CN109102035A - A kind of coking coal multidimensional index similitude classification method based on clustering - Google Patents
A kind of coking coal multidimensional index similitude classification method based on clustering Download PDFInfo
- Publication number
- CN109102035A CN109102035A CN201811055912.3A CN201811055912A CN109102035A CN 109102035 A CN109102035 A CN 109102035A CN 201811055912 A CN201811055912 A CN 201811055912A CN 109102035 A CN109102035 A CN 109102035A
- Authority
- CN
- China
- Prior art keywords
- coal
- classification
- index
- level
- 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.)
- Granted
Links
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 coking coal multidimensional index similitude classification method based on clustering that the present invention relates to a kind of, according to coal index and coke quality index, the multi-level multidimensional index similitude classification of coking coal is carried out in conjunction with cluster algorithm, is realized using application as the coking coal classification of target;It include: that the index that coking coal classification uses 1) is divided into five levels;2) the calculating analysis of coking coal similitude is carried out using clustering mathematical algorithm;3) classification of coking coal is combined by the classification indicators of different levels, and point 4 calculating levels are completed;The present invention solves the processing method for being influenced by subjective factor when refinement coking coal classification in the prior art and simplifying or integrate indices and loses the problems such as similarity relationships are indefinite between actual coal and coke quality index otherness and same group coal between coking coal;It is utilized for coal chemical enterprise coking coal classification and coal yard environmental transformation, effective use coking coal resource provides important support.
Description
Technical field
The present invention relates to iron and steel domain coal chemical industry more particularly to a kind of coking coal multidimensional indexes based on clustering
Similitude classification method.
Background technique
With the development of blast furnace enlargement and Oygen And Coal Blowing Technology, blast furnace does not require nothing more than coke quality height, and requires burnt
Charcoal quality is stablized;In the case where coking coal resource growing tension, the various in style of coke oven coal resource, coal quality are complicated, with state
The import coking coal usage amount of interior coal there is some difference on coking property property is also gradually increasing.Above situation is to coking coal
Many difficulties are brought using, organization of production and coke quality control.Coking coal resource Rational Classification is faced in coking production and is had
Utilizing question is imitated, and the coal yard environmental transformation implemented under current environmental protection pressure and the novel silo using storage with unification
The problems such as efficiently using coking coal resource.
Existing Chinese Coal Classification standard GB/T 5751-2009 is established by Research foundation of whole coal resources in China
Chinese Coal Classification standard, the standard be applied to specific coal chemical enterprise when be disadvantageous in that: coking coal divide range ratio
Broad, there may be larger difference, different classes of coking coal is dividing friendship for same category coking coal coal quality and coke quality
At boundary may also property it is close;When the classification for carrying out coal differentiates, it is desirable that selected coal sample is single coal bed coal or identical rotten journey
The mixing coal sample of degree, then be not included within the scope of it different rank coking coal.Coal chemical enterprise coking coal classification one
As several groups are further refined as according to itself coal source feature according to existing Chinese Coal Classification and on this basis, but due to
Coal resource is various in style, and the index of coal quality and coke quality is also more, refines coking coal classification by subjectivity by intuitive and experience
Factor is affected, and if the similitude between same group coal between dry coal do not obtain explicit evaluation.Therefore, exploitation is suitable
The coking coal classification method for closing coal chemical enterprise is very necessary.
A kind of " coking raw material application classification is proposed application No. is the Chinese patent of CN201510492987.8 (document 1)
With Quality evaluation method ".The solid product quality index that coking feedstock specifications and its coking obtain is integrated into and coheres energy
Power, coking ability, hot performance obtain an overall target, each index and valence with ash content and sulphur content aggregation to classify
Cost performance is obtained after lattice association to evaluate coking coal.The shortcoming of this method are as follows: preliminary classification still uses Chinese Coal Classification
Standard, although every ability can be carried out when exceeding the range of this type set classification adjustment and and recalculate every ability,
What classification adjustment relied on is to cohere three ability, coking ability, hot performance subitem capacity indexes after integrating, and pass through integration
Three subitem capacity indexes afterwards cannot show the difference of the inside every specific coal quality and coke quality index;And through state
Different types of coking coal after preliminary classification is marked, specifically used design parameter has been in three subitem capacity index value calculating process
Variant, the comparativity between different type coking coal three subitem capacity indexes reduces, and is specifically adjusted to inside which classification
It cannot define.In addition, if the similarity relationships between same group of dry coal are indefinite.
Application No. is the Chinese patent of CN201410195065.6 (document 2) propose it is " a kind of based on coking coal coking capacity
Vitrinite is averaged maximum reflectivity, Ji Shi maximum fluidity, solid-softening temperature section and coke by coal quality classification and blending method "
Single grade coal is divided into gas rich coal, bottle coal, rich coal, 1/3 coke as the index for coking coal coking capacity, and according to this by optical texture
Coal, coking coal, lean coal and mixed coal inferior or special genetic coal, give classification indicators and divide range.
Application No. is the Chinese patent of CN201410335178.1 (documents 3) to propose a kind of " coking coal based on coking capacity
Divided method ", according to the difference of coking rank of coal metamorphism, select vitrinite be averaged maximum reflectivity, Ji Shi maximum fluidity,
Gu several indexs in-softening temperature section and charred coal organization structure carry out classification to it, gives classification indicators and draw
Divide range.
Document 2 and document 3 are disadvantageous in that: lacking the volatile matter V controlled in coal blending production in the classification indicatorsdaf
Index, and the close coking coal of reflectivity is in volatile matter VdafOn be also likely to be present difference;Lack directly characterization refining in classification indicators
The crushing strength M of coking coal coking property40, wear-resistant strength M10With thermal property reactivity CRI, post reaction strength CSR index, due to
The complexity of coking coal, the coking coal investigated other than range for this method are divided according to this method and index, and same group
The coking property of coking coal is also likely to be present larger difference, is restricted the applicability of method.If in addition, same group of dry coal
Between similarity relationships it is indefinite.
Coking coal classification is carried out using polymorphic type difference coal quality and coke quality index and with modern mathematics algorithm,
Solve the processing side that the refinement coking coal classification of intuitive and experience was affected by subjective factor and simplified or integrate indices
If method loses the phase between coking coal between the dry coal of actual coal and coke quality index otherness and same group coal again
Like the technical problems such as sexual intercourse is indefinite, technology is had not been reported in this respect.
Summary of the invention
The coking coal multidimensional index similitude classification method based on clustering that the present invention provides a kind of is based on coal
Matter and coke quality index simultaneously combine cluster algorithm, carry out multi-level multidimensional index similitude classification to coking coal,
Solve in the prior art it is existing always by intuitive and experience refine coking coal classify be affected by subjective factor, it is simple
Change or integrate indices processing method lose between coking coal again actual coal and coke quality index otherness and
If the problems such as similitude between the dry coal of same group coal does not obtain explicit evaluation,
In order to achieve the above object, the present invention is implemented with the following technical solutions:
A kind of coking coal multidimensional index similitude classification method based on clustering, existing about coal point
On the basis of class national standard, according to coal index and coke quality index, coking coal is carried out in conjunction with clustering mathematical algorithm
Multi-level multidimensional index similitude classification, realize using application as the coking coal classification of target;Specifically include as
Lower step:
1) index that coking coal classification uses is divided into five levels;
The classification indicators T of first level1By coking rank of coal metamorphism index volatile matter Vdaf, vitrinite it is average maximum anti-
Penetrate rateAnd variance S composition;
The classification indicators T of second level2By coal cohesiveness index caking index G value, thickness of colloidal matter layer Y value, Ah swollen difficult to understand
Expansibility b value and Giseeler fluidity MF value composition;
The classification indicators T of third level3It is made of macerals, lazy ratio living;
The classification indicators T of 4th level4By Experiment Coke Oven coke quality index, including crushing strength M40, wear-resistant strength
M10, reactivity CRI, post reaction strength CSR, coke average grain diameter and ash component index composition;
The classification indicators T of 5th level5By coking coal and the ash content A of cokedWith sulphur content St,dIndex composition, to testing
In be unable to the coking coal of coking by obtaining coke quality index with addition of base bond coal coking;
2) the calculating analysis of coking coal similitude is carried out using clustering mathematical algorithm;
Mean value standardized method is selected in the standardization of clustering data, shown in transfer function z such as formula (1):
Transfer function z=(X-Mean)/(Standard deviation) (1)
Wherein, X indicates some coal quality or coke quality index data of some coking coal, and Mean indicates all coking coals
The mean value of some coal quality or coke quality index data, Standard deviation indicate all coking coals some coal quality or
The standard deviation of coke quality index data;
Clustering Measurement Method selects square Euclidean distance, as shown in formula (2);
Wherein, d (x, y) indicates the distance between two coking coals x and y, xiAnd yiRespectively indicate some of coking coal x and y
Numerical value after coal index or coke quality index standardization;
The method of average between connection or class between clustering method selection group, i.e., with distance between coking coals two-by-two all in two classes
Average value is as two between class distance Dpq, as shown in formula (3);
Wherein, Gp, GqTwo classifications for respectively indicating coking coal, respectively contain np、nqA coking coal;
3) classification of coking coal is combined by the classification indicators of different levels, and point 4 calculating levels are completed;First root
According to when needing to be arranged each calculating level analysis coal index and coke quality index range to determine quantity of classifying, or press mesh
The indication range that preceding industry coking coal uses directly sets classification quantity;Specifically:
1st calculating level uses the corresponding classification indicators of first level, and the 2nd calculating level uses first level
Classification indicators and the classification indicators of second level be combined, the 3rd calculate level using second level classification indicators,
The classification indicators of third level and the classification indicators of the 4th level are combined, and the 4th calculating level uses the 5th level
Classification indicators;The coking coal similarity that 4th calculating level obtains is maximum.
The existing national standard about grade of coal refers to GB/T 5751-2009 " Chinese Coal Classification ";It is described
Coking coal covers weakly caking coal therein, sticks coal, bottle coal, gas rich coal, 1/3 coking coal, rich coal, coking coal, lean coal and meager lean coal in 1/2.
The coking rank of coal metamorphism index volatile matter VdafBetween 10%~45%, vitrinite is averaged maximum reflection
Rate is between 0.6%~2.3%.
Value >=18 caking index G of the cohesiveness index.
In the Experiment Coke Oven coke quality index, crushing strength M40 < 90.0%, wear-resistant strength M10 < 35%, coke
Reactive CRI > 15%, post reaction strength CSR < 78%.
When dividing 4 classifications for calculating levels completion coking coal, specific classification method is as follows:
1st calculates level using the classification indicators of first level and combines clustering method to selected whole coking
Coal is classified, according to volatile matter V in organizingdafMaximum value and maximum value difference determine classification quantity, or classification number less than 10%
Amount is taken as by current industry coking coal service index range less than 5 classes;
2nd is calculated level and is combined using the classification indicators of first level with the classification indicators of second level, and tied
It closes clustering method and classification is carried out to each group obtained by the 1st calculating level, according to volatile matter V in organizingdafMaximum value and
Maximum value difference < 3%, caking index G value maximum value and maximum value difference < 10, thickness of colloidal matter layer Y value maximum value and maximum value
Its maximum value and maximum value difference < 40%, difficult to understand Ah's dilation b value > when difference < 5mm, difficult to understand Ah's dilation b value < 100%
Its maximum value and maximum value difference < 150% determine that classification quantity, or classification quantity press current industry coking when 100%
Coal service index range is taken as less than 5 classes;
3rd calculates level and uses the classification indicators, the classification indicators of third level and the 4th layer of second level
Secondary classification indicators combination, and clustering method is combined to carry out classification again to each group obtained by the 2nd calculating level, it presses
According to caking index G value maximum value and maximum value difference < 5, thickness of colloidal matter layer Y value maximum value and maximum value difference < 4mm in organizing,
Its maximum value and it is maximum when maximum value difference < 30%, difficult to understand Ah's dilation b value > 100% when difficult to understand Ah's dilation b value < 100%
Value and maximum value difference < 130%, Experiment Coke Oven coke quality index M40Maximum value and maximum value difference < 5% when > 60%,
M10Maximum value and maximum value difference < 3%, CSR maximum value and maximum value difference < 8% determine classification quantity when < 15%, or
Classification quantity is taken as by current industry coking coal service index range less than 5 classes;
4th calculates level using the classification indicators of the 5th level and combines clustering method to the 3rd computation layer
Each group obtained by face carries out classification again, according to coking coal or its coke ash A in organizingdMaximum value and maximum value difference <
1.0%, sulphur content AdMaximum value and maximum value difference < 0.5% determine that classification quantity, or classification quantity press current industry
Coking coal service index range is taken as less than 4 classes.
Compared with prior art, the beneficial effects of the present invention are:
1) present invention passes through the representative coal index and coke quality index for selecting coking coal at many levels, from different perspectives
Embody coking coal coal characteristic and coking property, avoid due to coking coal coal index and coke quality index carry out letter
Change or integration handles and weakens the difference between coking coal between actual coal and coke quality index;
2) present invention carries out multi-level multidimensional similitude classification to coking coal using clustering method, passes through mathematical algorithm
Whole similarity degree between integrated survey many index is able to solve in the prior art because of coking coal coal index and coke matter
Figureofmerit is more, and refines the problem of coking coal classification is affected by subjective factor by intuitive and experience;
3) present invention carries out multi-level multidimensional similitude classification to coking coal using clustering method, except what is needed
Outside classification schemes, moreover it is possible to the relational graph of similarity degree between coking coal is obtained, if specifying between same group coal dry coal
Similarity relationships;
4) present invention obtains coking coal point by clustering method and setting coal index and coke quality index range
Class can be used as the basis of coking coal properties evaluations as a result, with versatility;
5) under current environmental protection pressure, the present invention is the coal yard environmental transformation implemented and matches the novel of unification using storage
Silo effective use coking coal resource provides important technical support.
Detailed description of the invention
Fig. 1 is the process of the coking coal multidimensional index similitude classification method of the present invention based on clustering
Figure.
Fig. 2 is the 1st calculating level classification schematic diagram (entirety) in the embodiment of the present invention.
Fig. 3 is the 1st calculating level classification schematic diagram (the 2nd group) in the embodiment of the present invention.
Fig. 4 is the 2nd calculating level classification schematic diagram (the 3rd group) in the embodiment of the present invention.
Fig. 5 is the 3rd calculating level classification schematic diagram (the 2.1st group) in the embodiment of the present invention.
Fig. 6 is the 3rd calculating level classification schematic diagram (the 2.2nd group) in the embodiment of the present invention.
Specific embodiment
Specific embodiments of the present invention will be further explained with reference to the accompanying drawing:
As shown in Figure 1, a kind of coking coal multidimensional index similitude classification side based on clustering of the present invention
Method, on the basis of the existing national standard about grade of coal, according to coal index and coke quality index, in conjunction with cluster point
The multi-level multidimensional index similitude classification that mathematical algorithm carries out coking coal is analysed, is realized using application as the coking coal of target
Classification;Specifically comprise the following steps:
1) index that coking coal classification uses is divided into five levels;
The classification indicators T of first level1By coking rank of coal metamorphism index volatile matter Vdaf, vitrinite it is average maximum anti-
Penetrate rateAnd variance S composition;
The classification indicators T of second level2By coal cohesiveness index caking index G value, thickness of colloidal matter layer Y value, Ah swollen difficult to understand
Expansibility b value and Giseeler fluidity MF value composition;
The classification indicators T of third level3It is made of macerals, lazy ratio living;
The classification indicators T of 4th level4By Experiment Coke Oven coke quality index, including crushing strength M40, wear-resistant strength
M10, reactivity CRI, post reaction strength CSR, coke average grain diameter and ash component index composition;
The classification indicators T of 5th level5By coking coal and the ash content A of cokedWith sulphur content St,dIndex composition, to testing
In be unable to the coking coal of coking by obtaining coke quality index with addition of base bond coal coking;
2) the calculating analysis of coking coal similitude is carried out using clustering mathematical algorithm;
Mean value standardized method is selected in the standardization of clustering data, shown in transfer function z such as formula (1):
Transfer function z=(X-Mean)/(Standard deviation) (1)
Wherein, X indicates some coal quality or coke quality index data of some coking coal, and Mean indicates all coking coals
The mean value of some coal quality or coke quality index data, Standard deviation indicate all coking coals some coal quality or
The standard deviation of coke quality index data.
Clustering Measurement Method selects square Euclidean distance, as shown in formula (2);
Wherein, d (x, y) indicates the distance between two coking coals x and y, xiAnd yiRespectively indicate some of coking coal x and y
Numerical value after coal index or coke quality index standardization;
The method of average between connection or class between clustering method selection group, i.e., with distance between coking coals two-by-two all in two classes
Average value is as two between class distance Dpq, as shown in formula (3);
Wherein, Gp, GqTwo classifications for respectively indicating coking coal, respectively contain np、nqA coking coal;
3) classification of coking coal is combined by the classification indicators of different levels, and point 4 calculating levels are completed;First root
According to when needing to be arranged each calculating level analysis coal index and coke quality index range to determine quantity of classifying, or press mesh
The indication range that preceding industry coking coal uses directly sets classification quantity;Specifically:
1st calculating level uses the corresponding classification indicators of first level, and the 2nd calculating level uses first level
Classification indicators and the classification indicators of second level be combined, the 3rd calculate level using second level classification indicators,
The classification indicators of third level and the classification indicators of the 4th level are combined, and the 4th calculating level uses the 5th level
Classification indicators;The coking coal similarity that 4th calculating level obtains is maximum.
The existing national standard about grade of coal refers to GB/T 5751-2009 " Chinese Coal Classification ";It is described
Coking coal covers weakly caking coal therein, sticks coal, bottle coal, gas rich coal, 1/3 coking coal, rich coal, coking coal, lean coal and meager lean coal in 1/2.
The coking rank of coal metamorphism index volatile matter VdafBetween 10%~45%, vitrinite is averaged maximum reflection
Rate is between 0.6%~2.3%.
Value >=18 caking index G of the cohesiveness index.
In the Experiment Coke Oven coke quality index, crushing strength M40 < 90.0%, wear-resistant strength M10 < 35%, coke
Reactive CRI > 15%, post reaction strength CSR < 78%.
When dividing 4 classifications for calculating levels completion coking coal, specific classification method is as follows:
1st calculates level using the classification indicators of first level and combines clustering method to selected whole coking
Coal is classified, according to volatile matter V in organizingdafMaximum value and maximum value difference determine classification quantity, or classification number less than 10%
Amount is taken as by current industry coking coal service index range less than 5 classes;
2nd is calculated level and is combined using the classification indicators of first level with the classification indicators of second level, and tied
It closes clustering method and classification is carried out to each group obtained by the 1st calculating level, according to volatile matter V in organizingdafMaximum value and
Maximum value difference < 3%, caking index G value maximum value and maximum value difference < 10, thickness of colloidal matter layer Y value maximum value and maximum value
Its maximum value and maximum value difference < 40%, difficult to understand Ah's dilation b value > when difference < 5mm, difficult to understand Ah's dilation b value < 100%
Its maximum value and maximum value difference < 150% determine that classification quantity, or classification quantity press current industry coking when 100%
Coal service index range is taken as less than 5 classes;
3rd calculates level and uses the classification indicators, the classification indicators of third level and the 4th layer of second level
Secondary classification indicators combination, and clustering method is combined to carry out classification again to each group obtained by the 2nd calculating level, it presses
According to caking index G value maximum value and maximum value difference < 5, thickness of colloidal matter layer Y value maximum value and maximum value difference < 4mm in organizing,
Its maximum value and it is maximum when maximum value difference < 30%, difficult to understand Ah's dilation b value > 100% when difficult to understand Ah's dilation b value < 100%
Value and maximum value difference < 130%, Experiment Coke Oven coke quality index M40Maximum value and maximum value difference < 5% when > 60%,
M10Maximum value and maximum value difference < 3%, CSR maximum value and maximum value difference < 8% determine classification quantity when < 15%, or
Classification quantity is taken as by current industry coking coal service index range less than 5 classes;
4th calculates level using the classification indicators of the 5th level and combines clustering method to the 3rd computation layer
Each group obtained by face carries out classification again, according to coking coal or its coke ash A in organizingdMaximum value and maximum value difference <
1.0%, sulphur content AdMaximum value and maximum value difference < 0.5% determine that classification quantity, or classification quantity press current industry
Coking coal service index range is taken as less than 4 classes.
Following embodiment is implemented under the premise of the technical scheme of the present invention, gives detailed embodiment and tool
The operating process of body, but protection scope of the present invention is not limited to following embodiments.Method therefor is such as without spy in following embodiments
Not mentionleting alone bright is conventional method.
[embodiment]
The present embodiment by taking the coking coal of certain coke-oven plant of iron company and coke characteristic and pricing information as an example, to coking coal into
Row multidimensional index similitude classification, basic information are shown in Table 1.
1 coking coal coal quality of table, coke quality index and price
In this implementation case, the selection of level and classification indicators is calculated when coking coal is classified are as follows: the 1st calculatings level uses the
Classification indicators, that is, coking coal volatile matter V of one leveldafIndex;2nd calculating level is referred to using the classification of first level
Mark is the volatile matter V of coking coaldaf, second level classification indicators, that is, cohesiveness index caking index G value, thickness of colloidal matter layer Y
Value and difficult to understand Ah's dilation b value;3rd calculate level using second level classification indicators, that is, coking coal cohesiveness index G value,
B value and the classification indicators of the 4th level are the crushing strength M tested in coke quality index40, wear-resistant strength M10And reactivity
CRI, post reaction strength CSR.
It is obtained coking coal classification schemes (being shown in Table 2) using clustering method by 3 calculating levels, selected coking coal point
For 5 major class, 10 group.
2 coking coal classification results of table
Use coking coal volatile matter VdafWhen carrying out the 1st calculating level classification of coking coal, coking coal is divided into 3 classes, respectively
For the 1st class, the 2nd class and the 3rd class (as shown in Figure 2);Use coking coal volatile matter Vdaf, Y value and b value carry out the 2nd class coking coal
When the 2nd calculating level classification, it is divided into 3 classes, respectively 2.1 classes, 2.2 classes and 2.3 classes (as shown in Figure 3);Use coking coal Y
Value, b value and coke M40、M10, CRI, CSR the classification of the 3rd calculating level carried out to 2.1 classes and 2.2 class coking coals respectively, each point
For 2 classes, respectively 2.1.1 class, 2.1.2 class and 2.2.1 class, 2.2.2 class (as shown in Figure 4, Figure 5);Use coking coal G value and coke
Charcoal M40、M10, CRI, CSR value 3.1 class coking coals are divided into 3.1.1 class, 3.1.2 class and 3.1.3 class (as shown in Figure 6).Fig. 2-figure
Abscissa in 6 represents the number of coking coal, and ordinate represents the distance between all kinds of coking coals, shows simultaneously in Fig. 2-Fig. 6
The similarity degree relationship between coking coal is gone out.
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, according to the technique and scheme of the present invention and its
Inventive concept is subject to equivalent substitution or change, should be covered by the protection scope of the present invention.
Claims (6)
1. a kind of coking coal multidimensional index similitude classification method based on clustering, which is characterized in that existing
On the basis of grade of coal national standard, according to coal index and coke quality index, in conjunction with clustering mathematical algorithm
The multi-level multidimensional index similitude classification of coking coal is carried out, is realized using application as the coking coal classification of target;
Specifically comprise the following steps:
1) index that coking coal classification uses is divided into five levels;
The classification indicators T of first level1By coking rank of coal metamorphism index volatile matter Vdaf, vitrinite is averaged maximum reflectivityAnd variance S composition;
The classification indicators T of second level2By coal cohesiveness index caking index G value, thickness of colloidal matter layer Y value, difficult to understand Ah's dilation b
Value and Giseeler fluidity MF value composition;
The classification indicators T of third level3It is made of macerals, lazy ratio living;
The classification indicators T of 4th level4By Experiment Coke Oven coke quality index, including crushing strength M40, wear-resistant strength M10, it is anti-
Answering property CRI, post reaction strength CSR, coke average grain diameter and ash component index composition;
The classification indicators T of 5th level5By coking coal and the ash content A of cokedWith sulphur content St,dIndex composition, in an experiment not
The coking coal of energy coking is by obtaining coke quality index with addition of base bond coal coking;
2) the calculating analysis of coking coal similitude is carried out using clustering mathematical algorithm;
Mean value standardized method is selected in the standardization of clustering data, shown in transfer function z such as formula (1):
Transfer function z=(X-Mean)/(Standard deviation) (1)
Wherein, X indicates some coal quality or coke quality index data of some coking coal, and Mean indicates some of all coking coals
The mean value of coal quality or coke quality index data, Standard deviation indicate some coal quality or coke of all coking coals
The standard deviation of quality index data;
Clustering Measurement Method selects square Euclidean distance, as shown in formula (2);
Wherein, d (x, y) indicates the distance between two coking coals x and y, xiAnd yiRespectively indicate some coal quality of coking coal x and y
Numerical value after index or coke quality index standardization;
The method of average between connection or class, i.e., be averaged with distance between coking coals two-by-two all in two classes between clustering method selection group
Value is used as two between class distance Dpq, as shown in formula (3);
Wherein, Gp, GqTwo classifications for respectively indicating coking coal, respectively contain np、nqA coking coal;
3) classification of coking coal is combined by the classification indicators of different levels, and point 4 calculating levels are completed;First according to need
Coal index and coke quality index range when each calculating level analysis is arranged is to determine quantity of classifying, or presses current work
The indication range that coking coal used in industry uses directly sets classification quantity;Specifically:
1st calculating level uses the corresponding classification indicators of first level, and the 2nd calculates level and use point of first level
Class index and the classification indicators of second level are combined, and the 3rd calculates classification indicators, the third that level uses second level
The classification indicators of a level and the classification indicators of the 4th level are combined, and the 4th calculates level and use point of the 5th level
Class index;The coking coal similarity that 4th calculating level obtains is maximum.
2. a kind of coking coal multidimensional index similitude classification method based on clustering according to claim 1,
It is characterized in that, the existing national standard about grade of coal refers to GB/T 5751-2009 " Chinese Coal Classification ";
The coking coal covers weakly caking coal therein, sticks coal, bottle coal, gas rich coal, 1/3 coking coal, rich coal, coking coal, lean coal and poor thin in 1/2
Coal.
3. a kind of coking coal multidimensional index similitude classification method based on clustering according to claim 1,
It is characterized in that, the coking rank of coal metamorphism index volatile matter VdafBetween 10%~45%, vitrinite it is average maximum anti-
Rate is penetrated between 0.6%~2.3%.
4. a kind of coking coal multidimensional index similitude classification method based on clustering according to claim 1,
It is characterized in that, value >=18 caking index G of the cohesiveness index.
5. a kind of coking coal multidimensional index similitude classification method based on clustering according to claim 1,
It is characterized in that, in the Experiment Coke Oven coke quality index, crushing strength M40 < 90.0%, wear-resistant strength M10 < 35%,
Coke reactivity CRI > 15%, post reaction strength CSR < 78%.
6. a kind of coking coal multidimensional index similitude classification method based on clustering according to claim 1,
It is characterized in that, specific classification method is as follows when point 4 calculating levels complete the classification of coking coal:
1st calculate level using the classification indicators of first level and combine clustering method to selected whole coking coal into
Row classification, according to volatile matter V in organizingdafMaximum value and maximum value difference determine that classification quantity, or classification quantity are pressed less than 10%
Industry coking coal service index range is taken as less than 5 classes at present;
2nd is calculated level and is combined using the classification indicators of first level with the classification indicators of second level, and is combined poly-
Alanysis method carries out classification to each group obtained by the 1st calculating level, according to volatile matter V in organizingdafMaximum value and maximum
Value difference value < 3%, caking index G value maximum value and maximum value difference < 10, thickness of colloidal matter layer Y value maximum value and maximum value difference
Its maximum value and when maximum value difference < 40%, difficult to understand Ah's dilation b value > 100% when < 5mm, difficult to understand Ah's dilation b value < 100%
Its maximum value and maximum value difference < 150% determine that classification quantity, or classification quantity are used by current industry coking coal
Indication range is taken as less than 5 classes;
3rd calculates level and uses the classification indicators of second level, the classification indicators of third level and the 4th level
Classification indicators combination, and clustering method is combined to carry out classification again to each group obtained by the 2nd calculating level, according to group
Interior caking index G value maximum value and maximum value difference < 5, thickness of colloidal matter layer Y value maximum value and maximum value difference < 4mm, Ao A
When dilation b value < 100% its maximum value and when maximum value difference < 30%, difficult to understand Ah's dilation b value > 100% its maximum value and
Maximum value difference < 130%, Experiment Coke Oven coke quality index M40Maximum value and maximum value difference < 5%, M when > 60%10<
Maximum value and maximum value difference < 3%, CSR maximum value and maximum value difference < 8% determine classification quantity, or classification when 15%
Quantity is taken as by current industry coking coal service index range less than 5 classes;
4th calculates level using the classification indicators of the 5th level and combines clustering method to the 3rd calculating level institute
It obtains each group and carries out classification again, according to coking coal or its coke ash A in organizingdMaximum value and maximum value difference < 1.0%,
Sulphur content AdMaximum value and maximum value difference < 0.5% determine that classification quantity, or classification quantity press current industry coking coal
Service index range is taken as less than 4 classes.
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 true CN109102035A (en) | 2018-12-28 |
CN109102035B 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) |
Cited By (4)
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 |
CN112098263A (en) * | 2020-09-14 | 2020-12-18 | 山西亚鑫新能科技有限公司 | Method for parameter comprehensive prediction of coke thermal strength model |
CN114369471A (en) * | 2021-12-27 | 2022-04-19 | 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 |
Citations (8)
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 |
-
2018
- 2018-09-11 CN CN201811055912.3A patent/CN109102035B/en active Active
Patent Citations (8)
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)
Title |
---|
CHEN HONG-JUN等: "Weight Determination Method Based on Principal Component Analysis Coking", 《ADVANCES IN MANUFACTURING SCIENCE AND ENGINEERING》 * |
H.B. SAHU等: "An empirical approach for classification of coal seams with respect to the spontaneous heating susceptibility of Indian coals", 《INTERNATIONAL JOURNAL OF COAL GEOLOGY》 * |
单晓云等: "模糊聚类分析优化炼焦配煤的研究", 《煤炭科学技术》 * |
吕青: "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", 《中国优秀博硕士学位论文全文数据库(博士) 信息科技Ⅰ辑》 * |
高志芳: "开滦精煤炼焦特性及焦炭质量预测的研究", 《中国优秀博硕士学位论文全文数据库 (硕士) 工程科技Ⅰ辑》 * |
Cited By (4)
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 |
CN112098263A (en) * | 2020-09-14 | 2020-12-18 | 山西亚鑫新能科技有限公司 | Method for parameter comprehensive prediction of coke thermal strength model |
CN114369471A (en) * | 2021-12-27 | 2022-04-19 | 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 |
Also Published As
Publication number | Publication date |
---|---|
CN109102035B (en) | 2022-03-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109102035A (en) | A kind of coking coal multidimensional index similitude classification method based on clustering | |
CN105062531B (en) | Coking raw material application is classified and Quality evaluation and its instructs blending method | |
CN102890144B (en) | Method for predicting coke quality through nonlinear optimization coal blending based on coal rock vitrinite total reflectance | |
CN102890145B (en) | Method for performing nonlinear prediction on coke quality on basis of cohesiveness and coal-rock indexes of single coal | |
Gray et al. | Coke carbon forms: microscopic classification and industrial applications | |
CN101081989A (en) | Coal coking blending method | |
CN104140834B (en) | Based on the coking coal divided method of coking property and the application in coal blending | |
CN103952166B (en) | Based on ature of coal classification and the blending method of coking coal coking property | |
CN102690669B (en) | Coking coal quality classification method and coal blending and coking method with participation of coking coal | |
CN109064061A (en) | A kind of coking coal multidimensional property evaluation method based on AHP analytic hierarchy process (AHP) | |
CN102901802A (en) | Method for evaluating cost performance of coking coal | |
CN105243437B (en) | The method of prediction coke quality and Optimized Coal Blending ratio for tamping coking | |
CN101294948B (en) | Method for anthracology coal blending | |
CN103275740A (en) | Evaluation method of fat coal quality | |
CN104449778B (en) | A kind of method that integrated use coal petrography index carries out the exploitation of coal source | |
CN105316017B (en) | A kind of Blending of Coal Petrography method using coal for coking vitrinite reflectance as leading indicator | |
Chen et al. | Petrographic characteristics of Chinese coals and their application in coal utilization processes | |
CN112521965A (en) | Rapid coal blending method | |
CN106479549B (en) | Mixed coal Giseeler fluidity prediction technique | |
CN114662763A (en) | Method and system for evaluating cost performance of single coal for coking coal blending | |
CN110210000A (en) | The identification of industrial process efficiency and diagnostic method based on Multiple Non Linear Regression | |
CN104678075B (en) | The Forecasting Methodology of coal-blending coking coke scuff resistance | |
Liu et al. | Coal blend properties and evaluation on the quality of stamp charging coke from weakly coking blends | |
CN102095667A (en) | Coal quality assessment method of coking coal having volatile component between 27% and 29% | |
CN102807883A (en) | Gas-fat-doped coking and coal blending 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 |