass spectrometry in situ analysis technique and

Determine the ore species type of each sample.

The technical solution provided by the embodiment of the present invention has the following advantages: Analyzing the trace element content using the laser exfoliation inductively coupled plasma mass spectrometry LA-ICP-MS iIn situ analysis technique, performing a systematic cluster analysis, screening the

Trace Element Content, Establishing a Supervised Classification Model for the

Garnet ore species type and trace element content to finally assess the

Ore type to be realized by the chemical composition of the garnet microzones. The description of the change mineral - garnet - in metallogenic

The calc-silicate rock-type systems are transformed from a macroscopic qualitative interpretation to a microscopic quantitative interpretation of those contained therein

Compositional variations raised what the association of

Granal microzone compositional variations with potential

Allowed ore species type reactions on the mine scale. The procedure for determining potential ore types from the composition of garnet in calc-silicate

Deposits proposes quantitative indicators and formulas of garnet trace siemens for rapid identification of ore species types and has a short test time, low

Cost, convenient and environmentally friendly, can effectively shorten the mineral exploration cycle, does not harm the environment, can greatly improve the accuracy of rapid identification of ore species types and forecasting of the target area, overcome the problematic points of the traditional method of identifying

Types of ore species in the initial exploration phase with poor efficiency, a long period and high cost, allows the correspondence of the variation of the chemical composition of the garnet microzones with the reaction of

Ore species types, realized the combination of mineral geochemistry, evaluation of the

Mineralization potential and ore type prediction, provides methodological support for rapid exploration and evaluation of deposits, can be theoretical

basis for further optimization and selection of ore search methods

Deliver deposit level and reduce the exploration risks, the method represents a new and indispensable exploration means and method, offers important indications for the further analysis of the mineralization potential, the deposit genesis and the

ore exploration breakthroughs and exhibits important production and dissemination values,

BRIEF DESCRIPTION OF THE DRAWING

Figure 1 shows a flowchart of an embodiment of the method for

Distinguishing different calc-silicate rock deposits based on variations in garnet composition according to the present invention,

Figure 2 shows a diagram of the data processing of grenai trace elements from the Longen deposit.

Figure 3 shows a diagram of the spectrum of the cluster analysis of

Garnet trace elements from the Longen deposit.

Figure 4 shows a scatterplot of the classifications of the sifted function 1,

Function 2 and Function 3.

Figure 5 shows a diagram of the discriminant classification of different types of calc-silicate rock beds.

DETAILED DESCRIPTION

In connection with figures, the embodiment of the present invention is explained in more detail in the following, so that the object, the technical solutions and the advantages of the present invention become clearer. See FIG. 1, shows an exemplary embodiment of the present invention

Methods for distinguishing different calc-silicate rock deposits based on variations in garnet composition are available, including the following

Steps:

S1: Collect training sample data, where the training sample data is the

ore species type and the corresponding trace element content of each sample,

Carrying out a systematic cluster analysis for the trace element content,

Setting a threshold for the relative distance of the class, screening the

Trace element content, using the sifted trace element content as

model factor to determine ore species type;

The training sample data can be acquired from existing data or through the

Test of samples obtained after collection in the field, collecting from

garnet samples, grinding and fabricating laser discs from the samples, analyzing the trace element content in each sample using the laser exfoliation inductively coupled plasma mass spectrometry in situ analysis technique and determining the

ore species type of each sample.

In particular, the garnet samples are collected, The garnet must be from the

Bedrock is taken where the calc-silicate ice is formed, the sampling site is positioned using GPS with corrected parameters to determine the sampling coordinates, the macroscopic photos of the site are taken and are good

Description of garnet production condition (veined or solid etc.,

production condition measurements) as well as the petrographic features (color,

Comgròle, syngenstic combinations and mineralization features of the

garnet, etc.) is performed to aid in determining the ore species type for the sample.

Grinding and manufacturing of laser discs from the samples, through the microscopic observation, the microscopic petrographic features of the garnet are determined, the optical features, the syngeneic compositions and the peculiarities of the garnet are described in detail and recorded, by means of the laser exfoliation-inductively-coupled- plasma mass spectrometry in situ

Analysis technique, the trace element content of the garnet is analyzed to determine the ore species type of each sample.

The trace mineral content includes the rare earth element content and the content of

elements with high field strength, if the trace element types of the tested sample are not complete, the training sample is obtained by obtaining

Garnet trace siement data for a known calc-silicate ice deposit

Ore species type added to improve the accuracy of the analysis.

After receiving the trace element content, data must be pre-processed for the

Trace element content can be carried out prior to the systematic cluster analysis of the trace element content. in particular, using the ICPMSDataCal software

Calibration information, standard samples and calibration items set and

Element integration interval determined to filter out abnormal data, where the

Trace element content is corrected with the calibration procedure. It will

Silicate trace element 2951 selected as a calibration element, where the

Trace element content using a correction method with several external standards - without internal standard - or a simple correction method with external

standard in combination with a sensitivity drift correction.

Performing a systematic cluster analysis for the trace element content, first an appropriate distance measurement method (e.g. Euclidean distance, Euclidean squared distance, cosine distance, Pearson correlation distance,

Chebyshev distance, Minkowski distance, etc.) selected to define the distance dij for systematic cluster analysis and a suitable method for

Calculation of inter-class distance (e.g. intergroup linking method, intra-group linking method, method of nearest

Neighboring Element, Method of Farthest Neighboring Element, Center of Mass

clustering method, median clustering method, Wald method, etc.} in order to

Calculate class distance Dpg and Dkr and perform a systematic cluster analysis for

Carry out rare earth elements and elements with high field strength, The

Spectrograms that have been subjected to systematic cluster analysis are then thresholded for the relative distance of the class (e.g. 5, 10, 15, etc.). In accordance with the set thresholds, the trace element content is screened, and the elements of the first major class are used as model factors for the

Determination of ore species type screened. A logarithmic transformation is performed on the model factors to eliminate the problem of heteroscedasticity, and the transformed model factors are subjected to information extraction of the supervised classification model.

S2: Determine the canonical discriminant functions for the ore species type and the

trace element content using canonical discriminant analysis,

Sifting through the optimal combination of the canonical discriminant functions in

Conformity with the Wilk lambda value, the typical correlation of the functions, the significance of the differences between the groups, and the classification effect of each combination of the canonical discriminant functions, creating a

Discriminant classification diagram to establish a supervised classification module for garnet ore type and trace element content,

The amounts of data for each group of different ore types are counted. The canonical discriminant functions are in accordance with the cumulative

Percentage of eigenvalues are determined and the typical correlations are calculated.

The significance of the differences between the groups is in accordance with the

Wilk lambda values (the closer the value is to 0, the more significant it is

difference between groups, and the closer the value is to 1, the less significant the difference between groups), In accordance with the

number of discriminant functions are standardized Fisher

Calculated discriminant function coefficients to construct a supervised classification model for garnet ore species type and trace element content.

S3: Collecting garnet samples from the deposit to be examined, grinding and producing laser discs from the samples, analyzing the

Trace element content in each sample using laser exfoliation-inductively-coupled-

Plasma mass spectrometry in situ analysis technique, screening the

Trace element gshaits of the deposit to be examined according to the type of

Modslfactors and inclusion in the sifted discriminant function combination,

Inserting the calculated function variables into the classification diagram and

Determination of the ore species type of the deposit to be examined.

Working example 1. Collection of garnet samples

Selecting the Longen Deposit Drill Hole and Surface

garnet samples. During sampling, the following contents are recorded in detail: 2. Analysis and testing of trace elements in the sample

Grinding and making laser discs from the collected samples, through the microscopic observation, the microscopic petrographic features of the garnet are determined, the optical features, the syngeneic compositions and the peculiarities of the garnet are described in detail and recorded, using the laser exfoliation-inductive- coupled-plasma

Mass spectrometry (LA-ICP-MS) in situ analysis technique, the trace element content is analyzed, the trace element gshait includes the content of rare earth elements (such as La,

Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, etc.) and the content of high field strength elements (such as U, Y, Zr, Hf, etc.).

Table 1 Sampling table for the Longen deposit

drilling

sample sample production hole mineral | stand-taking- | X Y lithology tlons- ; i photo number sation | location number depth state { mer

fine-grained

Brass 365 | 3275 | red-brown i 4 solid mineral | ZPO1 | Pit 417 | 438 shell sation

Kalksijkatfeis

EN light yellow- green 365 | 3275 modified, sphalerite, | drilling 2 ZiK201 | 50 garnets- . | ZPC6 450 | 441 | 72465 Pyrite | hole

epidotal

lime silicate ice

Coarse-grained egg-green pyrite- 365 | 3275 Seogran World Cup | Bohr- 3 ZK101 70 Garnet- Massiv Minerali- | ZP08 465 | 413 | hole

actinolithation

Lime silicate rock 3. Data processing of garnet trace siements

The data files obtained are introduced into the ICPMSDataCalSoftware which

Correction information is adjusted for the element content analysis that

Standard assay for analysis of elemental content is set to the National Institute of Standards and Technology (NIST 610 or 612) standard assay that

Calibration element is set to the 2951 silicate trace element (see Figure 2}, the standard sample type and test method have been set, the

Element integration interval is determined to filter out abnormal data that

Trace element content is corrected using a multiple external standard correction method - no internal standard - or a simple external standard correction method in combination with a sensitivity drift correction to ensure the accuracy of the trace element data.

To ensure reliable data sifting, it must be prior to the

Performing the screening ensures that the content of

Rare earth elements (such as La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu etc} and the content of high field strength elements (such as U, Y, Zr, Hf etc .) have a unit of ppm. If the trace elementaries in the training sample data are not complete, the training sample will be supplemented by obtaining garnet trace element data for a calc-silicate rock deposit of a known orearian type. As this type of deposit has not been previously tested, additional data is required , plus North, East and Southwest China, South Tibet, Korea, Australia and others

Locate the data of the content of rare earth elements (La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy,

Ho, Er, Tm, Yb, Lu) and on high field strength elements (U, Y, Zr, HP for garnet LA-

ICP-MS of Cu-{Mo}, W-{Mo}Fe. and Pb-£n deposits collected from calcareous silicate rock. The supplemental data are selected from calcareous silicate rock deposits in China and around the world where the ore types are already known, and the

Ore types cover a wide range (including but not limited to

Cu-{Mo), W-{Mo}, Fe, Pb-Zn etc). 4, Determination of model factors and discriminant functions

The squared Euclidean distance is used as the distance measurement method, and after defining the distance for the systematic cluster analysis, the intergroup linking method is selected to obtain a systematic

Perform cluster analysis for garnet rare earth elements and high field strength elements. The threshold for the relative distance of the class is set to 5 (see Figure 3), and the elements in the first main class (La, Pr, Nd, Sm,

Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, HF, U} are screened as model factors for ore species type determination.

A logarithmic transformation is performed on the filtered model factors to eliminate the problem of heteroscedasticity, the data for each

Group of Cu-(Mo)-lime-silicate rock deposits (lime-silicate rock-copper(molybdenum)-

deposit }, W-(Mo)-lime-silica-fels deposits (calc-silicate-ice-tungsten(molybdenum)}-

Deposit), Fe-Calcosilicate rock (Calcosilicate rock-iron deposit) and Pb-Zn-

Lime-silicate rock deposits (calc-silicate rock-lead-zinc deposit) are counted, using the canonical discriminant analysis, the canonical discriminant functions for the ore species type and the trace element content are determined from the several adjusted

Functions, the first three functions with the minimum Wilk lambda values are selected (i.e. function y1, function y2 and function y3).

Function 1: pl = -0.9log La + 0.007logPr + 6.7631og Nd — 1.127log Sm +0.9910g Eu +0.34810g Gd — 0.609102 7h — 0.6871og Dy +1.069log Ho — 0.07 1log Er — 0.686log Tm + 0.628log yb — 1.622log Lu + 0.918log Hf +1 088log U — 0.7

Function 2: p2=0.271log La +1.655log Pr — 0.556 log Nd + 0.609 log Sm — 2.379 log Eu — 0.988 log Gd + 0.41610g 75 + 1.594 log Dy — 0.632 log Ho + 1.253 log Er — 8.81 1log Tm +0.67 l log Yb - 1.117 log Lu + 0.796 log Hf + 0.086 log U - 1.525

Function 3: Y3 = 0119LOG LA + 1.324 1OG PR— 0.741 Log ND - 1,784 Log + 1.168log EU —1.08310g GD + 1.904 Log TB + 0.236 Log DY + 0.14110G HO + 3.476LOG ER - 2.28LOG TM - 1,866 LOG Yb + 0.632log Lu + 0.369log Hf - 0.448log U +1.529

Sifting through the optimal combination of canonical discriminant functions In

Agreement with the typical correlation of the functions, the significance of the

Differences between the groups and the classification effect of each combination of the canonical discriminant functions (see Table 2, Table 3 and Table 4),

Creating a Discriminant Classification Chart to show a monitored

establish a classification model for garnet ore type and trace element content,

Table 2 Table of eigenvalues of discriminant analysis

variance percentage correlation 1 | 2677 | 77.2 TTET 70.853 2 0.528° 15.2 | 92.4 0.588

Table 3 Wilk Lambda Table of Discriminant Analysis 3 | 8781 | 68,489 13 0000il

BE2023/5019

By combining the cumulative percentage of the eigenvalues and the typical

Correlations is based on the classification effect of the scatterplot of

Function 1, Function 2 and Function 3 {see Figure 4) determines the best combination of the canonical discriminant functions as Function 1 and Function 2. After

Determination of the functions is based on the distribution area of the

scatter plot and the position of the center of mass of each group

discriminant classification chart determined {see Figure 5). After determining the discriminant function, the garnet trace elements of the Longen ore type are sifted by the type of model factor and introduced into the selected discriminant function, and the calculated variable values of the function are entered using the Excel

Software inserted into the cassification diagram to determine the ore species type of the deposit being examined, with the information about the

Mineralization exploration can be effectively extracted from the characteristic minerals, the direction of ore prospecting is clarified and hence the ore exploration and the

Exploration of the deposits are conducted, If the calculated values of the

Function variables are not within the scope of different classes of the

classification chart, the garnet trace element data for a

Added a known type of ore to the Kakisilicatefels deposit, then the

Steps 3-5 continued,

The garnet is the predominant mineral in the calc-silicate rock deposits and the

Garnet composition varies greatly between different ore types, offering the opportunity to determine the ore type using the

To determine garnet compositions, The present invention analyzes the

Trace element content using laser exfoilation inductively coupled plasma

Mass spectrometry LA-ICP-MS in situ analysis technique, performs a systematic

Run cluster analysis, sift through trace element levels, set up a monitor

Classification model for the granal ore species type and the trace element type, to finally assess the ore species type by the chemical composition of the

To realize garnet microzones, Description of the change mineral -

Garnet - Raised in calc-silica-type metallogenic systems from a macroscopic qualitative interpretation to a microscopic quantitative interpretation of the compositional variations therein, and the

Microzone composition variations are linked to ore species type responses at the mine scale.

Compared to conventional technology, the method for determining potential ore types based on the composition of garnet in calcareous silicate

Deposits provide quantitative indicators and formulas of garnet trace element for rapid identification of ore species types, and has a short test time, low

Cost, convenient and environmentally friendly, can effectively shorten the mineral exploration cycle, does not harm the environment, can greatly improve the accuracy of rapid identification of ore species types and target area prediction, overcome the problematic points of the traditional method of identifying

Types of ore species in the initial exploration phase with poor efficiency, a long period and high cost, allows the correspondence of the variation of the chemical composition of the garnet microzones with the reaction of

Ore species types, realized the combination of mineral gsochemistry, evaluation of

Mineralization potential and ore style prediction, provides methodological support for rapid exploration and assessment of deposits, may be theoretical

basis for further optimization and selection of ore search methods

Deposit level and reduce exploration risks, the method represents a new and indispensable exploration tool and method, offers important information for the further analysis of the mineralization potential, the deposit genesis and the

Ore exploration breakthroughs and exhibits important production and dissemination values.

In the description, the orientation terms front, back,

Top and bottom in relation to the location of the parts in the attached

Easily defined drawings and the position of the parts in relation to each other

For reasons of clarity and convenience in presenting the technical solution. It goes without saying that the use of the orientation terms mentioned in this

application should not limit the scope of protection claimed,

The above embodiments and the features in the embodiments can be combined with each other in the case of no conflicts.

Only preferred embodiments of the present invention are explained above, but the present invention is not limited thereto, all changes made under the spirit and principles of the present invention are equivalent

Replacements and improvements should be considered as being within the scope of the present

Invention covered

Claims (10)

EXPECTATIONS 1. A method for distinguishing different calc-silicate rock deposits based on variations in garnet composition, characterized in that it comprises the following steps: S1: collecting training sample data, the training sample data comprising the ore species type and the corresponding trace element content of each sample, performing a systematic cluster analysis for the trace element content, setting a threshold for the relative distance of the class, screening the trace element content, using the screened trace element content as a model factor for determining the ore species type; S2: Determine the canonical discriminant functions for ore species type and trace element content using canonical discriminant analysis, screening the optimal combination of canonical discriminant functions in accordance with the Wilk lambda value, the typical correlation of the functions, the significance of the differences between the groups and the classification effect of each combination of the canonical discriminant functions, constructing a discriminant classification diagram to establish a supervised classification model for garnet ore species type and trace element content; S3: Collect garnet samples from the deposit to be tested, grind and laser slice the samples, analyze the trace element content in each sample using the laser exfoliation inductively coupled plasma mass spectrometry in situ analysis technique, screen the trace element content of the to deposit to be examined according to the type of model factor and inclusion in the sifted discriminant function combination, insertion of the calculated function variables into the classification diagram and determination of the ore species type of the deposit to be examined. 2. Method for distinguishing different calc-silicate rock deposits based on the variations in garnet composition according to claim 1, characterized in that in step S1 the squared Euclidean distance is selected as the distance measurement method, after defining the distance for the systematic cluster analysis the intergroup linking method is selected used to perform a systematic cluster analysis for trace element content. 3. A method for distinguishing different calc-silicate rock deposits based on variations in garnet composition according to claim 1, characterized in that the trace element content comprises the rare earth element content and the high field strength element content. 4. Method for distinguishing different calc-silicate rock deposits based on variations in garnet composition according to claim 3, characterized in that in step S2 the equation of the discriminant function combination is: Function 1: yl = -0.910g La + 0.007logPr + 0.763log Nd -1.127 log Sm + 0.9910g Eu + 0.3481og Gd — 0.60910g7b — 0.687l0g Dy + 1.069l1o0g Ho — 0.07 1log Er — 0.686log Tm + 0.628 10g Yb — 1.02210g Lu +0.918l0g Hf +1.088logU — 0.7 Function 2: y2 = 0.271log La +1.655log Pr — 0.556log Nd + 0.60910g Sm — 2.379 log Eu — 0.988 log Gd + 0.41610g 7b + 1.594 log Dy — 0.632 log Ho + 1.253 log Er — 0.8111og Tm + 0.671log Yb — 1.1171l0g Lu + 0.796log Hf +0.086log U —1.525 5. Method for distinguishing different calc-silicate rock deposits based on the variations in garnet composition according to claim 1, characterized in that in step S1, if the trace element types in the training sample data are not complete, the training sample by obtaining garnet trace element data for a calc-silicate rock deposit of a known type of ore is added. 6. Method for distinguishing different calc-silicate rock deposits based on the variations in garnet composition according to claim 1, characterized in that a data pre-processing for the trace element content is performed before step S1 before the systematic cluster analysis of the trace element content. 7. Method for distinguishing different calc-silicate rock deposits based on the variations in garnet composition according to claim 6, characterized in that calibration information, standard samples and calibration elements are set using the ICPMSDataCal software and element integration intervals are determined in order to filter out anomalous data, the trace element content being compared with the Calibration procedure is corrected. 8. A method for distinguishing different calc-silicate rock deposits based on variations in garnet composition according to claim 7, characterized in that the silicate trace element 29Si is selected as the calibration element. 9. Method for distinguishing different calc-silicate rock deposits based on the variations in the garnet composition according to claim 7, characterized in that the trace element content is determined by means of a correction method with several external standards - without internal standard - or a simple correction method with external standard in combination with a correction the sensitivity drift is corrected. 10. Method for distinguishing different calc-silicate rock deposits based on the variations in the garnet composition according to claim 1, characterized in that the acquisition of training sample data in step S1 consists in particular in collecting garnet samples, grinding and producing laser discs from the samples, analyzing the trace element content in each sample using the laser exfoliation inductively coupled plasma mass spectrometry in situ analysis technique and determining the ore species type of each sample.