CN109102035B - Clustering analysis-based coking coal multi-dimensional index similarity refined classification method - Google Patents

Clustering analysis-based coking coal multi-dimensional index similarity refined classification method Download PDF

Info

Publication number
CN109102035B
CN109102035B CN201811055912.3A CN201811055912A CN109102035B CN 109102035 B CN109102035 B CN 109102035B CN 201811055912 A CN201811055912 A CN 201811055912A CN 109102035 B CN109102035 B CN 109102035B
Authority
CN
China
Prior art keywords
classification
index
coal
value
coking
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201811055912.3A
Other languages
Chinese (zh)
Other versions
CN109102035A (en
Inventor
刘洋
白金锋
张志华
李超
王宝科
臧孝
师德谦
钟祥云
张雅茹
徐君
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Science and Technology Liaoning USTL
Original Assignee
University of Science and Technology Liaoning USTL
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by University of Science and Technology Liaoning USTL filed Critical University of Science and Technology Liaoning USTL
Priority to CN201811055912.3A priority Critical patent/CN109102035B/en
Publication of CN109102035A publication Critical patent/CN109102035A/en
Application granted granted Critical
Publication of CN109102035B publication Critical patent/CN109102035B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/17—Function evaluation by approximation methods, e.g. inter- or extrapolation, smoothing, least mean square method
    • G06F17/175—Function evaluation by approximation methods, e.g. inter- or extrapolation, smoothing, least mean square method of multidimensional data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Data Mining & Analysis (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Educational Administration (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Strategic Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Development Economics (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Algebra (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Game Theory and Decision Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Software Systems (AREA)
  • Databases & Information Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Coke Industry (AREA)

Abstract

The invention relates to a clustering analysis-based coking coal multi-dimensional index similarity refined classification method, which combines a clustering analysis algorithm to carry out multi-level multi-dimensional index similarity refined classification on coking coal according to a coal quality index and a coke quality index so as to realize the refined classification of the coking coal with applicability as a target; the method comprises the following steps: 1) dividing indexes used by classifying the coking coal into five levels; 2) calculating and analyzing the similarity of the coking coal by adopting a clustering analysis mathematical algorithm; 3) the refined classification of the coking coal is combined by classification indexes of different levels and is finished by 4 calculation levels; the invention solves the problems that the processing method for simplifying or integrating various indexes is greatly influenced by subjective factors when the classification of coking coal is refined in the prior art, the difference of actual coal quality and coke quality indexes among the coking coals is lost, the similarity relation among the same subclass of coal is not clear, and the like; the method provides important support for refining and classifying the coking coal of the coking enterprise, improving the environmental protection of the coal yard and effectively utilizing the coking coal resources.

Description

Clustering analysis-based coking coal multi-dimensional index similarity refined classification method
Technical Field
The invention relates to the coal chemical industry in the field of steel, in particular to a clustering analysis-based method for refining and classifying similarity of multi-dimensional indexes of coking coal.
Background
With the development of large-scale blast furnaces and oxygen coal injection technology, blast furnaces not only require high coke quality, but also require stable coke quality; under the condition that coking coal resources are increasingly tense, the coke oven has various coal resources and complex coal quality, and the usage amount of imported coking coal which has certain difference with domestic coal in coking characteristics is gradually increased. The above-mentioned circumstances cause many difficulties in the use of coking coals, the production organization and the control of coke quality. The problems of reasonable classification and effective utilization of coking coal resources are faced in coking production, and the problems of environmental protection modification of a coal yard and effective utilization of the coking coal resources by adopting a novel silo integrating storage and matching and the like are implemented under the current environmental protection pressure.
The existing Chinese coal classification standard GB/T5751-containing 2009 is a Chinese coal classification standard established on the basis of the research of integral Chinese coal resources, and the standard is applied to specific coking enterprises and has the defects that: the dividing range of the coking coal is wider, the coal quality and the coke quality of the same type of coking coal may have larger difference, and the properties of the coking coal of different types may be similar at the dividing junction; when the coal type is discriminated, the selected coal sample is required to be single coal layer coal or mixed coal sample having the same deterioration degree, and coking coals having different deterioration degrees are not included in the range. Coking coal classification of coking enterprises is further refined into a plurality of subclasses according to the existing Chinese coal classification and the characteristics of coal sources of coking enterprises on the basis, but because of various coal resources and more indexes of coal quality and coke quality, the coking coal classification is refined intuitively and empirically, the influence of subjective factors is greater, and the similarity among a plurality of coals in the same subclass is not evaluated clearly. Therefore, it is necessary to develop a method for classifying coking coal in a refined manner, which is suitable for coking enterprises.
Chinese patent No. CN201510492987.8 (document 1) proposes "a method for classifying coking raw material applicability and evaluating comprehensive quality". The coking raw material indexes and the quality indexes of solid products obtained by coking are integrated into the binding capacity, the coking capacity and the thermal state performance for classification, and are integrated with the ash content and sulfur content indexes to obtain a comprehensive index, and the performance-price ratio is obtained after each index is associated with the price to evaluate the coking coal. The method has the following defects: the primary classification still uses Chinese coal classification standards, although various abilities exceed the set range of the type, classification adjustment can be carried out and various abilities can be recalculated, the classification adjustment depends on three sub-ability indexes of bonding ability, coking ability and thermal state performance after integration, and the three sub-ability indexes after integration can not display the difference of various specific coal quality and coke quality indexes inside; and after the coking coals of different types are preliminarily classified by national standards, specific parameters specifically used in the calculation process of the three subentry capacity index values have differences, comparability among the three subentry capacity indexes of the coking coals of different types is reduced, and specific adjustment to which category cannot be clear. In addition, the similarity relationship between several coals of the same group is unclear.
Chinese patent No. CN201410195065.6 (document 2) proposes "a method for classifying and blending coal based on coking property of coking coal", which uses vitrinite average maximum reflectance, kirschner maximum fluidity, solid-soft temperature range and coke optical structure as indexes of coking property of coking coal, and accordingly classifies single coal into gas fat coal, gas coal, fat coal, 1/3 coking coal, lean coal, poor mixed coal or special cause coal, and provides classification index classification range.
Chinese patent No. CN201410335178.1 (document 3) proposes "a coking coal subdivision method based on coking property", which selects several indexes in vitrinite average maximum reflectance, kirschner maximum fluidity, solid-soft temperature interval and coke optical organization structure to perform refined classification according to different coking coal deterioration degrees, and provides classification index division range.
The disadvantages of documents 2 and 3 are: the classification index lacks volatile component V controlled in coal blending productiondafIndex, and the volatile component V of coking coal with approximate reflectivitydafThere may also be differences in the above; the crushing strength M directly representing the coking characteristics of coking coal is lacked in the classification index40Abrasion resistance M10And thermal property reactivity CRI and post-reaction strength CSR indexes, and coking coals outside the investigation range of the method are divided according to the method and the indexes due to the complexity of the coking coals, so that the coking characteristics of the same group of coking coals can have larger difference, and the applicability of the method is limited. In addition, the similarity relationship between several coals of the same group is unclear.
The method uses various types of different coal quality and coke quality indexes and applies a modern mathematical algorithm to refine and classify the coking coal, and solves the technical problems that the visual and empirical refined coking coal classification is greatly influenced by subjective factors, the difference of the actual coal quality and coke quality indexes among the coking coals is lost by simplifying or integrating the processing method of each index, the similarity relation among a plurality of coals of the same subclass of coal is not clear, and the like, and the technology in the aspect is not reported.
Disclosure of Invention
The invention provides a method for refining and classifying the similarity of multi-dimensional indexes of coking coal based on cluster analysis, which is used for refining and classifying the similarity of multi-dimensional indexes of coking coal at multiple levels based on the quality indexes of coal and coke and by combining a cluster analysis algorithm, solves the problems that the classification of the coking coal is greatly influenced by subjective factors through intuitive and empirical refining, the processing method for simplifying or integrating various indexes loses the difference of the quality indexes of the actual coal and the coke among the coking coals and the similarity among a plurality of coals of the same subclass of coal is not clearly evaluated and the like in the prior art,
in order to achieve the purpose, the invention adopts the following technical scheme:
a clustering analysis-based coking coal multi-dimensional index similarity refined classification method is characterized in that on the basis of the existing national standard about coal classification, according to a coal quality index and a coke quality index, a clustering analysis mathematical algorithm is combined to carry out multilevel multi-dimensional index similarity refined classification on coking coal, so that the coking coal refined classification with applicability as a target is realized; the current national standard related to coal classification refers to GB/T5751-; weak sticky coal in 1/2, gas coal, gas fat coal, 1/3 coking coal, fat coal, coking coal, lean coal and lean coal; the method specifically comprises the following steps:
1) dividing indexes used by classifying the coking coal into five levels;
classification index T of the first level1Volatile component V is indicated by coking coal deterioration degree indexdafVitrinite mean maximum reflectance
Figure GDA0003425618220000032
And variance S;
classification index T of the second level2Consists of a coal cohesiveness index G value, a gelatinous layer thickness Y value, an Aoardomorphism degree b value and a Gieseler fluidity MF value;
classification index T of the third level3Consists of coal rock microscopic components and an active-inert ratio;
classification index T of the fourth level4From experimental coke oven coke quality indicators including crushing strength M40Abrasion resistance M10Reactive CRI, post-reaction strength CSR, coke average particle size and ash content index composition;
level fiveClassification index T of5From coking coal and ash A of cokedAnd sulfur content St,dIndex composition, namely, coking coal which cannot be coked in an experiment is subjected to coking by adding basic caking coal to obtain a coke quality index;
2) calculating and analyzing the similarity of the coking coal by adopting a clustering analysis mathematical algorithm;
the normalization of the cluster analysis data adopts a mean value normalization method, and the conversion function z is shown as formula (1):
conversion function z ═ (X-Mean)/(Standard deviation) (1)
Wherein X represents certain coal quality or coke quality index data of certain coking coal, Mean represents the Mean value of certain coal quality or coke quality index data of all coking coals, and Standard deviation represents the Standard deviation of certain coal quality or coke quality index data of all coking coals;
the method for measuring the clustering analysis selects the square Euclidean distance as shown in a formula (2);
Figure GDA0003425618220000031
wherein d (x, y) represents the distance between the two coking coals x and y, xiAnd yiRespectively representing a certain coal quality index of the coking coal x and y or a numerical value after the coke quality index is standardized;
the clustering method selects the intergroup connection or the inter-class average method, i.e. the average value of the distances between every two coking coals in two classes is used as the distance D between the two classespqAs shown in formula (3);
Figure GDA0003425618220000041
wherein G isp，GqRespectively represent two categories of coking coals, each containing np、nqCoking coal;
3) the refined classification of the coking coal is combined by classification indexes of different levels and is finished by 4 calculation levels; firstly, setting a coal quality index and a coke quality index range during analysis of each computational layer according to needs to determine the classification quantity, or directly setting the classification quantity according to the index range used by coking coal used in the current industry; the 1 st computing level uses the classification index corresponding to the first level, the 2 nd computing level uses the classification index of the first level to combine with the classification index of the second level, the 3 rd computing level uses the classification index of the second level, the classification index of the third level combines with the classification index of the fourth level, and the 4 th computing level uses the classification index of the fifth level; the coking coal similarity obtained by the 4 th calculation level is maximum;
the specific classification method is as follows:
the 1 st calculation layer classifies the selected integral coking coal by using the classification indexes of the first layer and combining a clustering analysis method according to the volatile component V in the groupdafDetermining the classification quantity when the maximum value and the maximum difference value in the group are less than 10%, or taking the classification quantity to be less than 5 according to the use index range of the coking coal used in the current industry;
the 2 nd calculation level combines the classification index of the first level and the classification index of the second level, combines a clustering analysis method to carry out thinning classification on each subclass obtained by the 1 st calculation level, and carries out thinning classification according to the volatile component V in the subclassdafThe maximum value and the maximum difference value in the group are less than 3 percent, the maximum value of the bond index G value and the maximum difference value in the group are less than 10, the maximum value of the thickness Y value of the colloidal layer and the maximum difference value in the group are less than 5mm, the maximum value when the Odok expansion b value is less than 100 percent, the maximum value when the Odok expansion b value is less than 40 percent and the maximum difference value in the group are less than 150 percent, determining the classification quantity, or taking the classification quantity as less than 5 types according to the use index range of the coking coal used in the industry at present;
the 3 rd calculation level uses the classification index of the second level, the classification index of the third level and the classification index of the fourth level to combine with a clustering analysis method to carry out refinement classification on each subclass obtained by the 2 nd calculation level, and the subclasses are classified according to the maximum value of the G value of the in-group bonding index and the maximum difference value in the group which are less than 5, the maximum value of the Y value of the thickness of the colloidal layer and the groupThe maximum value when the internal maximum difference value is less than 4mm and the Olympic expansion degree b value is less than 100 percent, the maximum value when the internal maximum difference value is less than 30 percent and the Olympic expansion degree b value is more than 100 percent, the maximum value when the internal maximum difference value is less than 130 percent and the coke quality index M of the experimental coke oven40Maximum value when more than 60% and maximum difference in group less than 5%, M10Determining the classification quantity when the maximum value is less than 15 percent and the maximum difference value in the group is less than 3 percent, the CSR maximum value and the maximum difference value in the group is less than 8 percent, or taking the classification quantity to be less than 5 classes according to the use index range of the coking coal used in the industry at present;
the 4 th calculation level uses the classification index of the fifth level and combines a clustering analysis method to carry out refined classification on each subclass obtained by the 3 rd calculation level, and the subclasses are classified according to the intraclass coking coal or the coke ash A thereofdMaximum value and maximum difference value in group less than 1.0%, sulfur content St,dThe maximum value and the maximum difference value in the group are less than 0.5 percent to determine the classification quantity, or the classification quantity is less than 4 types according to the use index range of the coking coal used in the current industry.
The coking coal deterioration index volatile component V daf10 to 45 percent and the average maximum reflectivity of vitrinite is 0.6 to 2.3 percent.
The cohesiveness index G value of the cohesiveness index is more than or equal to 18.
Among the coke quality indexes of the experimental coke oven, the crushing strength M40Less than 90.0 percent and abrasion resistance M10Less than 35%, the coke reactivity CRI is more than 15%, and the intensity CSR after reaction is less than 78%.
Compared with the prior art, the invention has the beneficial effects that:
1) according to the invention, the representative coal quality index and the coke quality index of the coking coal are selected in a multi-level manner, so that the coal quality characteristic and the coking characteristic of the coking coal are reflected from different angles, and the phenomenon that the difference between the actual coal quality and the coke quality index between the coking coals is weakened due to the simplification or integration treatment of the coal quality index and the coke quality index of the coking coal is avoided;
2) according to the method, a clustering analysis method is adopted to carry out multi-level and multi-dimensional similarity classification on the coking coal, the overall similarity degree among multiple indexes is comprehensively inspected through a mathematical algorithm, and the problem that the coking coal classification is influenced greatly by subjective factors through visual and empirical refinement due to more coking coal quality indexes and coke quality indexes in the prior art can be solved;
3) according to the method, a clustering analysis method is adopted to carry out multi-level and multi-dimensional similarity classification on the coking coal, besides a required classification scheme, a relation graph of similarity degrees among the coking coal can be obtained, and the similarity relation among a plurality of coals of the same subclass of coal is determined;
4) according to the method, the classification result of the coking coal is obtained by a clustering analysis method and setting the coal quality index and the coke quality index range, so that the method has universality and can be used as a basis for evaluating the properties of the coking coal;
5) under the current environmental protection pressure, the invention provides important technical support for the environmental protection reconstruction of the coal yard and the effective utilization of coking coal resources by adopting a novel silo integrating storage and matching.
Drawings
FIG. 1 is a flow chart of a clustering analysis-based method for refining and classifying the similarity of multi-dimensional indicators of coking coal.
Fig. 2 is a schematic diagram (overall) of the classification of the 1 st computational level in the embodiment of the present invention.
FIG. 3 is a schematic diagram of the 1 st computational level classification in an embodiment of the present invention (group 2).
FIG. 4 is a schematic diagram of the 2 nd computational level classification in an embodiment of the present invention (group 3).
FIG. 5 is a schematic diagram of the 3 rd computational level classification in an embodiment of the present invention (group 2.1).
FIG. 6 is a schematic diagram of the 3 rd computational level classification in an embodiment of the present invention (group 2.2).
Detailed Description
The following further describes embodiments of the present invention with reference to the accompanying drawings:
as shown in FIG. 1, the clustering analysis-based coking coal multi-dimensional index similarity refined classification method of the present invention combines a clustering analysis mathematical algorithm to perform multi-level multi-dimensional index similarity refined classification of coking coal based on the existing national standard about coal classification according to coal quality index and coke quality index, so as to realize application-targeted coking coal refined classification; the current national standard related to coal classification refers to GB/T5751-; weak sticky coal in 1/2, gas coal, gas fat coal, 1/3 coking coal, fat coal, coking coal, lean coal and lean coal; the method specifically comprises the following steps:
1) dividing indexes used by classifying the coking coal into five levels;
classification index T of the first level1Volatile component V is indicated by coking coal deterioration degree indexdafVitrinite mean maximum reflectance
Figure GDA0003425618220000062
And variance S;
classification index T of the second level2Consists of a coal cohesiveness index G value, a gelatinous layer thickness Y value, an Aoardomorphism degree b value and a Gieseler fluidity MF value;
classification index T of the third level3Consists of coal rock microscopic components and an active-inert ratio;
classification index T of the fourth level4From experimental coke oven coke quality indicators including crushing strength M40Abrasion resistance M10Reactive CRI, post-reaction strength CSR, coke average particle size and ash content index composition;
classification index T of fifth level5From coking coal and ash A of cokedAnd sulfur content St,dIndex composition, namely, coking coal which cannot be coked in an experiment is subjected to coking by adding basic caking coal to obtain a coke quality index;
2) calculating and analyzing the similarity of the coking coal by adopting a clustering analysis mathematical algorithm;
the normalization of the cluster analysis data adopts a mean value normalization method, and the conversion function z is shown as formula (1):
conversion function z ═ (X-Mean)/(Standard deviation) (1)
Wherein X represents certain coal quality or coke quality index data of certain coking coal, Mean represents the Mean value of certain coal quality or coke quality index data of all coking coals, and Standard deviation represents the Standard deviation of certain coal quality or coke quality index data of all coking coals;
the method for measuring the clustering analysis selects the square Euclidean distance as shown in a formula (2);
Figure GDA0003425618220000061
wherein d (x, y) represents the distance between the two coking coals x and y, xiAnd yiRespectively representing a certain coal quality index of the coking coal x and y or a numerical value after the coke quality index is standardized;
the clustering method selects the intergroup connection or the inter-class average method, i.e. the average value of the distances between every two coking coals in two classes is used as the distance D between the two classespqAs shown in formula (3);
Figure GDA0003425618220000071
wherein G isp，GqRespectively represent two categories of coking coals, each containing np、nqCoking coal;
3) the refined classification of the coking coal is combined by classification indexes of different levels and is finished by 4 calculation levels; firstly, setting a coal quality index and a coke quality index range during analysis of each computational layer according to needs to determine the classification quantity, or directly setting the classification quantity according to the index range used by coking coal used in the current industry; the 1 st computing level uses the classification index corresponding to the first level, the 2 nd computing level uses the classification index of the first level to combine with the classification index of the second level, the 3 rd computing level uses the classification index of the second level, the classification index of the third level combines with the classification index of the fourth level, and the 4 th computing level uses the classification index of the fifth level; the coking coal similarity obtained by the 4 th calculation level is maximum;
the specific classification method is as follows:
the 1 st calculation layer classifies the selected integral coking coal by using the classification indexes of the first layer and combining a clustering analysis method according to the volatile component V in the groupdafDetermining the classification quantity when the maximum value and the maximum difference value in the group are less than 10%, or taking the classification quantity to be less than 5 according to the use index range of the coking coal used in the current industry;
the 2 nd calculation level combines the classification index of the first level and the classification index of the second level, combines a clustering analysis method to carry out thinning classification on each subclass obtained by the 1 st calculation level, and carries out thinning classification according to the volatile component V in the subclassdafThe maximum value and the maximum difference value in the group are less than 3 percent, the maximum value of the bond index G value and the maximum difference value in the group are less than 10, the maximum value of the thickness Y value of the colloidal layer and the maximum difference value in the group are less than 5mm, the maximum value when the Odok expansion b value is less than 100 percent, the maximum value when the Odok expansion b value is less than 40 percent and the maximum difference value in the group are less than 150 percent, determining the classification quantity, or taking the classification quantity as less than 5 types according to the use index range of the coking coal used in the industry at present;
the 3 rd calculation level uses the classification index of the second level, the classification index of the third level and the classification index of the fourth level, and combines a clustering analysis method to carry out refined classification on all subclasses obtained by the 2 nd calculation level, according to the maximum value of the group bonding index G value and the maximum difference value in the group which are less than 5, the maximum value of the thickness Y value of the colloidal layer and the maximum difference value in the group which are less than 4mm, the maximum value of the opal expansion b value which is less than 100 percent, the maximum value of the difference value in the group which is less than 30 percent and the opal expansion b value which is more than 100 percent, the maximum value of the difference value in the group which is less than 130 percent and the coke quality index M of the experimental coke oven40Maximum value when more than 60% and maximum difference in group less than 5%, M10Determining the classification quantity when the maximum value is less than 15 percent and the maximum difference value in the group is less than 3 percent, the CSR maximum value and the maximum difference value in the group is less than 8 percent, or taking the classification quantity to be less than 5 classes according to the use index range of the coking coal used in the industry at present;
the 4 th calculation level uses the classification index of the fifth level and combines the clustering analysis method to refine each subclass obtained by the 3 rd calculation levelClassification according to the intragroup coking coal or its coke ash content AdMaximum value and maximum difference value in group less than 1.0%, sulfur content St,dThe maximum value and the maximum difference value in the group are less than 0.5 percent to determine the classification quantity, or the classification quantity is less than 4 types according to the use index range of the coking coal used in the current industry.
The coking coal deterioration index volatile component V daf10 to 45 percent and the average maximum reflectivity of vitrinite is 0.6 to 2.3 percent.
The cohesiveness index G value of the cohesiveness index is more than or equal to 18.
Among the coke quality indexes of the experimental coke oven, the crushing strength M40Less than 90.0 percent and abrasion resistance M10Less than 35%, the coke reactivity CRI is more than 15%, and the intensity CSR after reaction is less than 78%.
The following examples are carried out on the premise of the technical scheme of the invention, and detailed embodiments and specific operation processes are given, but the scope of the invention is not limited to the following examples. The methods used in the following examples are conventional methods unless otherwise specified.
[ examples ] A method for producing a compound
In the embodiment, the characteristics and price information of coking coal and coke of a coking plant of a certain iron and steel company are taken as an example, the multi-dimensional index similarity refinement classification is carried out on the coking coal, and the basic information is shown in table 1.
TABLE 1 coking coal quality, coke quality index and price
Figure GDA0003425618220000081
Figure GDA0003425618220000091
In this embodiment, the calculation level and the classification index are selected when classifying the coking coal: the 1 st calculation level uses the classification index of the first level, namely the volatile component V of the coking coaldafIndexes; the 2 nd calculation level uses the classification index of the first level, i.e. coking coalVolatile component VdafThe classification indexes of the second level are a cohesiveness index G value, a colloidal layer thickness Y value and an Aoargon expansion degree b value; the 3 rd calculation level uses the classification index of the second level, namely the coking coal cohesiveness index G value and the b value, and the classification index of the fourth level, namely the crushing strength M in the experimental coke quality index40Abrasion resistance M10And reactive CRI, post-reaction intensity CSR.
And (3) obtaining a coking coal classification scheme (see table 2) by using a clustering analysis method through 3 calculation layers, wherein the selected coking coal is classified into 5 major classes and 10 minor classes.
TABLE 2 coking coal classification results
Figure GDA0003425618220000092
Using coking coal volatiles VdafWhen the 1 st computational layer classification of the coking coal is carried out, classifying the coking coal into 3 classes which are respectively the 1 st class, the 2 nd class and the 3 rd class (as shown in figure 2); using coking coal volatiles VdafWhen the 2 nd computational level classification is carried out on the 2 nd type coking coal by the Y value and the b value, the classification is carried out into 3 types, namely 2.1 types, 2.2 types and 2.3 types (shown in figure 3); using coking coal Y value, b value and coke M40、M10The CRI and the CSR respectively classify the coking coals of 2.1 types and 2.2 types into 2 types, namely 2.1.1 types, 2.1.2 types, 2.2.1 types and 2.2.2.2 types (as shown in figures 4 and 5); using coking coal G value and coke M40、M10CRI, CSR values classified the 3.1-class coking coals into 3.1.1-class, 3.1.2-class and 3.1.3-class (as shown in fig. 6). The abscissa in fig. 2 to 6 represents the number of coking coals and the ordinate represents the distance between various types of coking coals, and the relationship of the degree of similarity between the coking coals is also shown in fig. 2 to 6.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.

Claims (4)

1. A clustering analysis-based coking coal multi-dimensional index similarity refined classification method is characterized in that based on the existing national standard about coal classification, according to coal quality indexes and coke quality indexes, a clustering analysis mathematical algorithm is combined to carry out multilevel multi-dimensional index similarity refined classification on coking coal, and the refined classification of the coking coal with applicability as a target is realized; the current national standard related to coal classification refers to GB/T5751-; weak sticky coal in 1/2, gas coal, gas fat coal, 1/3 coking coal, fat coal, coking coal, lean coal and lean coal; the method specifically comprises the following steps:
1) dividing indexes used by classifying the coking coal into five levels;
classification index T of the first level1Volatile component V is indicated by coking coal deterioration degree indexdafVitrinite mean maximum reflectance
Figure FDA0003425618210000012
And variance S;
classification index T of the second level2Consists of a coal cohesiveness index G value, a gelatinous layer thickness Y value, an Aoardomorphism degree b value and a Gieseler fluidity MF value;
classification index T of the third level3Consists of coal rock microscopic components and an active-inert ratio;
classification index T of the fourth level4From experimental coke oven coke quality indicators including crushing strength M40Abrasion resistance M10Reactive CRI, post-reaction strength CSR, coke average particle size and ash content index composition;
classification index T of fifth level5From coking coal and ash A of cokedAnd sulfur content St,dIndex composition, namely, coking coal which cannot be coked in an experiment is subjected to coking by adding basic caking coal to obtain a coke quality index;
2) calculating and analyzing the similarity of the coking coal by adopting a clustering analysis mathematical algorithm;
the normalization of the cluster analysis data adopts a mean value normalization method, and the conversion function z is shown as formula (1):
conversion function z ═ (X-Mean)/(Standard deviation) (1)
Wherein X represents certain coal quality or coke quality index data of certain coking coal, Mean represents the Mean value of certain coal quality or coke quality index data of all coking coals, and Standard deviation represents the Standard deviation of certain coal quality or coke quality index data of all coking coals;
the method for measuring the clustering analysis selects the square Euclidean distance as shown in a formula (2);
Figure FDA0003425618210000011
wherein d (x, y) represents the distance between the two coking coals x and y, xiAnd yiRespectively representing a certain coal quality index of the coking coal x and y or a numerical value after the coke quality index is standardized;
the clustering method selects the intergroup connection or the inter-class average method, i.e. the average value of the distances between every two coking coals in two classes is used as the distance D between the two classespqAs shown in formula (3);
Figure FDA0003425618210000021
wherein G isp，GqRespectively represent two categories of coking coals, each containing np、nqCoking coal;
3) the refined classification of the coking coal is combined by classification indexes of different levels and is finished by 4 calculation levels; firstly, setting a coal quality index and a coke quality index range during analysis of each computational layer according to needs to determine the classification quantity, or directly setting the classification quantity according to the index range used by coking coal used in the current industry; the 1 st computing level uses the classification index corresponding to the first level, the 2 nd computing level uses the classification index of the first level to combine with the classification index of the second level, the 3 rd computing level uses the classification index of the second level, the classification index of the third level combines with the classification index of the fourth level, and the 4 th computing level uses the classification index of the fifth level; the coking coal similarity obtained by the 4 th calculation level is maximum;
the specific classification method is as follows:
the 1 st calculation layer classifies the selected integral coking coal by using the classification indexes of the first layer and combining a clustering analysis method according to the volatile component V in the groupdafDetermining the classification quantity when the maximum value and the maximum difference value in the group are less than 10%, or taking the classification quantity to be less than 5 according to the use index range of the coking coal used in the current industry;
the 2 nd calculation level combines the classification index of the first level and the classification index of the second level, combines a clustering analysis method to carry out thinning classification on each subclass obtained by the 1 st calculation level, and carries out thinning classification according to the volatile component V in the subclassdafThe maximum value and the maximum difference value in the group are less than 3 percent, the maximum value of the bond index G value and the maximum difference value in the group are less than 10, the maximum value of the thickness Y value of the colloidal layer and the maximum difference value in the group are less than 5mm, the maximum value when the Odok expansion b value is less than 100 percent, the maximum value when the Odok expansion b value is less than 40 percent and the maximum difference value in the group are less than 150 percent, determining the classification quantity, or taking the classification quantity as less than 5 types according to the use index range of the coking coal used in the industry at present;
the 3 rd calculation level uses the classification index of the second level, the classification index of the third level and the classification index of the fourth level, and combines a clustering analysis method to carry out refined classification on all subclasses obtained by the 2 nd calculation level, according to the maximum value of the group bonding index G value and the maximum difference value in the group which are less than 5, the maximum value of the thickness Y value of the colloidal layer and the maximum difference value in the group which are less than 4mm, the maximum value of the opal expansion b value which is less than 100 percent, the maximum value of the difference value in the group which is less than 30 percent and the opal expansion b value which is more than 100 percent, the maximum value of the difference value in the group which is less than 130 percent and the coke quality index M of the experimental coke oven40Maximum value when more than 60% and maximum difference in group less than 5%, M10Determining the classification quantity when the maximum value is less than 15 percent and the maximum difference value in the group is less than 3 percent, the CSR maximum value and the maximum difference value in the group is less than 8 percent, or taking the classification quantity to be less than 5 classes according to the use index range of the coking coal used in the industry at present;
the 4 th calculation level uses the classification index of the fifth level and combines a clustering analysis method to carry out refined classification on each subclass obtained by the 3 rd calculation level, and the subclasses are classified according to the intraclass coking coal or the coke ash A thereofdMaximum value and maximum difference value in group less than 1.0%, sulfur content St,dThe maximum value and the maximum difference value in the group are less than 0.5 percent to determine the classification quantity, or the classification quantity is less than 4 types according to the use index range of the coking coal used in the current industry.
2. The clustering-analysis-based coking coal multi-dimensional index similarity refined classification method according to claim 1, characterized in that the coking coal deterioration degree index volatile component Vdaf10 to 45 percent and the average maximum reflectivity of vitrinite is 0.6 to 2.3 percent.
3. The clustering analysis-based coking coal multi-dimensional index similarity refined classification method according to claim 1, characterized in that the caking index G value of the caking property index is not less than 18.
4. The clustering analysis based refined classification method for similarity of multi-dimensional indicators of coking coal according to claim 1, characterized in that the crushing strength M in the quality indicators of experimental coke oven coke40Less than 90.0 percent and abrasion resistance M10Less than 35%, the coke reactivity CRI is more than 15%, and the intensity CSR after reaction is less than 78%.
CN201811055912.3A 2018-09-11 2018-09-11 Clustering analysis-based coking coal multi-dimensional index similarity refined classification method Active CN109102035B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811055912.3A CN109102035B (en) 2018-09-11 2018-09-11 Clustering analysis-based coking coal multi-dimensional index similarity refined classification method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811055912.3A CN109102035B (en) 2018-09-11 2018-09-11 Clustering analysis-based coking coal multi-dimensional index similarity refined classification method

Publications (2)

Publication Number Publication Date
CN109102035A CN109102035A (en) 2018-12-28
CN109102035B true CN109102035B (en) 2022-03-04

Family

ID=64865825

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811055912.3A Active CN109102035B (en) 2018-09-11 2018-09-11 Clustering analysis-based coking coal multi-dimensional index similarity refined classification method

Country Status (1)

Country Link
CN (1) CN109102035B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112070368A (en) * 2020-08-20 2020-12-11 中国煤炭地质总局勘查研究总院 Method and system for determining clean utilization mode of coal
CN112098263B (en) * 2020-09-14 2022-07-01 山西亚鑫新能科技有限公司 Method for parameter comprehensive prediction of coke thermal strength model
CN114369471B (en) * 2021-12-27 2022-11-01 乌海市华信煤焦化有限公司 Chamber type coking method for improving coke strength by using lean coal
CN115646848A (en) * 2022-10-21 2023-01-31 微山金源煤矿 Intelligent classification method and system for coal mine
CN116047023A (en) * 2023-02-10 2023-05-02 淮北矿业股份有限公司淮北选煤厂 Coking coal cohesiveness detection system based on big data

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4310412A (en) * 1977-10-07 1982-01-12 Nippon Steel Corporation Method for classification of coals for coke production
SU1833755A1 (en) * 1991-03-04 1993-08-15 Kh Polt I Im V I Lenina Method of obtaining coking coal burden
CN102690669A (en) * 2012-05-23 2012-09-26 武汉钢铁(集团)公司 Coking coal quality classification method and coal blending and coking method with participation of coking coal
CN102890145A (en) * 2012-10-22 2013-01-23 辽宁科技大学 Method for performing nonlinear prediction on coke quality on basis of cohesiveness and coal-rock indexes of single coal
CN103952166A (en) * 2014-05-09 2014-07-30 武汉钢铁(集团)公司 Coal quality sorting and coal distributing method based on cokeability of coking coal
CN104140834A (en) * 2014-07-15 2014-11-12 武汉钢铁(集团)公司 Coking coal subdividing method based on cokeability and application of method in coal blending
CN105062531A (en) * 2015-08-12 2015-11-18 中钢集团鞍山热能研究院有限公司 Coking raw material applicability classification, comprehensive quality evaluation and coal blending guiding method
CN108256747A (en) * 2017-12-29 2018-07-06 中国大唐集团科学技术研究院有限公司火力发电技术研究所 Thermal power plant's coal storage coal blending intelligent management method based on K mean cluster algorithm

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4310412A (en) * 1977-10-07 1982-01-12 Nippon Steel Corporation Method for classification of coals for coke production
SU1833755A1 (en) * 1991-03-04 1993-08-15 Kh Polt I Im V I Lenina Method of obtaining coking coal burden
CN102690669A (en) * 2012-05-23 2012-09-26 武汉钢铁(集团)公司 Coking coal quality classification method and coal blending and coking method with participation of coking coal
CN102890145A (en) * 2012-10-22 2013-01-23 辽宁科技大学 Method for performing nonlinear prediction on coke quality on basis of cohesiveness and coal-rock indexes of single coal
CN103952166A (en) * 2014-05-09 2014-07-30 武汉钢铁(集团)公司 Coal quality sorting and coal distributing method based on cokeability of coking coal
CN104140834A (en) * 2014-07-15 2014-11-12 武汉钢铁(集团)公司 Coking coal subdividing method based on cokeability and application of method in coal blending
CN105062531A (en) * 2015-08-12 2015-11-18 中钢集团鞍山热能研究院有限公司 Coking raw material applicability classification, comprehensive quality evaluation and coal blending guiding method
CN108256747A (en) * 2017-12-29 2018-07-06 中国大唐集团科学技术研究院有限公司火力发电技术研究所 Thermal power plant's coal storage coal blending intelligent management method based on K mean cluster algorithm

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
An empirical approach for classification of coal seams with respect to the spontaneous heating susceptibility of Indian coals;H.B. Sahu等;《International Journal of Coal Geology》;20091013;第80卷(第3-4期);第175-180页 *
Weight Determination Method Based on Principal Component Analysis Coking;Chen Hong-jun等;《Advances in Manufacturing Science and Engineering》;20131231;第2469-2473页 *
开滦精煤炼焦特性及焦炭质量预测的研究;高志芳;《中国优秀博硕士学位论文全文数据库 (硕士) 工程科技Ⅰ辑》;20051215;第B017-2页 *
思维进化和支持向量机理论及其在炼焦配煤优化中的应用研究;吕青;《中国优秀博硕士学位论文全文数据库(博士) 信息科技Ⅰ辑》;20110815(第08期);第B017-3页 *
模糊聚类分析优化炼焦配煤的研究;单晓云等;《煤炭科学技术》;20050630;第33卷(第6期);第68-71页 *

Also Published As

Publication number Publication date
CN109102035A (en) 2018-12-28

Similar Documents

Publication Publication Date Title
CN109102035B (en) Clustering analysis-based coking coal multi-dimensional index similarity refined classification method
CN105062531B (en) Coking raw material application is classified and Quality evaluation and its instructs blending method
CN102890145B (en) Method for performing nonlinear prediction on coke quality on basis of cohesiveness and coal-rock indexes of single coal
CN102890144B (en) Method for predicting coke quality through nonlinear optimization coal blending based on coal rock vitrinite total reflectance
CN104449778B (en) A kind of method that integrated use coal petrography index carries out the exploitation of coal source
CN111647416A (en) Coal blending method based on vitrinite reflectance distribution of blended coal
CN104140834A (en) Coking coal subdividing method based on cokeability and application of method in coal blending
CN112001117A (en) Coking single coal proportioning optimization method
CN111117669B (en) Method for distinguishing classification of coking coal types and application thereof
CN110491454B (en) Blast furnace smelting cost management method and system and computer-storable medium
CN109064061A (en) A kind of coking coal multidimensional property evaluation method based on AHP analytic hierarchy process (AHP)
CN114662763A (en) Method and system for evaluating cost performance of single coal for coking coal blending
CN116479195B (en) Method for optimizing blast furnace injection coal blending based on coal rock parameters
CN104678075A (en) Predicating method for abrasive resistance of coal-blending coking coke
CN112322314B (en) Method for improving coke lumpiness by using coal colloid performance and functional group structural characteristics
Ghosh et al. Understanding the impact of coal blend properties on the coke strength
CN115359851A (en) Multi-objective prediction optimization method for sintering burdening based on extreme random tree-NSGA-II
CN114757543A (en) Compost process efficiency evaluation method
CN105713632B (en) Coal blending method mainly using fat coal
Chia A comprehensive set of idiograms representing all interpretive levels of resolution: ISCN (2009).
CN113969177B (en) Coke matched with oxidative deterioration coking coal and coking method
CN113817489B (en) Coal blending method based on coke optical characteristic index
Singh et al. Microtextural analysis of coke from single and binary blend and its impact on coke quality
CN115368917B (en) Method for predicting tamping coke coal blending structure
CN111932143A (en) Coking coal mixing judgment method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant