CN112926270A - Gas multi-factor coupling relation analysis and early warning model construction method - Google Patents

Gas multi-factor coupling relation analysis and early warning model construction method Download PDF

Info

Publication number
CN112926270A
CN112926270A CN202110290687.7A CN202110290687A CN112926270A CN 112926270 A CN112926270 A CN 112926270A CN 202110290687 A CN202110290687 A CN 202110290687A CN 112926270 A CN112926270 A CN 112926270A
Authority
CN
China
Prior art keywords
data
gas
early warning
clustering
support
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
Application number
CN202110290687.7A
Other languages
Chinese (zh)
Other versions
CN112926270B (en
Inventor
黄玉鑫
范京道
闫振国
王延平
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.)
Xian University of Science and Technology
Original Assignee
Xian University of Science and Technology
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 Xian University of Science and Technology filed Critical Xian University of Science and Technology
Priority to CN202110290687.7A priority Critical patent/CN112926270B/en
Publication of CN112926270A publication Critical patent/CN112926270A/en
Application granted granted Critical
Publication of CN112926270B publication Critical patent/CN112926270B/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
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/08—Probabilistic or stochastic CAD
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Probability & Statistics with Applications (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a method for analyzing a gas multi-factor coupling relationship and constructing an early warning model, which mainly comprises two parts, wherein one part is used for selecting and detecting an abnormal value of an optimized k-mean algorithm based on an initial value of a cluster center, and the other part is used for mining an association rule of an Apriori algorithm based on item set support degree weight optimization. The main idea is as follows: the method comprises the steps of collecting working face, upper corner, coal seam gas concentration data and working face hydraulic support pressure data, conducting abnormal value detection on four dimensional data through a k-means algorithm selected and optimized based on a cluster center initial value, extracting abnormal values to serve as original data mined by association rules, conducting dimensionality reduction on the original data, converting the original data into 0-1 Boolean matrixes, conducting association rule mining on the matrixes through an Apriori algorithm optimized based on a support degree weight, finally determining coupling relations among 4 gas influence factors, exploring the association rules, determining gas early warning levels, and effectively improving gas data abnormal value analysis capacity and gas disaster risk early warning performance.

Description

Gas multi-factor coupling relation analysis and early warning model construction method
Technical Field
The invention relates to the technical field of coal mine safety, in particular to a method for analyzing a gas multi-factor coupling relationship and constructing an early warning model.
Background
In the process of coal mine informatization and intelligentization construction, "mass data and lack of information" are ubiquitous problems. The method has the advantages that the mine monitoring data are analyzed and researched, the production is guided according to the analysis result, and the method has important significance for realizing high-quality development of the coal industry and related industries in a new period of coal mine enterprises. The gas monitoring data is important data related to coal mine production safety, a large amount of abnormal data often exists in the gas monitoring data, and although the coal mine underground gas monitoring system is gradually improved, the analysis and the processing of the data acquired by the system are insufficient. Most of the gas data are subjected to prediction analysis and rule extraction based on time series, and mining analysis of association rules among dimensions of different influence factors is lacked for gas abnormal value data, so that the requirement of coal mine safety production is difficult to meet.
The gas monitoring data is important data about coal mine production safety, but a large amount of abnormal data often exist in the gas monitoring data, and the abnormal data may be caused by noise and may also be a precursor of danger, but the current mine monitoring system cannot effectively distinguish the abnormal data, only carries out disaster early warning according to a preset gas concentration threshold index, and has certain risk. The underground gas abnormity causes are various, and the danger degrees caused by gas abnormity are different, so that a grading early warning mechanism aiming at the gas danger under different danger degrees needs to be constructed urgently, the existence of the gas danger can be early warned in advance, and a plan can be set in advance to deal with the gas danger of different levels.
Most of the prior art is to analyze and extract the monitoring data of a certain monitoring point fixed underground, and establish an early warning model for gas outburst and gas overrun, and certain research results still have some defects:
(1) at present, abnormal value data uploaded by a gas sensor is abandoned in the data processing process under most conditions of regular analysis and prediction of gas monitoring data, the partial abnormal values are not fully utilized and mined, potential characteristics and rules of the partial abnormal values are found, and secondary utilization of the partial data is lacked.
(2) In the aspect of gas early warning, the existing technology only considers factors such as gas outburst and emission quantity of a working face, gas risks are considered and evaluated in single-dimensional data, and influences of other factors on the gas concentration of the working face are not considered, for example, the gas concentration of an upper corner, the gas concentration of a coal seam and the pressure fluctuation of a bottom plate caused by the formation of a mining area are not considered, and the geological structure of the coal seam can change along with the pressure fluctuation, so that the pressure born by a hydraulic support of the working face fluctuates, the release rate of the gas of the coal seam is influenced, and the gas concentration of the working face is further influenced.
Disclosure of Invention
In the process of researching the gas characteristics of the working face in the past, the deep analysis of abnormal values in monitored gas data and the establishment of a grading early warning mechanism of gas dangerous cases through analysis are less based on gas concentration prediction, gas occurrence rules and the like. The invention aims to construct a multi-factor coupling relation analysis and grading early warning model of gas on a working face, construct an association rule learning set by extracting and analyzing abnormal values in a gas monitoring process, design a grading early warning mechanism and solve the technical problem of grading early warning of dangerous gas on the working face.
In order to achieve the purpose, the invention provides the following technical scheme: the method for analyzing the gas multi-factor coupling relationship and constructing the early warning model specifically comprises the following steps:
collecting working face, upper corner, coal bed gas concentration data and working face hydraulic support pressure data;
step two, abnormal value detection of a k-means algorithm selected and optimized based on a new cluster initial value is carried out on the four dimensional data;
extracting abnormal values as original data mined by association rules;
step four, performing dimensionality reduction on the original data, and converting the original data into a 0-1 Boolean matrix;
fifthly, mining association rules of the matrix by using an apriori algorithm based on support degree weight optimization;
and step six, finally determining the coupling relation among the four gas influence factors, exploring the association rule of the four gas influence factors, and determining the early warning grade of the gas.
Preferably, the Apriori algorithm searches the (k +1) term set by using the k term set by using a layer-by-layer search iteration method to find out all frequent term sets, and the frequent set generates a strong association rule.
Preferably, let Φ ═ I1，I2，…，ImIs a set of items, and task-related data D is a set of database transactions, where each transaction T is a non-empty set of items, such that
Figure BDA0002982463420000031
Each transaction has an identifier, called TID;
let A be a set of items, transaction T contain A, if and only if
Figure BDA0002982463420000032
The association rule is in the form of
Figure BDA0002982463420000033
Is of the connotation type wherein
Figure BDA0002982463420000034
And is
Figure BDA0002982463420000035
Rules
Figure BDA0002982463420000036
Is true in the transaction set, with a support degree s, where s is the percentage of transactions in D that contain AuB, as the probability P (AuB);
rules
Figure BDA0002982463420000037
There is a confidence level c in the transaction set D, where c is the percentage of transactions in D that contain a, while also containing B. This is the conditional probability P (B | a). Namely:
Figure BDA0002982463420000038
Figure DA00029824634237196480
Figure BDA0002982463420000042
where support _ count (atou B) is the number of transactions containing the set of items atou B, and support _ count (a) is the number of transactions containing the set of items a. The rule that satisfies both the minimum support threshold (min _ sup) and the minimum confidence threshold (min _ conf) is a strong rule.
Preferably, in the process of mining the frequent item sets, the traditional Apriori algorithm deletes the item set lower than the minimum support threshold standard in the pruning process, and if partial data which has a large influence on the working face gas dangerous case but has a small detected gas abnormal value and cannot reach the set support threshold size exist, the algorithm deletes the partial data, so that the dangerous case rule of the partial data is lost. In view of this, the algorithm support is optimized, and a weight parameter is added in the support calculation process, so that the gas abnormal value with low occurrence frequency is added into the frequent item set. The design support threshold function is:
Figure BDA0002982463420000043
wherein support (X) is the support of a given set of items, λXD is the transaction set. For any x ═ x1,x2,…,xrIn which xiE I (I ═ 1,2, …, r), if x is a single term, its weighting coefficient is assigned after the term set is generated, otherwise, the weighting parameter needs to be obtained from the contained terms, i.e.:
λx＝F(λx1,λx2,…,λxr)＝min(λx1,λx2,…,λxr) (4)
the calculation function of the weight parameter is expressed as the minimum value of the weight parameters.
Preferably, the outliers in the gas data can be regarded as outliers in the clustering process, so that the outliers can be extracted by detecting the outliers by considering the relationship between the object and the clusters based on the clustering method, and each object o can be given an outlier score according to the distance between the object and the nearest cluster center by using k-means clustering.
Preferably, assume that the nearest center to o is coThen o and coThe distance between is dist (o, c)o)，coAnd is assigned to coHas an average distance between the objects of
Figure BDA0002982463420000051
Ratio of
Figure BDA0002982463420000052
The metric dist (o, c)o) The degree of difference from the mean, points that are far from the corresponding center are suspected to be outliers;
the k-means algorithm aims to convert n m-dimensional data X to X1,x2,...,xn}，xi∈Rm(1 ≦ i ≦ n), clustering to k sets.
Preferably, the algorithm clusters data by randomly extracting an initial cluster center, which may result in different initial values corresponding to different clustering results, so that the method for selecting the initial cluster center needs to be optimized. Because the production of the coal mining working face has periodicity, namely the processes of feeding, cutting coal, loading coal, transporting coal, moving a frame, performing three-shift cycle operation every day, and performing the processes of overhauling, drilling, supporting and the like in the period, and the production processes every day are relatively fixed, the cluster centers of the pre-order data and the post-order data in the collected data set have similarity, the cluster centers of the pre-order data subset are used for optimizing the initial cluster centers of all the data, and the algorithm is described as follows:
setting a gas monitoring data set as n m-dimensional data X ═ X1,x2,...,xn}，xi∈Rm(i is more than or equal to 1 and less than or equal to n), clustering to k sets, extracting preamble data subset sets of the gas data set X,
Figure BDA0002982463420000053
determining a preamble data set X using contour coefficientproFor the preamble data set XproRandomly determining k cluster centers cpro＝{cpro1,cpro2,...,cprokThen, according to the data contained in each cluster set, adopting a formula
Figure BDA0002982463420000054
Calculating the distance between the data to be clustered and the cluster center by adopting a formula
Figure BDA0002982463420000061
Updating the iterative cluster center value as a criterion function until a final clustering result is determined to obtain a final clustering center set c of the preamble data subsetpro＝{cpro1,cpro2,...,cprok}；
Set k initial cluster centers c ═ c of dataset XproAnd determining the final clustering center by applying a k-means algorithm. And giving a threshold value theta, and if the distance between a certain data value and the clustering center is greater than theta, judging that the data value is an abnormal value.
Compared with the prior art, the invention has the beneficial effects that:
the invention breaks through the traditional gas early warning method, analyzes the gas monitoring data by adopting an abnormal value detection and data mining method, constructs a grading early warning mechanism aiming at the gas dangerous case under different dangerous degrees, can not only early warn the gas dangerous case in advance and grade the gas dangerous case according to the dangerous degree, but also can set a plan in advance to deal with the gas dangerous cases of different grades.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a three-dimensional distribution diagram of gas concentration data;
FIG. 2 is a pressure data scatter plot;
FIG. 3 is a thermal correlation coefficient plot of preceding and following gas concentration data;
FIG. 4a is a line drawing of the contour coefficient under different values of k in the contour coefficient method;
FIG. 4b is a schematic diagram of clustering under different k values and contour coefficients in the contour coefficient method;
FIG. 5 is a graph of the clustering effect of the data sets of the first two days;
FIG. 6 is a graph of data outlier identifiers for the first two days;
FIG. 7 is a graph value detection of a seven day data set clustering effect;
FIG. 8 is a seven day data outlier identification plot;
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
Referring to fig. 1 to 8, the present invention provides the following technical solutions:
example one
And (3) verifying the model by taking a certain working face of a certain mine as a test working face, and arranging a gas concentration sensor and a pressure sensor on a hydraulic support of the working face respectively for the drainage pipeline, the stope working face and the upper corner of the coal seam.
Gas data of a working face, a coal bed and an upper corner are collected at intervals of 5 seconds, the maximum value of the gas in 5 minutes is taken as an experimental data value, the pressure data is subjected to value taking once every 5 minutes and corresponds to a time node of a gas data extraction value, and the data of 7 days are arranged and respectively recorded as: u shapeX、WX、CXAnd SXEach of the 4 dimensions collects a data 2016 set. Therefore, the concentration of the gas extracted from the upper corner, the working face and the coal bed is obtained, the three-dimensional data distribution diagram is shown in figure 1, the pressure data distribution diagram is shown in figure 2, and the abnormal points are partially detected by an abnormal value detection algorithm. The method is characterized in that gas dangerous case alarm information appearing in a gas monitoring and monitoring system is recorded while gas is collected, and 55 gas dangerous cases are recorded.
Setting the preceding data sets of the upper corner, the working face and the coal bed gas concentration data of the previous 2 days as
Figure BDA0002982463420000071
And
Figure BDA0002982463420000072
576 groups, respectively, for the last 5 days
Figure BDA0002982463420000073
And
Figure BDA0002982463420000081
1440 groups each. The maximum, minimum and mean characteristics of the preceding and following gas data sets are shown in table 1 below,
Figure BDA0002982463420000082
TABLE 1 initial Cluster center comparison
The thermal correlation analysis is performed on the front and rear sequence data to obtain a thermal correlation coefficient graph as shown in fig. 3, and it can be seen that the absolute values of the correlation coefficients of the front and rear sequence data of the upper corner, the working face and the coal seam are all greater than 0.1, and the correlation is strong.
Dividing the data to be classified into k clusters by adopting a k-means clustering algorithm, respectively calculating the distance a (i) from each vector in a cluster to other points in all the clusters to which the vector belongs and the average distance b (i) from all the points in a cluster which is adjacent and nearest to the vector by adopting a contour coefficient method, and then calculating the round of i vectorsThe contour coefficient is:
Figure BDA0002982463420000083
the value of the visible contour coefficient is between-1, 1]The closer to 1, the better the cohesion and separation. As shown in fig. 4a and 4b, the maximum value of the calculated contour coefficient is s (i) 0.89, and the optimum clustering effect is obtained by the contour coefficient method.
From the above, the clustering of the preamble data sets is performed, and the clustering effect is shown in fig. 6. Abnormal values and discrete points of the clustered scattered points are detected, 29 groups of abnormal data are detected when the threshold value is 2, and the detection effect is shown in fig. 5.
The clustering centers of the pre-order data sets are adopted to optimize the clustering centers of the whole data sets, abnormal value detection is carried out on the whole data sets, a three-dimensional graph of the clustering effect of the whole data sets is shown as the following graph 7, an abnormal value marking graph of the whole data sets is shown as the following graph 8, and the ratio of the cluster center change conditions of the optimized cluster centers to the random initial cluster centers is shown as a table 2.
The initial values of the support pressure data in the data acquisition process are all 24Mpa, and the support pressure changes along with the mining influence when the coal bed stress structure changes. And detecting the support pressure data by adopting the abnormal value detection method. This time, 253 sets of abnormal data were detected by abnormal value detection, which included: and (D)55 groups of dangerous cases (D) are recorded together with 56 groups of working face gas concentration abnormal values, 58 groups of upper corner gas concentration abnormal values, 56 groups of coal seam gas concentration abnormal values and 83 groups of support pressure data abnormal values. And converting the detected abnormal data values and the dangerous case records into 0-1 Boolean matrixes, recording the abnormal data as 1, recording the normal data as 0, recording the dangerous case as 1 and recording the no dangerous case as 0, wherein 125 groups of abnormal data combinations are obtained. The partially transformed 0-1 boolean matrix is shown in table 2, and is subjected to association rule mining based on the optimized Apriori algorithm constructed as described above.
Figure BDA0002982463420000091
Figure BDA0002982463420000101
TABLE 2 partial anomaly data set 0-1 Boolean matrix table
And analyzing the coupling relation among the abnormal values detected by the working face, the upper corner, the gas concentration of the coal bed and the support pressure by using an Apriori algorithm. Setting a minimum support threshold value to be 20%, setting a minimum confidence threshold value to be 60%, obtaining association rule results as shown in table 3, obtaining 16 association rules in total, wherein Con in the table represents preconditions of the association rules, 1,2 and 3 preconditions respectively exist in the association rules generated by mining at this time, Re represents results caused by the preconditions in the association rules, Su represents support of the association rules, and Cof represents confidence of the association rules.
Figure BDA0002982463420000102
TABLE 3 working face, upper corner, coal seam and support pressure data coupling relation rule table
The table and the figures show that the working surface, the upper corner, the gas concentration of the coal bed and the support pressure have stronger correlation. The data of any two dimensions are abnormal, which can lead to the abnormality of other dimension data except the coal bed, the gas concentration of the coal bed is mainly influenced by mine pressure and natural gas occurrence factors in the mining process, and therefore 2 association rules which lead to the abnormality of the coal bed in the association rules are the precondition that the pressure dimension data are abnormal.
Similarly, the method is adopted to explore the association relationship between four dimensions of the working face, the upper corner, the coal seam and the support pressure and the underground gas dangerous case, the minimum support degree threshold value is set to be 20%, the minimum confidence degree threshold value is set to be 60%, and 14 groups of association rules are obtained, and the following table 4 is shown.
Figure BDA0002982463420000111
TABLE 4 partial association rules of underground gas dangerous case and coal mine abnormal data training set
According to the results, when the data with more than three dimensions in the gas concentration data and the pressure data of the upper corner, the coal bed and the working face are detected to be abnormal, the gas dangerous case is caused by the confidence coefficient of 100%. In the association rules of the results caused by the two factors, the confidence coefficient of the dangerous case is 100% only under the precondition that the pressure of the coal bed, the working face and the working face is abnormal, and the confidence coefficient of the association rules of the results caused by the other two factors is more than 95%; the confidence of the association rule of the dangerous case caused by a single factor is lower than that of the former two, namely, about 75%. Therefore, it can be seen that multiple abnormal factors in the well are more likely to cause the gas dangerous case than a single abnormal factor.
Strong, weak and weak association rules are obtained by setting different confidence degrees, and the rules are sequentially divided into four levels, namely a level I, a level II, a level III and a level IV from strong to weak according to the confidence degrees from high to low, which is specifically shown in the following table 5.
Figure BDA0002982463420000112
Figure BDA0002982463420000121
TABLE 5 underground gas dangerous situation early warning grade table
The above table shows that the early warning confidence of the I-level (Sup is more than or equal to 0.2, and Cof is more than or equal to 0.99) dangerous case is 100%, and most of the early warning confidence is caused by the simultaneous abnormality of 3 dimensional data, when the abnormal values are detected by the pressure data of the coal bed, the upper corner, the working face and the support, the structure of the coal bed is possibly damaged in the mining process, the conditions such as hydrology are also changed, and gas gushes out from multiple places; the II-level (Sup is more than or equal to 0.2, and Cof is more than or equal to 0.99 and more than or equal to 0.7) dangerous case early warning is mainly caused by 2 dimensional data anomalies, and the occurrence frequency of the gas concentration anomalies of the coal bed and the upper corner can be seen to be higher from the precondition of the association rule; the dangerous case causes of class III (Su is more than or equal to 0.2, 0.7 is more than or equal to 0.5) and class IV (Su is more than or equal to 0.2, 0.5 is more than or equal to 0.4) are single, and the dangerous case of gas is caused only by data abnormality of one dimension.
The preferred embodiments of the invention disclosed above are intended to be illustrative only. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise embodiments disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims (7)

1. The method for analyzing the gas multi-factor coupling relationship and constructing the early warning model is characterized by comprising the following steps of:
collecting working face, upper corner, coal bed gas concentration data and working face hydraulic support pressure data;
step two, abnormal value detection of a k-means algorithm selected and optimized based on a new cluster initial value is carried out on the four dimensional data;
extracting abnormal values as original data mined by association rules;
step four, performing dimensionality reduction on the original data, and converting the original data into a 0-1 Boolean matrix;
fifthly, mining association rules of the matrix by using an Apriori algorithm based on support degree weight optimization;
step six, finally determining fourDifferent dimensionsAnd (3) exploring the association rule of the coupling relation among the gas influence factors, and determining the early warning grade of the gas.
2. The method for analyzing and constructing the early warning model according to the gas multifactor coupling relationship of claim 1, wherein the Apriori algorithm searches the (k +1) term set by using the k term set by using a layer-by-layer search iteration method to find out all frequent term sets, and the frequent set generates a strong association rule.
3. The method for analyzing and pre-warning a gas multi-factor coupling relationship according to claim 2, wherein the gas multi-factor coupling relationship is a gas multi-factor coupling relationship,
let phi ═ I1，I2，…，ImIs a set of items, and task-related data D is a set of database transactions, where each transaction T is a non-empty set of items, such that
Figure FDA0002982463410000011
Each transaction has an identifier, called TID;
let A be a set of items, transaction T contain A, if and only if
Figure FDA0002982463410000012
The association rule is in the form of
Figure FDA0002982463410000021
Is of the connotation type wherein
Figure FDA0002982463410000022
And is
Figure FDA0002982463410000023
Rules
Figure FDA0002982463410000024
Is true in the transaction set, with a support degree s, where s is the percentage of transactions in D that contain AuB, as the probability P (AuB);
rules
Figure FDA0002982463410000025
With confidence c in transaction set D, where c is the hundred of transactions in D that contain both A and BAnd (4) dividing the ratio. This is the conditional probability P (B | a). Namely:
Figure FDA0002982463410000026
Figure FDA0002982463410000027
where support _ count (atou B) is the number of transactions containing the set of items atou B, and support _ count (a) is the number of transactions containing the set of items a. The rule that satisfies both the minimum support threshold (min _ sup) and the minimum confidence threshold (min _ conf) is a strong rule.
4. The method for analyzing the gas multi-factor coupling relationship and constructing the early warning model according to claim 3, wherein the Apriori algorithm support is optimized, and a weight parameter is added in the support calculation process, so that a gas abnormal value with low occurrence frequency can be added into a frequent item set. The design support threshold function is:
Figure FDA0002982463410000028
wherein support (X) is the support of a given set of items, λXD is the transaction set. For any x ═ x1,x2,…,xrIn which xiE I (I ═ 1,2, …, r), if x is a single term, its weighting coefficient is assigned after the term set is generated, otherwise, the weighting parameter needs to be obtained from the contained terms, i.e.:
λx＝F(λx1,λx2,…,λxr)＝min(λx1,λx2,…,λxr) (4)
the calculation function of the weight parameter is expressed as the minimum value of the weight parameters.
5. The method as claimed in claim 1, wherein the outliers in the gas data are considered to be outliers in the clustering process, so that the outliers can be extracted by detecting the outliers by examining the relationship between the object and the clusters based on the clustering method, and k-means clustering is used, and for each object o, an outlier score can be given to the object according to the distance between the object and the nearest cluster center.
6. The method for analyzing and pre-warning the gas multifactor coupling relationship according to claim 5, wherein c is the nearest center to ooThen o and coThe distance between is dist (o, c)o)，coAnd is assigned to coHas an average distance between the objects of
Figure FDA0002982463410000031
Ratio of
Figure FDA0002982463410000032
The metric dist (o, c)o) The degree of difference from the mean, points that are far from the corresponding center are suspected to be outliers;
the k-means algorithm aims to convert n m-dimensional data X to X1,x2,...,xn}，xi∈Rm(1 ≦ i ≦ n), clustering to k sets.
7. The method for analyzing the gas multi-factor coupling relationship and constructing the early warning model according to claim 6, wherein the initial clustering centers of all data are optimized by using the clustering centers of the preamble data subsets, and the algorithm is described as follows:
setting a gas monitoring data set as n m-dimensional data X ═ X1,x2,...,xn}，xi∈Rm(i is more than or equal to 1 and less than or equal to n), clustering to k sets, extracting preamble data subset sets of the gas data set X,
Figure FDA0002982463410000033
determining a preamble data set X using contour coefficientproFor the preamble data set XproRandomly determining k cluster centers cpro＝{cpro1,cpro2,...,cprokThen, according to the data contained in each cluster set, adopting a formula
Figure FDA0002982463410000041
Calculating the distance between the data to be clustered and the cluster center by adopting a formula
Figure FDA0002982463410000042
Updating the iterative cluster center value as a criterion function until a final clustering result is determined to obtain a final clustering center set c of the preamble data subsetpro＝{cpro1,cpro2,...,cprok}；
Set k initial cluster centers c ═ c of dataset XproAnd determining the final clustering center by applying a k-means algorithm. And giving a threshold value theta, and if the distance between a certain data value and the clustering center is greater than theta, judging that the data value is an abnormal value.
CN202110290687.7A 2021-03-18 2021-03-18 Gas multi-factor coupling relation analysis and early warning model construction method Active CN112926270B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110290687.7A CN112926270B (en) 2021-03-18 2021-03-18 Gas multi-factor coupling relation analysis and early warning model construction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110290687.7A CN112926270B (en) 2021-03-18 2021-03-18 Gas multi-factor coupling relation analysis and early warning model construction method

Publications (2)

Publication Number Publication Date
CN112926270A true CN112926270A (en) 2021-06-08
CN112926270B CN112926270B (en) 2023-05-05

Family

ID=76175054

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110290687.7A Active CN112926270B (en) 2021-03-18 2021-03-18 Gas multi-factor coupling relation analysis and early warning model construction method

Country Status (1)

Country Link
CN (1) CN112926270B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120254242A1 (en) * 2011-03-31 2012-10-04 Infosys Technologies Limited Methods and systems for mining association rules
CN103437808A (en) * 2013-08-30 2013-12-11 刘正魁 Method for preparing coal mine gas instant eliminating liquid and method for using coal mine gas instant eliminating liquid
US20140019047A1 (en) * 2011-03-30 2014-01-16 China University Of Mining And Technology Multi-information coupling prediction method of coal and gas outburst danger
CN109752504A (en) * 2019-01-25 2019-05-14 西安科技大学 A kind of face gas sensor adjustment aid decision-making method
WO2020006818A1 (en) * 2018-07-02 2020-01-09 山东科技大学 Method for area-based coal rock water injection seepage-damage-stress coupling value simulation
CN111582603A (en) * 2020-05-19 2020-08-25 中煤科工集团重庆研究院有限公司 Intelligent early warning method for coal and gas outburst based on multi-source information fusion
US20200311099A1 (en) * 2017-04-18 2020-10-01 Chinese Academy Of Inspection And Quarantine Method of online tracing pesticide residues and visualizing warning on basis of high resolution mass spectrum, internet and geographic information

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140019047A1 (en) * 2011-03-30 2014-01-16 China University Of Mining And Technology Multi-information coupling prediction method of coal and gas outburst danger
US20120254242A1 (en) * 2011-03-31 2012-10-04 Infosys Technologies Limited Methods and systems for mining association rules
CN103437808A (en) * 2013-08-30 2013-12-11 刘正魁 Method for preparing coal mine gas instant eliminating liquid and method for using coal mine gas instant eliminating liquid
US20200311099A1 (en) * 2017-04-18 2020-10-01 Chinese Academy Of Inspection And Quarantine Method of online tracing pesticide residues and visualizing warning on basis of high resolution mass spectrum, internet and geographic information
WO2020006818A1 (en) * 2018-07-02 2020-01-09 山东科技大学 Method for area-based coal rock water injection seepage-damage-stress coupling value simulation
CN109752504A (en) * 2019-01-25 2019-05-14 西安科技大学 A kind of face gas sensor adjustment aid decision-making method
CN111582603A (en) * 2020-05-19 2020-08-25 中煤科工集团重庆研究院有限公司 Intelligent early warning method for coal and gas outburst based on multi-source information fusion

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
ZHAO XUSHENG 等: "Research on comprehensive early-warning technology of Coal and gas outburst", 《ELSEVIER》 *
刘斌: "基于最小二乘法的上隅角瓦斯浓度变化规律分析", 《洁净煤技术》 *
钱宇虹: "数据挖掘算法在瓦斯安全预测中的应用", 《煤炭技术》 *

Also Published As

Publication number Publication date
CN112926270B (en) 2023-05-05

Similar Documents

Publication Publication Date Title
CN110213222B (en) Network intrusion detection method based on machine learning
WO2019237492A1 (en) Semi-supervised learning-based abnormal electricity utilization user detection method
CN112288014B (en) Data mining-based equipment full life cycle management method
CN107578149B (en) Power grid enterprise key data analysis method
CN113298165A (en) Coal mine gas disaster advanced prediction method based on multi-source information fusion
Kotanchek et al. Symbolic regression via genetic programming as a discovery engine: Insights on outliers and prototypes
Zhou et al. A novel stacking heterogeneous ensemble model with hybrid wrapper-based feature selection for reservoir productivity predictions
Gan et al. A new spatial modeling method for 3D formation drillability field using fuzzy c-means clustering and random forest
Zixian et al. Outburst prediction and influencing factors analysis based on Boruta-Apriori and BO-SVM algorithms
CN114328663A (en) High-dimensional theater data dimension reduction visualization processing method based on data mining
CN115860582B (en) Impact risk intelligent early warning method based on self-adaptive lifting algorithm
CN117312919A (en) Rock burst early warning method based on BO-CNN-LSTM model
CN112926270A (en) Gas multi-factor coupling relation analysis and early warning model construction method
CN115272798B (en) Method and system for predicting training positive sample by using information quantity to expand ore formation
CN109857783B (en) Sonar seepage motion law discovery method based on deep learning
Yin et al. Time series based data explorer and stream analysis for anomaly prediction
Yuan et al. Application of Active Learning in Carbonate Lithologic Identification
CN115438239A (en) Abnormity detection method and device for automatic abnormal sample screening
CN114722081A (en) Streaming data time sequence transmission method and system based on transfer library mode
CN114064723A (en) Association rule mining method and device, computer equipment and storage medium
Borah et al. Rare association rule mining: a systematic review
Gao et al. Anomaly detection for time series with difference rate sample entropy and generative adversarial networks
Zhang et al. A new outlier detection algorithm based on fast density peak clustering outlier factor.
Chen et al. Identification of Wellbore Flow Abnormal Working Conditions Based on Deep Learning
CN113486088A (en) Data mining method based on complex technology

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