WO2019091134A1 - Method for predicting water-rich levels of sandstone in coal seam roof - Google Patents
Method for predicting water-rich levels of sandstone in coal seam roof Download PDFInfo
- Publication number
- WO2019091134A1 WO2019091134A1 PCT/CN2018/095623 CN2018095623W WO2019091134A1 WO 2019091134 A1 WO2019091134 A1 WO 2019091134A1 CN 2018095623 W CN2018095623 W CN 2018095623W WO 2019091134 A1 WO2019091134 A1 WO 2019091134A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- water
- rich
- inf
- coal seam
- sandstone
- Prior art date
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 110
- 239000003245 coal Substances 0.000 title claims abstract description 79
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000005070 sampling Methods 0.000 claims abstract description 49
- 238000012549 training Methods 0.000 claims abstract description 21
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 238000012360 testing method Methods 0.000 claims abstract description 7
- 230000035945 sensitivity Effects 0.000 claims description 16
- 238000010276 construction Methods 0.000 claims description 6
- 239000002002 slurry Substances 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 3
- 239000002689 soil Substances 0.000 claims 1
- 238000013528 artificial neural network Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000007792 addition Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
- G01V1/30—Analysis
- G01V1/306—Analysis for determining physical properties of the subsurface, e.g. impedance, porosity or attenuation profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V2210/00—Details of seismic processing or analysis
- G01V2210/60—Analysis
- G01V2210/62—Physical property of subsurface
- G01V2210/624—Reservoir parameters
Definitions
- the invention relates to the technical field of underground coal mining safety production, in particular to a method for predicting the water-rich grade of coal seam roof sandstone.
- the water damage in the coal seam roof is one of the mine water hazards, which seriously threatens the safe recovery of coal.
- the gushing of the roof water of the mine not only worsens the production environment of the working face, but also brings unsafe factors to the mine production. Because the water content of the roof sandstone of the mine could not be ascertained, the prediction of the water discharge from the roof is improper, or the safety factor is excessively increased, which greatly increases the cost of the prevention and control project; or the safety factor is too small, and the mine (mining area) has the ability to prevent and discharge water. Insufficient, causing major accidents in flooded areas and even casualties.
- the single factor indicator method is based on the establishment of a single influencing factor for water-rich assessment, although the method is simple, easy to operate and fast.
- the method of establishing the water-rich assessment based on the individual influencing factors has great limitations. Sex and one-sidedness.
- Field measurement is the most accurate and reliable method. However, due to its large amount of engineering, time-consuming and labor-intensive, high labor costs, and the constraints of the underground geography environment and complex roadway complex, many times, large-scale or all-round on-site measurement is impossible.
- the neural network method is based on the analysis of the factors affecting the water-richness of the roof sandstone in the coal seam and the field measured data.
- the neural network prediction model for the water-rich roof of the coal seam roof sandstone is constructed.
- the neural network prediction model is easy to fall into the local optimal solution when the weight and threshold are updated, the prediction accuracy is relatively low, and the promotion ability is poor.
- the object of the present invention is to provide a method for predicting the water-rich grade of coal seam roof sandstone, and to solve the technical problem that the prediction of the water-rich grade of the coal seam roof sandstone is inaccurate and the prediction efficiency is low.
- the invention provides a method for predicting the water-rich grade of coal seam roof sandstone, comprising the following steps: Step 1: sampling the N sampling sites of the coal seam roof sandstone in the underground, and analyzing the coal seam roof sandstone at each sampling point to obtain the three rich
- the water-based grade and the six influencing factors affecting the water-rich grade, the three water-rich grades include weak, medium, and strong, and the six influencing factors include sandstone thickness C 1 , mudstone thickness C 2 , core take-up rate C 3 , and slurry
- the leakage amount C 4 , the fracture density C 5 and the breaking strength C 6 are used to establish a sample database of the above sampling data at the N sampling points, wherein the water-rich grade of the coal seam roof sandstone at each sampling point needs to be determined after the coal mine construction;
- Step 2 After preprocessing the data of the sample database, the six influencing factors of the sampled data at each sampling point are taken as input vectors, and the three water-rich levels are used as output vectors to establish a coal seam roof sandstone based on Bayesian classifier. Water-rich grade prediction model;
- Step 3 Training and testing the prediction model of the water-rich grade of the coal seam roof sandstone based on Bayesian classifier
- Step 4 Sampling the sampling site of the sandstone roof of the underground coal seam, and analyzing the coal seam roof sandstone of the sampling site to obtain the six influencing factors affecting the water-rich grade, and inputting the six influencing factors into the training and testing.
- the prediction model of the water-rich grade of coal seam roof sandstone is realized by Bayesian formula method to predict the water-rich grade of coal seam roof sandstone to be tested.
- preprocessing the data of the sample database specifically includes the following steps:
- Step 23 Attribute reduction, and the repeated data in the discretized sample data is deleted according to the rough set theory.
- step 3 training and testing the prediction model of the water-rich grade of the coal seam roof sandstone based on the Bayesian classifier includes the following steps:
- Step 31 sensitivity, calculated by following formula (4),
- Step 32 specificity, calculated according to the following formula (5),
- Step 33 accuracy, calculated by following equation (6),
- step 4 using the Bayesian formula method to realize the prediction of the water-rich grade of the coal seam roof sandstone to be tested includes the following steps:
- Step 41 Calculate the probability of occurrence of each water-rich level
- Step 42 Calculate the probability of occurrence of each attribute level separately.
- S ik - the k-th attribute C k has an attribute level of X k and the corresponding number of training samples with a water-rich level of Y i ,
- Step 43 using formula (9), obtaining the attribution water enrichment level of the sample to be tested at the sampling site to be tested,
- Y(X) the water enrichment level of the sample X to be tested at the sampling site to be tested.
- the method for predicting the water-rich grade of the coal seam roof sandstone of the invention has the following characteristics and advantages:
- the method for predicting the water-rich grade of the coal seam roof sandstone of the invention has the main workload arranged in the early stage, and only needs to be sampled at the sampling place to be tested in the later stage, and according to the six influencing factors affecting the water-rich level, the accurate and efficient manner can be accurately and efficiently It is predicted that the water-rich grade of the coal seam roof sandstone at the sampling site will provide guarantee for the efficient and safe construction of the coal mine and has a high promotion value.
- FIG. 1 is a flow chart of a method for predicting the water-rich grade of a coal seam roof sandstone according to an embodiment of the present invention.
- the method for predicting the water-rich grade of the coal seam roof sandstone of the present embodiment is analyzed and verified in more than 10 coal mines such as Caozhuang Coal Mine, Xinyi Coal Mine and Zhaozhuang Mine Field, and the prediction accuracy is very high, and is now collected in Xinyi Coal Mine. 38 sets of samples are illustrated.
- the present embodiment provides a method for predicting the water-rich grade of a coal seam roof sandstone, which includes the following four steps.
- Step 1 Sampling the 38 sampling sites of the coal seam roof sandstone in the underground, and analyzing the coal seam roof sandstone at each sampling site to obtain three water-rich grades and six influencing factors affecting the water-rich grade, three water-rich grades. Including weak (indicated by "1" in the table), medium (indicated by “2” in the table), and strong (indicated by “3” in the table), the six influencing factors include sandstone thickness C 1 , mudstone thickness C 2 , core take-up rate C 3 , slurry loss C 4 , fracture density C 5 and breaking strength C 6 , the above-mentioned sampling data of the first 27 sampling points are input into the computer to establish a sample database, wherein the coal seam roof at each sampling point The water-rich grade of sandstone needs to be determined after the construction of the coal mine.
- Table 1 The original sample data is shown in Table 1:
- Step 2 After preprocessing the 38 sets of data in the sample database, the six influencing factors of the sampled data at each sampling are used as input vectors, and the three water-rich levels are used as output vectors to establish a coal seam based on Bayesian classifier. Prediction model for water-rich grade of roof sandstone.
- the preprocessing of the data of the sample database specifically includes the following steps:
- Step 23 Attribute reduction, and the repeated data in the discretized sample data is deleted according to the rough set theory.
- the first 27 sets of discretized sample data are selected as the training sample U1, and the remaining 11 sets of discretized sample data are used as the sample V to be tested, and the discretized sample data is opened in the UltraEdit software to be "-inf-0.3333", "0.3333- The three segments of 0.6667” and "0.6667-inf" are replaced by 1, 2, and 3, respectively. After the editing is completed, the attribute reduction results are shown in Table 4:
- Step 3 The Bayesian classifier-based coal seam roof sandstone water-rich grade prediction model is trained and tested. Weka software is used to train the Bayesian classifier based coal seam roof sandstone water-rich grade prediction model and obtain training results. Confusion matrix, the model is tested according to the confusion matrix, the specific steps are as follows:
- the confusion matrix for the prediction model obtained in the Confusion Matrix is as follows:
- Step 31 sensitivity, calculated by following formula (4),
- Step 32 specificity, calculated according to the following formula (5),
- Step 33 accuracy, calculated by following equation (6),
- Step 4 On-site sampling of the sampling site to be tested in the roof sandstone of the underground coal seam.
- 28 to 38 sets of samples are taken as samples to be tested, and the coal seam roof sandstone to be tested is analyzed to obtain the six factors affecting the water-rich level. Influencing factors, the six influencing factors are input into the trained and tested Bayesian classifier-based coal seam roof sandstone water-rich grade prediction model, and the Bayesian formula method is used to realize the coal seam roof sandstone rich in the sampling area to be tested. Water grade prediction.
- the first 27 sets of data are taken as the new training sample U 2 , and the remaining 11 sets of data are taken as the sample V to be tested.
- a total of 6 numerical properties are normalized and discretized and divided into three segments, using (X 11 , X 12 , X 13 ), (X 21 , X 22 , X 23 ), (X 31 , X 32 , X 33 ), (X 41 , X 42 , X 43 ), (X 51 , X 52 , X 53 ), (X 61 , X 62 , X 63
- Step 41 Calculate the probability of occurrence of each water-rich level
- Step 42 Calculate the probability of occurrence of each attribute level separately.
- S ik - the k-th attribute C k has an attribute level of X k and the corresponding number of training samples with a water-rich level of Y i ,
- Step 43 using formula (9), obtaining the attribution water enrichment level of the sample to be tested at the sampling site to be tested,
- Y(X) the water enrichment level of the sample X to be tested at the sampling site to be tested.
- the water-rich grade of the coal seam roof sandstone of the sample V to be tested is predicted.
- the 24th group of samples to be tested in Table 6 is taken as an example, and each attribute level is (X 11 , X 23 , X 31 , X 43 , X 51 , X 63 ), the prediction of the water-rich grade of the coal seam roof sandstone, the specific steps are as follows:
- Step1 Calculate the probability of occurrence of each water-rich grade of the coal seam roof:
- Step 2 respectively calculate the probability of occurrence of each attribute level:
- the probability of occurrence of the first attribute sandstone thickness (C 1 ) attribute level of X 11 is:
- the probability of occurrence of the second attribute mudstone thickness (C 2 ) attribute level of X 23 is:
- the probability of occurrence of the third attribute core take-off rate (C 3 ) attribute level is X 31 is:
- the probability of occurrence of the fourth attribute slurry loss (C 4 ) attribute level of X 43 is:
- the probability of occurrence of the fifth attribute fracture density (C 5 ) attribute level of X 43 is:
- the probability of occurrence of the sixth attribute breaking strength (C 6 ) attribute level of X 43 is:
- Step 3 Predict the water enrichment level of the sample to be tested (Group 24) at the sampling site to be tested:
- the prediction model predicts 10 correct samples and 1 error, and the prediction accuracy is 90.9%. It can be concluded that the prediction method of the water-rich grade of the coal seam roof sandstone of the present embodiment has a high prediction accuracy, and provides guarantee for efficient and safe construction of the coal mine, and has high promotion value.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Remote Sensing (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Acoustics & Sound (AREA)
- Environmental & Geological Engineering (AREA)
- Geology (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geophysics (AREA)
- Analysing Materials By The Use Of Radiation (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A method for predicting water-rich levels of sandstone in a coal seam roof comprises the following steps: A. performing on-site sampling of sandstone at N sampling locations in an underground coal seam roof, and analyzing, at each of the sampling locations, the sandstone in the coal seam roof to obtain three water-rich levels thereof and six influencing factors influencing the water-rich levels, so as to establish a sample database; B. after pre-processing data of the sample database, establishing a Bayes classifier based model for predicting water-rich levels of sandstone in a coal seam roof; C. training and testing the Bayes classifier based model for predicting water-rich levels of sandstone in a coal seam roof; and D. performing on-site sampling at a sampling location of the sandstone in the underground coal seam roof, and using a Bayes formula to predict water-rich levels of the sandstone at the sampling location in the coal seam roof on the basis of the Bayes classifier based model.
Description
本发明涉及地下煤炭开采安全生产技术领域,特别是涉及一种煤层顶板砂岩富水性等级预测方法。The invention relates to the technical field of underground coal mining safety production, in particular to a method for predicting the water-rich grade of coal seam roof sandstone.
煤矿煤层顶板水害是矿井水害之一,严重威胁了煤炭的安全回采。矿井顶板水的涌出,不仅恶化了工作面的生产环境,而且给矿井生产带来不安全因素。由于矿井顶板砂岩富水性未能查清,对顶板水涌出量的预测不当,要么过分增加安全系数而极大地增加了防治工程的费用;要么安全系数过小,矿井(采区)防排水能力不足,造成淹井淹采区甚至人员伤亡的重大事故。The water damage in the coal seam roof is one of the mine water hazards, which seriously threatens the safe recovery of coal. The gushing of the roof water of the mine not only worsens the production environment of the working face, but also brings unsafe factors to the mine production. Because the water content of the roof sandstone of the mine could not be ascertained, the prediction of the water discharge from the roof is improper, or the safety factor is excessively increased, which greatly increases the cost of the prevention and control project; or the safety factor is too small, and the mine (mining area) has the ability to prevent and discharge water. Insufficient, causing major accidents in flooded areas and even casualties.
现有技术中,煤层顶板砂岩富水性等级预测方法有多种,主要包括单因素指标法、现场实测法和神经网络法。但是,上述方法均存在不同程度的局限性或先天的不足之处:In the prior art, there are various methods for predicting the water-rich grade of coal seam roof sandstone, including single factor index method, field measurement method and neural network method. However, all of the above methods have different degrees of limitations or innate shortcomings:
单因素指标法,由于建立在单个的影响因素的基础上进行富水性评估,尽管方法简单、操作简便、快捷。但由于煤层顶板砂岩富水性是诸多因素共同作用的结果,并且多个影响因素之间相互制约,相互影响,这种建立在单个的影响因素的基础上进行富水性评估方法,存在较大的局限性与片面性。The single factor indicator method is based on the establishment of a single influencing factor for water-rich assessment, although the method is simple, easy to operate and fast. However, since the water-richness of the roof sandstone of the coal seam is the result of many factors, and the multiple influencing factors mutually restrict each other and influence each other, the method of establishing the water-rich assessment based on the individual influencing factors has great limitations. Sex and one-sidedness.
现场实测法是最为准确、可靠的方法。但是,其工程量大、费时费力,人工成本高,加上井下地理环境特殊、巷道综错复杂等因素的制约,很多时候,无法进行大规模或全方位的现场实测。Field measurement is the most accurate and reliable method. However, due to its large amount of engineering, time-consuming and labor-intensive, high labor costs, and the constraints of the underground geography environment and complex roadway complex, many times, large-scale or all-round on-site measurement is impossible.
神经网络法,是建立在分析煤层顶板砂岩富水性的影响因素及现场实测数据的基础上,构建煤层顶板砂岩富水性的神经网络预测模型。但是,神经网络预测模型进行权值与阈值更新时易陷入局部最优解,预测准确率相对不高,并且推广能力较差。The neural network method is based on the analysis of the factors affecting the water-richness of the roof sandstone in the coal seam and the field measured data. The neural network prediction model for the water-rich roof of the coal seam roof sandstone is constructed. However, the neural network prediction model is easy to fall into the local optimal solution when the weight and threshold are updated, the prediction accuracy is relatively low, and the promotion ability is poor.
发明内容Summary of the invention
本发明的目的在于提供一种煤层顶板砂岩富水性等级预测方法,解决目前煤层顶板砂岩富水性等级预测不准确、预测效率较低的技术问题。The object of the present invention is to provide a method for predicting the water-rich grade of coal seam roof sandstone, and to solve the technical problem that the prediction of the water-rich grade of the coal seam roof sandstone is inaccurate and the prediction efficiency is low.
本发明提供一种煤层顶板砂岩富水性等级预测方法,包括以下步骤:步骤一、对井下煤层顶板砂岩的N个取样处进行现场取样,对每一取样处的煤层顶板砂岩分析得到其三个富水性等级和影响富水性等级的六个影响因素,三个富水性等级包括较弱、中等和较强,六个影响因素包括砂岩厚度C
1、泥岩厚度C
2、岩芯采取率C
3、浆液漏失量C
4、断裂密度C
5以及断裂强度C
6,将N个取样处的上述取样数据建立样本数据库,其中,每一取样处的煤层顶板砂岩的富水性等级需要在煤矿施工后进行确定;
The invention provides a method for predicting the water-rich grade of coal seam roof sandstone, comprising the following steps: Step 1: sampling the N sampling sites of the coal seam roof sandstone in the underground, and analyzing the coal seam roof sandstone at each sampling point to obtain the three rich The water-based grade and the six influencing factors affecting the water-rich grade, the three water-rich grades include weak, medium, and strong, and the six influencing factors include sandstone thickness C 1 , mudstone thickness C 2 , core take-up rate C 3 , and slurry The leakage amount C 4 , the fracture density C 5 and the breaking strength C 6 are used to establish a sample database of the above sampling data at the N sampling points, wherein the water-rich grade of the coal seam roof sandstone at each sampling point needs to be determined after the coal mine construction;
步骤二、对样本数据库的数据进行预处理后,将每一取样处的取样数据的六个影响因素作为输入向量,三个富水性等级作为输出向量,建立基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型;Step 2: After preprocessing the data of the sample database, the six influencing factors of the sampled data at each sampling point are taken as input vectors, and the three water-rich levels are used as output vectors to establish a coal seam roof sandstone based on Bayesian classifier. Water-rich grade prediction model;
步骤三、对基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型进行训练和检验;Step 3: Training and testing the prediction model of the water-rich grade of the coal seam roof sandstone based on Bayesian classifier;
步骤四、对井下煤层顶板砂岩的待测取样处进行现场取样,对待测取样处的煤层顶板砂岩分析得到其影响富水性等级的六个影响因素,将其六个影响因素输入到训练、检验好的基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型中,利用贝叶斯公式法实现对待测取样处的煤层顶板砂岩富水性等级预测。Step 4: Sampling the sampling site of the sandstone roof of the underground coal seam, and analyzing the coal seam roof sandstone of the sampling site to obtain the six influencing factors affecting the water-rich grade, and inputting the six influencing factors into the training and testing. Based on the Bayesian classifier, the prediction model of the water-rich grade of coal seam roof sandstone is realized by Bayesian formula method to predict the water-rich grade of coal seam roof sandstone to be tested.
进一步的,步骤二中,对样本数据库的数据进行预处理具体包括如下步骤:Further, in step 2, preprocessing the data of the sample database specifically includes the following steps:
步骤21、归一化,按下式(1)计算得出,Step 21, normalization, calculated by following formula (1),
上式(1)中,In the above formula (1),
x
ij——归一化前样本数据,
x ij - normalized pre-sample data,
s
ij——归一化后样本数据,
s ij - normalized sample data,
min(x
j)——原始样本数据中的最小值,
Min(x j ) - the minimum value in the original sample data,
max(x
j)——原始样本数据中的最大值;
Max(x j ) - the maximum value in the original sample data;
步骤22、离散化,按下式(2)和(3)计算得出,Step 22, discretization, calculated by following equations (2) and (3),
上式(2)中,In the above formula (2),
z
ij——离散化后的样本数据,
z ij - discretized sample data,
min(s
j)——归一化后样本数据的最小值,
Min(s j ) - the minimum value of the normalized sample data,
max(s
j)——归一化后样本数据的最大值,
Max(s j ) - the maximum value of the normalized sample data,
上式(3)中,In the above formula (3),
Q——步长;Q - step size;
步骤23、属性约简,根据粗糙集理论删除离散化后样本数据中重复的数据。Step 23. Attribute reduction, and the repeated data in the discretized sample data is deleted according to the rough set theory.
进一步的,步骤三中,对基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型进行训练和检验具体包括如下步骤:Further, in step 3, training and testing the prediction model of the water-rich grade of the coal seam roof sandstone based on the Bayesian classifier includes the following steps:
步骤31、敏感性,按下式(4)计算得出,Step 31, sensitivity, calculated by following formula (4),
上式(4)中,In the above formula (4),
Sensitivity——敏感性,Sensitivity - sensitivity,
positive——已知正样本数目,Positive - the number of known positive samples,
true_positive——相应于正确分类的正样本数目;True_positive - the number of positive samples corresponding to the correct classification;
步骤32、专一性,按下式(5)计算得出,Step 32, specificity, calculated according to the following formula (5),
上式(5)中,In the above formula (5),
Specificity——专一性,Specificity - specificity,
negative——已知负样本的数目,Negative - the number of known negative samples,
true_negative——相应于正确分类的负样本的数目;True_negative - the number of negative samples corresponding to the correct classification;
步骤33、精准度,按下式(6)计算得出,Step 33, accuracy, calculated by following equation (6),
上式(6)中,In the above formula (6),
Accuracy——精准度,Accuracy - precision,
Sensitivity——敏感性,Sensitivity - sensitivity,
positive——已知正样本数目,Positive - the number of known positive samples,
negative——已知负样本的数目,Negative - the number of known negative samples,
Specificity——专一性。Specificity - specificity.
进一步的,步骤四中,利用贝叶斯公式法实现对待测取样处的煤层顶板砂岩富水性等级 预测具体包括如下步骤:Further, in step 4, using the Bayesian formula method to realize the prediction of the water-rich grade of the coal seam roof sandstone to be tested includes the following steps:
步骤41、分别计算各富水性等级的发生概率,Step 41: Calculate the probability of occurrence of each water-rich level,
上式(7)中,In the above formula (7),
Y
i——富水性等级,
Y i - water-rich grade,
S
i——富水性等级为Y
i的样本个数,
S i -- the number of samples with a water-rich grade of Y i ,
S——模型中训练样本的个数,S——the number of training samples in the model,
S
c——所有富水性等级个数,
S c - the number of all water-rich grades,
P(Y
i)——富水性等级Y
i的概率;
P(Y i ) - probability of water-rich grade Y i ;
步骤42、分别计算每个属性等级的发生概率,Step 42: Calculate the probability of occurrence of each attribute level separately.
上式(8)中,In the above formula (8),
k——第k个属性,k - the kth attribute,
C
k——第k个属性的影响指标,
C k - the impact indicator of the kth attribute,
X
k——第k个属性C
k的属性级别,
X k - the attribute level of the kth attribute C k ,
S
ik——第k个属性C
k的属性级别为X
k且对应的富水性等级为Y
i的训练样本个数,
S ik - the k-th attribute C k has an attribute level of X k and the corresponding number of training samples with a water-rich level of Y i ,
S
k——训练样本中X
k的富水性等级个数,
S k —— the number of water-rich grades of X k in the training sample,
P(C
k=X
k|Y
i)——富水性等级Y
i中属性值为X
k的概率;
P(C k =X k |Y i )——the probability that the attribute value is X k in the water-rich level Y i ;
步骤43、利用公式(9)得出待测取样处的待测样本的归属富水性等级,Step 43, using formula (9), obtaining the attribution water enrichment level of the sample to be tested at the sampling site to be tested,
上式(9)中,In the above formula (9),
X——待测取样处的待测样本,X——the sample to be tested at the sampling site to be tested,
Y(X)——待测取样处的待测样本X的归属富水性等级。Y(X)——the water enrichment level of the sample X to be tested at the sampling site to be tested.
与现有技术相比,本发明的煤层顶板砂岩富水性等级预测方法及具有以下特点和优点:Compared with the prior art, the method for predicting the water-rich grade of the coal seam roof sandstone of the invention has the following characteristics and advantages:
本发明的煤层顶板砂岩富水性等级预测方法,将主工作量安排在前期,后期只需要在待 测取样处进行现场取样,根据其影响富水性等级的六个影响因素,就可准确、高效地预测取样处煤层顶板砂岩富水性等级,为煤矿的高效、安全施工提供保障,具有较高的推广价值。The method for predicting the water-rich grade of the coal seam roof sandstone of the invention has the main workload arranged in the early stage, and only needs to be sampled at the sampling place to be tested in the later stage, and according to the six influencing factors affecting the water-rich level, the accurate and efficient manner can be accurately and efficiently It is predicted that the water-rich grade of the coal seam roof sandstone at the sampling site will provide guarantee for the efficient and safe construction of the coal mine and has a high promotion value.
结合附图阅读本发明的具体实施方式后,本发明的特点和优点将变得更加清楚。The features and advantages of the present invention will become more apparent from the detailed description of the embodiments.
为了更清楚地说明本发明实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图是本发明的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below. Obviously, the drawings in the following description are Some embodiments of the present invention may also be used to obtain other drawings based on these drawings without departing from the art.
图1为本发明实施例煤层顶板砂岩富水性等级预测方法的流程图。1 is a flow chart of a method for predicting the water-rich grade of a coal seam roof sandstone according to an embodiment of the present invention.
本实施例的煤层顶板砂岩富水性等级预测方法,在曹庄煤矿、新驿煤矿、赵庄井田等10余家煤矿进行分析验证,其预测准确率很高,现以在新驿煤矿采集到的38组样本举例说明。其中,以1至27组样本作为现场取样待测样本,分析得到1至27组样本三个富水性等级和影响富水性等级的六个影响因素,并以1至27组取样数据建立样本数据库;以28至38组样本作为待测取样样本,分析得到28至38组样本各样本三个富水性等级和影响富水性等级的六个影响因素,需要说明的是,28至38组样本中的“三个富水性等级”在工程实践中不需要分析,而且也只能是在煤矿施工后进行分析,在此处分析得到其“三个富水性等级”,仅仅用于验证本实施例的煤层顶板砂岩富水性等级预测方法的预测准确率。The method for predicting the water-rich grade of the coal seam roof sandstone of the present embodiment is analyzed and verified in more than 10 coal mines such as Caozhuang Coal Mine, Xinyi Coal Mine and Zhaozhuang Mine Field, and the prediction accuracy is very high, and is now collected in Xinyi Coal Mine. 38 sets of samples are illustrated. Among them, 1 to 27 samples were taken as samples for on-site sampling, and the three influencing factors of the water-rich grades and the water-influence grades of the samples from 1 to 27 were analyzed, and the sample database was established with 1 to 27 sets of sampling data; Using 28 to 38 samples as the sample to be tested, the analysis of the three water-rich grades of each sample from 28 to 38 samples and the six influencing factors affecting the water-rich grade, it should be noted that in the 28 to 38 samples The three water-rich grades do not need to be analyzed in engineering practice, and can only be analyzed after coal mine construction. The “three water-rich grades” are analyzed here, which is only used to verify the coal seam roof of this embodiment. Predictive accuracy of prediction methods for sandstone water-rich grades.
参照图1,本实施例提供一种煤层顶板砂岩富水性等级预测方法,包括以下四个步骤。Referring to FIG. 1 , the present embodiment provides a method for predicting the water-rich grade of a coal seam roof sandstone, which includes the following four steps.
步骤一、对井下煤层顶板砂岩的38个取样处进行现场取样,对每一取样处的煤层顶板砂岩分析得到其三个富水性等级和影响富水性等级的六个影响因素,三个富水性等级包括较弱(表中用“1”表示)、中等(表中用“2”表示)和较强(表中用“3”表示),六个影响因素包括砂岩厚度C
1、泥岩厚度C
2、岩芯采取率C
3、浆液漏失量C
4、断裂密度C
5以及断裂强度C
6,将前27个取样处的上述取样数据输入计算机,建立样本数据库,其中,每一取样处的煤层顶板砂岩的富水性等级需要在煤矿施工后进行确定。原始样本数据如表1所示:
Step 1. Sampling the 38 sampling sites of the coal seam roof sandstone in the underground, and analyzing the coal seam roof sandstone at each sampling site to obtain three water-rich grades and six influencing factors affecting the water-rich grade, three water-rich grades. Including weak (indicated by "1" in the table), medium (indicated by "2" in the table), and strong (indicated by "3" in the table), the six influencing factors include sandstone thickness C 1 , mudstone thickness C 2 , core take-up rate C 3 , slurry loss C 4 , fracture density C 5 and breaking strength C 6 , the above-mentioned sampling data of the first 27 sampling points are input into the computer to establish a sample database, wherein the coal seam roof at each sampling point The water-rich grade of sandstone needs to be determined after the construction of the coal mine. The original sample data is shown in Table 1:
表1 原始样本数据Table 1 Raw sample data
序号Serial number | C 1(mm) C 1 (mm) | C 2(mm) C 2 (mm) | C 3(m 3/h) C 3 (m 3 /h) | C 4(%) C 4 (%) | C 5(MPa) C 5 (MPa) | C 6(kg/m 3) C 6 (kg/m 3 ) | YY |
11 | 530530 | 364364 | 0.0270.027 | 0.6290.629 | 0.0040.004 | 0.50.5 | 11 |
22 | 383383 | 635635 | 0.0040.004 | 0.7880.788 | 0.0020.002 | 0.50.5 | 11 |
33 | 209209 | 149149 | 00 | 11 | 0.0020.002 | 0.50.5 | 11 |
44 | 575575 | 573573 | 00 | 0.8180.818 | 0.0520.052 | 0.50.5 | 11 |
55 | 786786 | 602602 | 0.0350.035 | 0.3260.326 | 0.0760.076 | 0.50.5 | 11 |
66 | 595595 | 795795 | 0.0470.047 | 0.8170.817 | 0.0440.044 | 0.50.5 | 11 |
77 | 303303 | 859859 | 0.0070.007 | 0.960.96 | 00 | 00 | 11 |
88 | 430430 | 756756 | 00 | 11 | 00 | 00 | 11 |
99 | 342342 | 578578 | 0.0110.011 | 0.8990.899 | 00 | 00 | 11 |
1010 | 416416 | 781781 | 0.0090.99 | 0.930.93 | 00 | 00 | 11 |
1111 | 518518 | 596596 | 0.0780.078 | 0.5340.534 | 00 | 00 | 22 |
1212 | 513513 | 108108 | 0.5930.593 | 0.0880.088 | 00 | 00 | 22 |
1313 | 513513 | 123123 | 0.6110.611 | 0.0480.048 | 00 | 00 | 22 |
1414 | 516516 | 700700 | 0.2360.236 | 0.0610.061 | 00 | 00 | 22 |
1515 | 512512 | 467467 | 0.3370.337 | 0.0550.055 | 00 | 00 | 22 |
1616 | 580580 | 248248 | 0.0290.029 | 0.8290.829 | 00 | 00 | 22 |
1717 | 527527 | 581581 | 0.0410.041 | 0.7890.789 | 00 | 00 | 22 |
1818 | 553553 | 616616 | 0.0360.036 | 0.8230.823 | 00 | 00 | 22 |
1919 | 520520 | 526526 | 0.0530.053 | 0.7860.786 | 00 | 00 | 22 |
2020 | 217217 | 597597 | 0.2440.244 | 0.0890.089 | 00 | 00 | 33 |
21twenty one | 210210 | 606606 | 0.23410.2341 | 0.0630.063 | 00 | 00 | 33 |
22twenty two | 181181 | 560560 | 0.7530.753 | 0.340.34 | 00 | 00 | 33 |
23twenty three | 180180 | 317317 | 0.7070.707 | 0.0980.098 | 00 | 00 | 33 |
24twenty four | 521521 | 171171 | 0.7180.718 | 0.0520.052 | 0.00110.0011 | 0.1110.111 | 33 |
2525 | 296296 | 818818 | 0.5430.543 | 0.0720.072 | 0.00080.0008 | 0.3330.333 | 33 |
2626 | 704704 | 398398 | 0.2010.201 | 0.2150.215 | 0.00030.0003 | 0.1110.111 | 33 |
2727 | 475475 | 589589 | 0.8640.864 | 0.0140.014 | 0.00130.0013 | 0.1110.111 | 33 |
2828 | 309309 | 610610 | 0.0070.007 | 0.960.96 | 0.0020.002 | 0.50.5 | 11 |
2929 | 253253 | 600600 | 00 | 11 | 00 | 00 | 11 |
3030 | 408408 | 185185 | 0.0050.005 | 0.9320.932 | 00 | 00 | 11 |
3131 | 526526 | 238238 | 0.0250.025 | 0.8290.829 | 00 | 00 | 11 |
3232 | 518518 | 711711 | 0.0470.047 | 0.6620.662 | 00 | 00 | 22 |
3333 | 502502 | 192192 | 0.0620.062 | 0.6130.613 | 00 | 00 | 22 |
3434 | 713713 | 583583 | 0.2450.245 | 0.0280.028 | 00 | 00 | 22 |
3535 | 516516 | 237237 | 0.0650.065 | 0.6240.624 | 00 | 00 | 22 |
3636 | 445445 | 585585 | 0.2440.244 | 0.1510.151 | 0.00040.0004 | 0.2220.222 | 33 |
3737 | 228228 | 569569 | 0.2230.223 | 0.0720.072 | 00 | 00 | 33 |
3838 | 425425 | 648648 | 0.1320.132 | 0.2390.239 | 0.00020.0002 | 0.1110.111 | 33 |
步骤二、对样本数据库的38组数据进行预处理后,将每一取样处的取样数据的六个影响因素作为输入向量,三个富水性等级作为输出向量,建立基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型。Step 2: After preprocessing the 38 sets of data in the sample database, the six influencing factors of the sampled data at each sampling are used as input vectors, and the three water-rich levels are used as output vectors to establish a coal seam based on Bayesian classifier. Prediction model for water-rich grade of roof sandstone.
其中,对样本数据库的数据进行预处理具体包括如下步骤:The preprocessing of the data of the sample database specifically includes the following steps:
步骤21、归一化,按下式(1)计算得出,Step 21, normalization, calculated by following formula (1),
上式(1)中,In the above formula (1),
x
ij——归一化前样本数据,
x ij - normalized pre-sample data,
s
ij——归一化后样本数据,
s ij - normalized sample data,
min(x
j)——原始样本数据中的最小值,
Min(x j ) - the minimum value in the original sample data,
max(x
j)——原始样本数据中的最大值;
Max(x j ) - the maximum value in the original sample data;
归一化结果如表2所示:The normalized results are shown in Table 2:
表2 归一化结果Table 2 normalized results
序号Serial number | C 1(mm) C 1 (mm) | C 2(mm) C 2 (mm) | C 3(m 3/h) C 3 (m 3 /h) | C 4(%) C 4 (%) | C 5(MPa) C 5 (MPa) | C 6(kg/m 3) C 6 (kg/m 3 ) | YY |
11 | 0.57760.5776 | 0.34090.3409 | 0.03130.0313 | 0.62370.6237 | 0.07690.0769 | 11 | 11 |
22 | 0.3350.335 | 0.70170.7017 | 0.00460.0046 | 0.7850.785 | 0.03850.0385 | 11 | 11 |
33 | 0.04790.0479 | 0.05460.0546 | 00 | 11 | 0.03850.0385 | 11 | 11 |
44 | 0.65180.6518 | 0.61920.6192 | 00 | 0.81540.8154 | 11 | 11 | 11 |
55 | 11 | 0.65780.6578 | 0.04050.0405 | 0.31640.3164 | 0.88460.8846 | 11 | 11 |
66 | 0.68480.6848 | 0.91480.9148 | 0.05440.0544 | 0.81440.8144 | 0.84620.8462 | 11 | 11 |
77 | 0.2030.203 | 11 | 0.00810.0081 | 0.95940.9594 | 00 | 00 | 11 |
88 | 0.41250.4125 | 0.86280.8628 | 00 | 11 | 00 | 00 | 11 |
99 | 0.26730.2673 | 0.62580.6258 | 0.01270.0127 | 0.89760.8976 | 00 | 00 | 11 |
1010 | 0.38940.3894 | 0.89610.8961 | 0.01040.0104 | 0.9290.929 | 00 | 00 | 11 |
1111 | 0.55780.5578 | 0.64980.6498 | 0.09030.0903 | 0.52740.5274 | 00 | 00 | 22 |
1212 | 0.54950.5495 | 00 | 0.68630.6863 | 0.07510.0751 | 00 | 00 | 22 |
1313 | 0.54950.5495 | 0.020.02 | 0.70720.7072 | 0.03450.0345 | 00 | 00 | 22 |
1414 | 0.55450.5545 | 0.78830.7883 | 0.27310.2731 | 0.04770.0477 | 00 | 00 | 22 |
1515 | 0.54790.5479 | 0.4780.478 | 0.390.39 | 0.04160.0416 | 00 | 00 | 22 |
1616 | 0.66010.6601 | 0.18640.1864 | 0.03360.0336 | 0.82660.8266 | 00 | 00 | 22 |
1717 | 0.57260.5726 | 0.62980.6298 | 0.04750.0475 | 0.7860.786 | 00 | 00 | 22 |
1818 | 0.61550.6155 | 0.67640.6764 | 0.04170.0417 | 0.82050.8205 | 00 | 00 | 22 |
1919 | 0.56110.5611 | 0.55660.5566 | 0.06130.0613 | 0.7830.783 | 00 | 00 | 22 |
2020 | 0.06110.0611 | 0.65110.6511 | 0.28240.2824 | 0.07610.0761 | 00 | 00 | 33 |
21twenty one | 0.04950.0495 | 0.66310.6631 | 0.27090.2709 | 0.04970.0497 | 00 | 00 | 33 |
22twenty two | 0.00170.0017 | 0.60190.6019 | 0.87150.8715 | 0.33060.3306 | 00 | 00 | 33 |
23twenty three | 00 | 0.27830.2783 | 0.81830.8183 | 0.08520.0852 | 00 | 00 | 33 |
24twenty four | 0.56270.5627 | 0.08390.0839 | 0.8310.831 | 0.03850.0385 | 0.02120.0212 | 0.2220.222 | 33 |
2525 | 0.19140.1914 | 0.94540.9454 | 0.62850.6285 | 0.05880.0588 | 0.01540.0154 | 0.6660.666 | 33 |
2626 | 0.86470.8647 | 0.38620.3862 | 0.23260.2326 | 0.20390.2039 | 0.00580.0058 | 0.2220.222 | 33 |
2727 | 0.48680.4868 | 0.64050.6405 | 11 | 00 | 0.0250.025 | 0.2220.222 | 33 |
2828 | 0.21290.2129 | 0.66840.6684 | 0.00810.0081 | 0.95940.9594 | 0.03850.0385 | 11 | 11 |
2929 | 0.12050.1205 | 0.65510.6551 | 00 | 11 | 00 | 00 | 11 |
3030 | 0.37620.3762 | 0.10250.1025 | 0.00580.0058 | 0.9310.931 | 00 | 00 | 11 |
3131 | 0.5710.571 | 0.17310.1731 | 0.02890.0289 | 0.82660.8266 | 00 | 00 | 11 |
3232 | 0.55780.5578 | 0.80290.8029 | 0.05440.0544 | 0.65720.6572 | 00 | 00 | 22 |
3333 | 0.53140.5314 | 0.11190.1119 | 0.07180.0718 | 0.60750.6075 | 00 | 00 | 22 |
3434 | 0.87950.8795 | 0.63250.6325 | 0.28360.2836 | 0.01420.0142 | 00 | 00 | 22 |
3535 | 0.55450.5545 | 0.17180.1718 | 0.07520.0752 | 0.61870.6187 | 00 | 00 | 22 |
3636 | 0.43730.4373 | 0.63520.6352 | 0.28240.2824 | 0.13890.1389 | 0.00770.0077 | 0.4440.444 | 33 |
3737 | 0.07920.0792 | 0.61380.6138 | 0.25810.2581 | 0.05880.0588 | 00 | 00 | 33 |
3838 | 0.40430.4043 | 0.7190.719 | 0.15280.1528 | 0.22820.2282 | 0.00380.0038 | 0.2220.222 | 33 |
步骤22、离散化,按下式(2)和(3)计算得出,Step 22, discretization, calculated by following equations (2) and (3),
上式(2)中,In the above formula (2),
z
ij——离散化后的样本数据,
z ij - discretized sample data,
min(s
j)——归一化后样本数据的最小值,
Min(s j ) - the minimum value of the normalized sample data,
max(s
j)——归一化后样本数据的最大值,
Max(s j ) - the maximum value of the normalized sample data,
上式(3)中,In the above formula (3),
Q——步长;Q - step size;
借助Weka软件将归一化后的样本数据进行离散化处理,离散化结果如表3所示:The normalized sample data is discretized by Weka software. The discretization results are shown in Table 3:
表3 离散化结果Table 3 Discretization results
序号Serial number | C 1(mm) C 1 (mm) | C 2(mm) C 2 (mm) | C 3(m 3/h) C 3 (m 3 /h) | C 4(%) C 4 (%) | C 5(MPa) C 5 (MPa) | C 6(kg/m 3) C 6 (kg/m 3 ) | YY |
11 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | 11 |
22 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | 11 |
33 | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | 11 |
44 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (0.6667-inf)'(0.6667-inf)' | (0.6667-inf)'(0.6667-inf)' | 11 |
55 | (0.6667-inf)'(0.6667-inf)' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (0.6667-inf)'(0.6667-inf)' | 11 |
66 | (0.6667-inf)'(0.6667-inf)' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (0.6667-inf)'(0.6667-inf)' | (0.6667-inf)'(0.6667-inf)' | 11 |
77 | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 11 |
88 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 11 |
99 | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 11 |
1010 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 11 |
1111 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
1212 | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
1313 | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
1414 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
1515 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.3333-0.6667]'(0.3333-0.6667]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
1616 | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
1717 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
1818 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
1919 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
2020 | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 33 |
21twenty one | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 33 |
22twenty two | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 33 |
23twenty three | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 33 |
24twenty four | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 33 |
2525 | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667](0.3333-0.6667] | 33 |
2626 | (0.6667-inf)'(0.6667-inf)' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 33 |
2727 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.3333-0.6667]'(0.3333-0.6667]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 33 |
2828 | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | 11 |
2929 | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 11 |
3030 | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 11 |
3131 | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 11 |
3232 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
3333 | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
3434 | (0.6667-inf)'(0.6667-inf)' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
3535 | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 22 |
3636 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | 33 |
3737 | (-inf-0.3333]'(-inf-0.3333]' | (0.3333-0.6667]'(0.3333-0.6667]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 33 |
3838 | (0.3333-0.6667]'(0.3333-0.6667]' | (0.6667-inf)'(0.6667-inf)' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | (-inf-0.3333]'(-inf-0.3333]' | 33 |
步骤23、属性约简,根据粗糙集理论删除离散化后样本数据中重复的数据。Step 23. Attribute reduction, and the repeated data in the discretized sample data is deleted according to the rough set theory.
选取前27组离散化后的样本数据作为训练样本U1,剩余11组离散化后的样本数据作为待测样本V,在UltraEdit软件中打开离散化样本数据将“-inf-0.3333”、“0.3333-0.6667”、“0.6667-inf”三段分别替换为1、2、3。编辑完成后,属性约简结果见表4所示:The first 27 sets of discretized sample data are selected as the training sample U1, and the remaining 11 sets of discretized sample data are used as the sample V to be tested, and the discretized sample data is opened in the UltraEdit software to be "-inf-0.3333", "0.3333- The three segments of 0.6667" and "0.6667-inf" are replaced by 1, 2, and 3, respectively. After the editing is completed, the attribute reduction results are shown in Table 4:
表4 属性约简结果Table 4 attribute reduction results
序号Serial number | C 1(mm) C 1 (mm) | C 2(mm) C 2 (mm) | C 3(m 3/h) C 3 (m 3 /h) | C 4(%) C 4 (%) | C 5(MPa) C 5 (MPa) | C 6(kg/m 3) C 6 (kg/m 3 ) | YY |
11 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | 11 |
22 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | 11 |
33 | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | 11 |
44 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(3)''(3)' | '(3)''(3)' | 11 |
55 | '(3)''(3)' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(2)''(2)' | '(3)''(3)' | 11 |
66 | '(3)''(3)' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(3)''(3)' | '(3)''(3)' | 11 |
77 | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 11 |
88 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 11 |
99 | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 11 |
1010 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 11 |
1111 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | 22 |
1212 | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 22 |
1313 | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 22 |
1414 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 22 |
1515 | '(2]''(2]' | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 22 |
1616 | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 22 |
1717 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 22 |
1818 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 22 |
1919 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 22 |
2020 | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
21twenty one | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
22twenty two | '(1]''(1]' | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
23twenty three | '(1]''(1]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
24twenty four | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
2525 | '(1]''(1]' | '(3)''(3)' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(2]''(2]' | 33 |
2626 | '(3)''(3)' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
2727 | '(2]''(2]' | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
2828 | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | ?? |
2929 | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | ?? |
3030 | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | ?? |
3131 | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | ?? |
3232 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | ?? |
3333 | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | ?? |
3434 | '(3)''(3)' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | ?? |
3535 | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | ?? |
3636 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(2]''(2]' | ?? |
3737 | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | ?? |
3838 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | ?? |
由表4可知,属性约简后发现第8组和第10组数据重复、第12组和第13组数据重复、第17组和第19组数据以及第20组和第21组数据重复,根据粗糙集理论删除第8组、第12组、第17组、以及第20组重复的数据,剩下的23组数据作为新的训练样本U2,如表5所 示:As can be seen from Table 4, after the attribute reduction, it is found that the 8th and 10th sets of data are repeated, the 12th and 13th sets of data are repeated, the 17th and 19th sets of data, and the 20th and 21st sets of data are repeated, according to The rough set theory deletes the data of the eighth group, the 12th group, the 17th group, and the 20th group, and the remaining 23 sets of data are used as the new training sample U2, as shown in Table 5:
表5 新的样本数据Table 5 New sample data
序号Serial number | C 1(mm) C 1 (mm) | C 2(mm) C 2 (mm) | C 3(m 3/h) C 3 (m 3 /h) | C 4(%) C 4 (%) | C 5(MPa) C 5 (MPa) | C 6(kg/m 3) C 6 (kg/m 3 ) | YY |
11 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | 11 |
22 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | 11 |
33 | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | 11 |
44 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(3)''(3)' | '(3)''(3)' | 11 |
55 | '(3)''(3)' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(2)''(2)' | '(3)''(3)' | 11 |
66 | '(3)''(3)' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(3)''(3)' | '(3)''(3)' | 11 |
77 | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 11 |
88 | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 11 |
99 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 11 |
1010 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | 22 |
1111 | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 22 |
1212 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 22 |
1313 | '(2]''(2]' | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 22 |
1414 | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 22 |
1515 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 22 |
1616 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | 22 |
1717 | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
1818 | '(1]''(1]' | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
1919 | '(1]''(1]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
2020 | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
21twenty one | '(1]''(1]' | '(3)''(3)' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(2]''(2]' | 33 |
22twenty two | '(3)''(3)' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
23twenty three | '(2]''(2]' | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | 33 |
24twenty four | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(3)''(3)' | ?? |
2525 | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | ?? |
2626 | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | ?? |
2727 | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | ?? |
2828 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | ?? |
2929 | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | ?? |
3030 | '(3)''(3)' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | ?? |
3131 | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | ?? |
3232 | '(2]''(2]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(2]''(2]' | ?? |
3333 | '(1]''(1]' | '(2]''(2]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | ?? |
3434 | '(2]''(2]' | '(3)''(3)' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | '(1]''(1]' | ?? |
步骤三、对基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型进行训练和检验,通过weka软件对基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型进行训练并得到训练结果的混淆矩阵,根据混淆矩阵对该模型进行检验,具体步骤如下:Step 3: The Bayesian classifier-based coal seam roof sandstone water-rich grade prediction model is trained and tested. Weka software is used to train the Bayesian classifier based coal seam roof sandstone water-rich grade prediction model and obtain training results. Confusion matrix, the model is tested according to the confusion matrix, the specific steps are as follows:
在Confusion Matrix中得到预测模型的混淆矩阵如下所示:The confusion matrix for the prediction model obtained in the Confusion Matrix is as follows:
步骤31、敏感性,按下式(4)计算得出,Step 31, sensitivity, calculated by following formula (4),
上式(4)中,In the above formula (4),
Sensitivity——敏感性,Sensitivity - sensitivity,
positive——已知正样本数目,Positive - the number of known positive samples,
true_positive——相应于正确分类的正样本数目;True_positive - the number of positive samples corresponding to the correct classification;
步骤32、专一性,按下式(5)计算得出,Step 32, specificity, calculated according to the following formula (5),
上式(5)中,In the above formula (5),
Specificity——专一性,Specificity - specificity,
negative——已知负样本的数目,Negative - the number of known negative samples,
true_negative——相应于正确分类的负样本的数目;True_negative - the number of negative samples corresponding to the correct classification;
步骤33、精准度,按下式(6)计算得出,Step 33, accuracy, calculated by following equation (6),
上式(6)中,In the above formula (6),
Accuracy——精准度,Accuracy - precision,
Sensitivity——敏感性,Sensitivity - sensitivity,
positive——已知正样本数目,Positive - the number of known positive samples,
negative——已知负样本的数目,Negative - the number of known negative samples,
Specificity——专一性;Specificity - specificity;
综上,基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型的敏感性、专一性和精准度均在85%以上,说明该模型训练效果很好,适合本课题研究。In summary, the Bayesian classifier based coal seam roof sandstone water-rich grade prediction model sensitivity, specificity and accuracy are above 85%, indicating that the model training effect is very good, suitable for the research of this topic.
步骤四、对井下煤层顶板砂岩的待测取样处进行现场取样,本实施例中以28至38组样本作为待测取样样本,对待测取样处的煤层顶板砂岩分析得到其影响富水性等级的六个影响因素,将其六个影响因素输入到训练、检验好的基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型中,利用贝叶斯公式法实现对待测取样处的煤层顶板砂岩富水性等级预测。Step 4: On-site sampling of the sampling site to be tested in the roof sandstone of the underground coal seam. In this example, 28 to 38 sets of samples are taken as samples to be tested, and the coal seam roof sandstone to be tested is analyzed to obtain the six factors affecting the water-rich level. Influencing factors, the six influencing factors are input into the trained and tested Bayesian classifier-based coal seam roof sandstone water-rich grade prediction model, and the Bayesian formula method is used to realize the coal seam roof sandstone rich in the sampling area to be tested. Water grade prediction.
其中,利用贝叶斯公式法实现对待测取样处的煤层顶板砂岩富水性等级预测具体包括如下步骤:Among them, using the Bayesian formula method to predict the water-rich grade of the coal seam roof sandstone to be tested includes the following steps:
将前27组数据作为新的训练样本U
2,剩余11组数据作为待测样本V,煤层顶板砂岩富水性等级分为三类,记为Y={Y
1,Y
2,Y
3}={较弱,中等,较强},对于砂岩厚度(C
1)、泥岩厚度(C
2)、岩芯采取率(C
3)、浆液漏失量(C
4)、断裂密度(C
5)、断裂强度(C
6)共6个数值型属性我们进行归一化、离散化后分为三段,分别用(X
11,X
12,X
13)、(X
21,X
22,X
23)、(X
31,X
32,X
33)、(X
41,X
42,X
43)、(X
51,X
52,X
53)、(X
61,X
62,X
63)表示砂岩厚度,泥岩厚度,岩芯采取率,浆液漏失量,断裂密度,断裂强度的三个级别,样本数据如表6所示:
The first 27 sets of data are taken as the new training sample U 2 , and the remaining 11 sets of data are taken as the sample V to be tested. The water-rich grade of the coal seam roof sandstone is divided into three categories, which are recorded as Y={Y 1 , Y 2 , Y 3 }={ Weak, medium, strong}, for sandstone thickness (C 1 ), mudstone thickness (C 2 ), core take-up rate (C 3 ), slurry loss (C 4 ), fracture density (C 5 ), breaking strength (C 6 ) A total of 6 numerical properties are normalized and discretized and divided into three segments, using (X 11 , X 12 , X 13 ), (X 21 , X 22 , X 23 ), (X 31 , X 32 , X 33 ), (X 41 , X 42 , X 43 ), (X 51 , X 52 , X 53 ), (X 61 , X 62 , X 63 ) represent sandstone thickness, mudstone thickness, core The three rates of rate, slurry loss, fracture density, and fracture strength are shown in Table 6.
表6 样本数据Table 6 sample data
序号Serial number | C 1(mm) C 1 (mm) | C 2(mm) C 2 (mm) | C 3(m 3/h) C 3 (m 3 /h) | C 4(%) C 4 (%) | C 5(MPa) C 5 (MPa) | C 6(kg/m 3) C 6 (kg/m 3 ) | YY |
11 | X 12 X 12 | X 22 X 22 | X 31 X 31 | X 42 X 42 | X 51 X 51 | X 63 X 63 | Y 1 Y 1 |
22 | X 12 X 12 | X 23 X 23 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 63 X 63 | Y 1 Y 1 |
33 | X 11 X 11 | X 21 X 21 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 63 X 63 | Y 1 Y 1 |
44 | X 12 X 12 | X 22 X 22 | X 31 X 31 | X 43 X 43 | X 53 X 53 | X 63 X 63 | Y 1 Y 1 |
55 | X 13 X 13 | X 22 X 22 | X 31 X 31 | X 41 X 41 | X 52 X 52 | X 63 X 63 | Y 1 Y 1 |
66 | X 13 X 13 | X 23 X 23 | X 31 X 31 | X 43 X 43 | X 53 X 53 | X 63 X 63 | Y 1 Y 1 |
77 | X 11 X 11 | X 23 X 23 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 61 X 61 | Y 1 Y 1 |
88 | X 11 X 11 | X 22 X 22 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 61 X 61 | Y 1 Y 1 |
99 | X 12 X 12 | X 23 X 23 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 61 X 61 | Y 1 Y 1 |
1010 | X 12 X 12 | X 22 X 22 | X 31 X 31 | X 42 X 42 | X 51 X 51 | X 61 X 61 | Y 2 Y 2 |
1111 | X 12 X 12 | X 21 X 21 | X 33 X 33 | X 41 X 41 | X 51 X 51 | X 61 X 61 | Y 2 Y 2 |
1212 | X 12 X 12 | X 23 X 23 | X 31 X 31 | X 41 X 41 | X 51 X 51 | X 61 X 61 | Y 2 Y 2 |
1313 | X 12 X 12 | X 22 X 22 | X 32 X 32 | X 41 X 41 | X 51 X 51 | X 61 X 61 | Y 2 Y 2 |
1414 | X 12 X 12 | X 21 X 21 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 61 X 61 | Y 2 Y 2 |
1515 | X 12 X 12 | X 23 X 23 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 61 X 61 | Y 2 Y 2 |
1616 | X 12 X 12 | X 22 X 22 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 61 X 61 | Y 2 Y 2 |
1717 | X 11 X 11 | X 22 X 22 | X 31 X 31 | X 41 X 41 | X 51 X 51 | X 61 X 61 | Y 3 Y 3 |
1818 | X 11 X 11 | X 22 X 22 | X 33 X 33 | X 41 X 41 | X 51 X 51 | X 61 X 61 | Y 3 Y 3 |
1919 | X 11 X 11 | X 21 X 21 | X 33 X 33 | X 41 X 41 | X 51 X 51 | X 61 X 61 | Y 3 Y 3 |
2020 | X 12 X 12 | X 21 X 21 | X 33 X 33 | X 41 X 41 | X 51 X 51 | X 61 X 61 | Y 3 Y 3 |
21twenty one | X 11 X 11 | X 23 X 23 | X 32 X 32 | X 41 X 41 | X 51 X 51 | X 62 X 62 | Y 3 Y 3 |
22twenty two | X 13 X 13 | X 22 X 22 | X 31 X 31 | X 41 X 41 | X 51 X 51 | X 61 X 61 | Y 3 Y 3 |
23twenty three | X 12 X 12 | X 22 X 22 | X 33 X 33 | X 41 X 41 | X 51 X 51 | X 61 X 61 | Y 3 Y 3 |
24twenty four | X 11 X 11 | X 23 X 23 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 63 X 63 | ?? |
2525 | X 11 X 11 | X 22 X 22 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 61 X 61 | ?? |
2626 | X 12 X 12 | X 21 X 21 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 61 X 61 | ?? |
2727 | X 12 X 12 | X 21 X 21 | X 31 X 31 | X 43 X 43 | X 51 X 51 | X 61 X 61 | ?? |
2828 | X 12 X 12 | X 23 X 23 | X 31 X 31 | X 42 X 42 | X 51 X 51 | X 61 X 61 | ?? |
2929 | X 12 X 12 | X 21 X 21 | X 31 X 31 | X 42 X 42 | X 51 X 51 | X 61 X 61 | ?? |
3030 | X 13 X 13 | X 22 X 22 | X 31 X 31 | X 41 X 41 | X 51 X 51 | X 61 X 61 | ?? |
3131 | X 12 X 12 | X 21 X 21 | X 31 X 31 | X 42 X 42 | X 51 X 51 | X 61 X 61 | ?? |
3232 | X 12 X 12 | X 22 X 22 | X 31 X 31 | X 41 X 41 | X 51 X 51 | X 62 X 62 | ?? |
3333 | X 11 X 11 | X 22 X 22 | X 31 X 31 | X 41 X 41 | X 51 X 51 | X 61 X 61 | ?? |
3434 | X 12 X 12 | X 23 X 23 | X 31 X 31 | X 41 X 41 | X 51 X 51 | X 61 X 61 | ?? |
步骤41、分别计算各富水性等级的发生概率,Step 41: Calculate the probability of occurrence of each water-rich level,
上式(7)中,In the above formula (7),
Y
i——富水性等级,
Y i - water-rich grade,
S
i——富水性等级为Y
i的样本个数,
S i -- the number of samples with a water-rich grade of Y i ,
S——模型中训练样本的个数,S——the number of training samples in the model,
S
c——所有富水性等级个数,
S c - the number of all water-rich grades,
P(Y
i)——富水性等级Y
i的概率;
P(Y i ) - probability of water-rich grade Y i ;
步骤42、分别计算每个属性等级的发生概率,Step 42: Calculate the probability of occurrence of each attribute level separately.
上式(8)中,In the above formula (8),
k——第k个属性,k - the kth attribute,
C
k——第k个属性的影响指标,
C k - the impact indicator of the kth attribute,
X
k——第k个属性C
k的属性级别,
X k - the attribute level of the kth attribute C k ,
S
ik——第k个属性C
k的属性级别为X
k且对应的富水性等级为Y
i的训练样本个数,
S ik - the k-th attribute C k has an attribute level of X k and the corresponding number of training samples with a water-rich level of Y i ,
S
k——训练样本中X
k的富水性等级个数,
S k —— the number of water-rich grades of X k in the training sample,
P(C
k=X
k|Y
i)——富水性等级Y
i中属性值为X
k的概率;
P(C k =X k |Y i )——the probability that the attribute value is X k in the water-rich level Y i ;
步骤43、利用公式(9)得出待测取样处的待测样本的归属富水性等级,Step 43, using formula (9), obtaining the attribution water enrichment level of the sample to be tested at the sampling site to be tested,
上式(9)中,In the above formula (9),
X——待测取样处的待测样本,X——the sample to be tested at the sampling site to be tested,
Y(X)——待测取样处的待测样本X的归属富水性等级。Y(X)——the water enrichment level of the sample X to be tested at the sampling site to be tested.
对上述步骤41、步骤42和步骤43进行举例阐述:The above steps 41, 42 and 43 are illustrated as examples:
根据表6和贝叶斯公式预测待测样本V的煤层顶板砂岩富水性等级,在此以表6中第24组待测样本为例,各属性级别为(X
11,X
23,X
31,X
43,X
51,X
63),进行煤层顶板砂岩富水性等级预测, 具体步骤如下所示:
According to Table 6 and Bayesian formula, the water-rich grade of the coal seam roof sandstone of the sample V to be tested is predicted. Here, the 24th group of samples to be tested in Table 6 is taken as an example, and each attribute level is (X 11 , X 23 , X 31 , X 43 , X 51 , X 63 ), the prediction of the water-rich grade of the coal seam roof sandstone, the specific steps are as follows:
Step1、分别计算煤层顶板各富水性等级的发生概率:Step1: Calculate the probability of occurrence of each water-rich grade of the coal seam roof:
Step2、分别计算每个属性等级的发生概率:Step 2, respectively calculate the probability of occurrence of each attribute level:
第1个属性砂岩厚度(C
1)属性级别为X
11的发生概率为:
The probability of occurrence of the first attribute sandstone thickness (C 1 ) attribute level of X 11 is:
第2个属性泥岩厚度(C
2)属性级别为X
23的发生概率为:
The probability of occurrence of the second attribute mudstone thickness (C 2 ) attribute level of X 23 is:
第3个属性岩芯采取率(C
3)属性级别为X
31的发生概率为:
The probability of occurrence of the third attribute core take-off rate (C 3 ) attribute level is X 31 is:
第4个属性浆液漏失量(C
4)属性级别为X
43的发生概率为:
The probability of occurrence of the fourth attribute slurry loss (C 4 ) attribute level of X 43 is:
第5个属性断裂密度(C
5)属性级别为X
43的发生概率为:
The probability of occurrence of the fifth attribute fracture density (C 5 ) attribute level of X 43 is:
第6个属性断裂强度(C
6)属性级别为X
43的发生概率为:
The probability of occurrence of the sixth attribute breaking strength (C 6 ) attribute level of X 43 is:
Step3、预测待测取样处的待测样本(第24组)的归属富水性等级:Step 3. Predict the water enrichment level of the sample to be tested (Group 24) at the sampling site to be tested:
由上得知,
的值在三个结果中较大,所以,第24组待测样本的煤层顶板砂岩富水性等级为较弱。同理,计算出另外10组待测样本预测结果。
From the above, The value of the sandstone roof sandstone of the 24th group to be tested is relatively weak. Similarly, another 10 sets of samples to be tested are calculated.
综上所述,11组待测样本的预测结果如表7所示:In summary, the prediction results of 11 groups of samples to be tested are shown in Table 7:
表7 预测结果Table 7 Prediction results
将前面利用贝斯分类器预测模型的结果与实际情况的结果相比较,如表8所示:Compare the results of the previous prediction model using the bass classifier with the actual results, as shown in Table 8:
表8 预测结果与实际结果对比Table 8 Comparison of predicted results with actual results
通过分析上述实施例可知,11组待测样本仅有第30组预测错误,预测结果为较强,实际预测结果为中等。因此,预测模型对待测样本预测正确10个、错误1个,预测准确率为90.9%。由此可以得出,本实施例的煤层顶板砂岩富水性等级预测方法预测准确率很高,为煤矿的高效、安全施工提供保障,具有较高的推广价值。By analyzing the above embodiments, it can be seen that the 11 groups of samples to be tested only have the 30th group of prediction errors, the prediction result is strong, and the actual prediction result is medium. Therefore, the prediction model predicts 10 correct samples and 1 error, and the prediction accuracy is 90.9%. It can be concluded that the prediction method of the water-rich grade of the coal seam roof sandstone of the present embodiment has a high prediction accuracy, and provides guarantee for efficient and safe construction of the coal mine, and has high promotion value.
当然,上述说明并非是对本发明的限制,本发明也并不仅限于上述举例,本技术领域的技术人员在本发明的实质范围内所做出的变化、改型、添加或替换,也应属于本发明的保护范围。The above description is not intended to limit the present invention, and the present invention is not limited to the above examples, and variations, modifications, additions or substitutions made by those skilled in the art within the scope of the present invention should also belong to the present invention. The scope of protection of the invention.
Claims (4)
- 一种煤层顶板砂岩富水性等级预测方法,其特征在于,包括以下步骤:A method for predicting water-rich grade of coal seam roof sandstone, characterized in that the method comprises the following steps:步骤一、对井下煤层顶板砂岩的N个取样处进行现场取样,对每一取样处的煤层顶板砂岩分析得到其三个富水性等级和影响富水性等级的六个影响因素,三个富水性等级包括较弱、中等和较强,六个影响因素包括砂岩厚度C 1、泥岩厚度C 2、岩芯采取率C 3、浆液漏失量C 4、断裂密度C 5以及断裂强度C 6,将N个取样处的上述取样数据建立样本数据库,其中,每一取样处的煤层顶板砂岩的富水性等级需要在煤矿施工后进行确定; Step 1. Sampling the N sampling sites of the coal seam roof sandstone in the underground, and analyzing the coal seam roof sandstone at each sampling site to obtain three water-rich grades and six influencing factors affecting the water-rich grade, three water-rich grades. Including weak, medium and strong, the six influencing factors include sandstone thickness C 1 , mudstone thickness C 2 , core take-up rate C 3 , slurry loss C 4 , fracture density C 5 and breaking strength C 6 , N The sample data of the sampling site is used to establish a sample database, wherein the water-rich grade of the coal seam roof sandstone at each sampling location needs to be determined after the coal mine construction;步骤二、对样本数据库的数据进行预处理后,将每一取样处的取样数据的六个影响因素作为输入向量,三个富水性等级作为输出向量,建立基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型;Step 2: After preprocessing the data of the sample database, the six influencing factors of the sampled data at each sampling point are taken as input vectors, and the three water-rich levels are used as output vectors to establish a coal seam roof sandstone based on Bayesian classifier. Water-rich grade prediction model;步骤三、对基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型进行训练和检验;Step 3: Training and testing the prediction model of the water-rich grade of the coal seam roof sandstone based on Bayesian classifier;步骤四、对井下煤层顶板砂岩的待测取样处进行现场取样,对待测取样处的煤层顶板砂岩分析得到其影响富水性等级的六个影响因素,将其六个影响因素输入到训练、检验好的基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型中,利用贝叶斯公式法实现对待测取样处的煤层顶板砂岩富水性等级预测。Step 4: Sampling the sampling site of the sandstone roof of the underground coal seam, and analyzing the coal seam roof sandstone of the sampling site to obtain the six influencing factors affecting the water-rich grade, and inputting the six influencing factors into the training and testing. Based on the Bayesian classifier, the prediction model of the water-rich grade of coal seam roof sandstone is realized by Bayesian formula method to predict the water-rich grade of coal seam roof sandstone to be tested.
- 根据权利要求1所述的煤层顶板砂岩富水性等级预测方法,其特征在于,步骤二中,对样本数据库的数据进行预处理具体包括如下步骤:The method for predicting the water-rich grade of the coal seam roof sandstone according to claim 1, wherein in step 2, pre-processing the data of the sample database comprises the following steps:步骤21、归一化,按下式(1)计算得出,Step 21, normalization, calculated by following formula (1),上式(1)中,In the above formula (1),x ij——归一化前样本数据, x ij - normalized pre-sample data,s ij——归一化后样本数据, s ij - normalized sample data,min(x j)——原始样本数据中的最小值, Min(x j ) - the minimum value in the original sample data,max(x j)——原始样本数据中的最大值; Max(x j ) - the maximum value in the original sample data;步骤22、离散化,按下式(2)和(3)计算得出,Step 22, discretization, calculated by following equations (2) and (3),上式(2)中,In the above formula (2),z ij——离散化后的样本数据, z ij - discretized sample data,min(s j)——归一化后样本数据的最小值, Min(s j ) - the minimum value of the normalized sample data,max(s j)——归一化后样本数据的最大值, Max(s j ) - the maximum value of the normalized sample data,上式(3)中,In the above formula (3),Q——步长;Q - step size;步骤23、属性约简,根据粗糙集理论删除离散化后样本数据中重复的数据。Step 23. Attribute reduction, and the repeated data in the discretized sample data is deleted according to the rough set theory.
- 根据权利要求2所述的煤层顶板砂岩富水性等级预测方法,其特征在于,步骤三中,对基于贝叶斯分类器的煤层顶板砂岩富水性等级预测模型进行训练和检验具体包括如下步骤:The method for predicting water-rich grade of coal seam roof sandstone according to claim 2, wherein in step 3, training and testing the water-soil level prediction model of the coal seam roof sandstone based on the Bayesian classifier comprises the following steps:步骤31、敏感性,按下式(4)计算得出,Step 31, sensitivity, calculated by following formula (4),上式(4)中,In the above formula (4),Sensitivity——敏感性,Sensitivity - sensitivity,positive——已知正样本数目,Positive - the number of known positive samples,true_positive——相应于正确分类的正样本数目;True_positive - the number of positive samples corresponding to the correct classification;步骤32、专一性,按下式(5)计算得出,Step 32, specificity, calculated according to the following formula (5),上式(5)中,In the above formula (5),Specificity——专一性,Specificity - specificity,negative——已知负样本的数目,Negative - the number of known negative samples,true_negative——相应于正确分类的负样本的数目;True_negative - the number of negative samples corresponding to the correct classification;步骤33、精准度,按下式(6)计算得出,Step 33, accuracy, calculated by following equation (6),上式(6)中,In the above formula (6),Accuracy——精准度,Accuracy - precision,Sensitivity——敏感性,Sensitivity - sensitivity,positive——已知正样本数目,Positive - the number of known positive samples,negative——已知负样本的数目,Negative - the number of known negative samples,Specificity——专一性。Specificity - specificity.
- 根据权利要求3所述的煤层顶板砂岩富水性等级预测方法,其特征在于,步骤四中,利用贝叶斯公式法实现对待测取样处的煤层顶板砂岩富水性等级预测具体包括如下步骤:The method for predicting water-rich grade of coal seam roof sandstone according to claim 3, wherein in step 4, using the Bayesian formula method to predict the water-rich grade of the coal seam roof sandstone to be tested includes the following steps:步骤41、分别计算各富水性等级的发生概率,Step 41: Calculate the probability of occurrence of each water-rich level,上式(7)中,In the above formula (7),Y i——富水性等级, Y i - water-rich grade,S i——富水性等级为Y i的样本个数, S i -- the number of samples with a water-rich grade of Y i ,S——模型中训练样本的个数,S——the number of training samples in the model,S c——所有富水性等级个数, S c - the number of all water-rich grades,P(Y i)——富水性等级Y i的概率; P(Y i ) - probability of water-rich grade Y i ;步骤42、分别计算每个属性等级的发生概率,Step 42: Calculate the probability of occurrence of each attribute level separately.上式(8)中,In the above formula (8),k——第k个属性,k - the kth attribute,C k——第k个属性的影响指标, C k - the impact indicator of the kth attribute,X k——第k个属性C k的属性级别, X k - the attribute level of the kth attribute C k ,S ik——第k个属性C k的属性级别为X k且对应的富水性等级为Y i的训练样本个数, S ik - the k-th attribute C k has an attribute level of X k and the corresponding number of training samples with a water-rich level of Y i ,S k——训练样本中X k的富水性等级个数, S k —— the number of water-rich grades of X k in the training sample,P(C k=X k|Y i)——富水性等级Y i中属性值为X k的概率; P(C k =X k |Y i )——the probability that the attribute value is X k in the water-rich level Y i ;步骤43、利用公式(9)得出待测取样处的待测样本的归属富水性等级,Step 43, using formula (9), obtaining the attribution water enrichment level of the sample to be tested at the sampling site to be tested,上式(9)中,In the above formula (9),X——待测取样处的待测样本,X——the sample to be tested at the sampling site to be tested,Y(X)——待测取样处的待测样本X的归属富水性等级。Y(X)——the water enrichment level of the sample X to be tested at the sampling site to be tested.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711111513.X | 2017-11-13 | ||
CN201711111513.XA CN107942383A (en) | 2017-11-13 | 2017-11-13 | Roof sandstone watery grade prediction technique |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2019091134A1 true WO2019091134A1 (en) | 2019-05-16 |
Family
ID=61933826
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2018/095623 WO2019091134A1 (en) | 2017-11-13 | 2018-07-13 | Method for predicting water-rich levels of sandstone in coal seam roof |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN107942383A (en) |
WO (1) | WO2019091134A1 (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109034528A (en) * | 2018-06-14 | 2018-12-18 | 安徽理工大学 | A kind of roof sandstone watery evaluation method that correlation method can be opened up based on improved grey model |
CN109190292A (en) * | 2018-09-26 | 2019-01-11 | 山东科技大学 | Aquifer water well prediction technique based on well-log information |
CN109254138B (en) * | 2018-10-22 | 2021-10-22 | 陕西煤业化工技术研究院有限责任公司 | Water-rich evaluation method based on sandstone microcosmic characteristics |
CN116955966B (en) * | 2023-09-20 | 2023-12-19 | 山东科技大学 | Method for judging water-rich grade of mine roof |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2005052639A1 (en) * | 2003-10-28 | 2005-06-09 | Western Geco, Llc | A method for estimating porosity and saturation in a subsurface reservoir |
CN102194056A (en) * | 2011-05-05 | 2011-09-21 | 中国矿业大学(北京) | BN-GIS (Bayesian Network-Geographic Information System) method for evaluating and predicting water inrush danger of coal-seam roof and floor |
CN105005712A (en) * | 2015-08-21 | 2015-10-28 | 山东科技大学 | Method for evaluating water yield property of limestone aquifer |
CN106250681A (en) * | 2016-07-26 | 2016-12-21 | 山东科技大学 | A kind of Forecasting Methodology for immediate roof stability |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106528707A (en) * | 2016-10-26 | 2017-03-22 | 中国矿业大学 | Coal seam roof sandstone aquifer water-abundance evaluation method |
-
2017
- 2017-11-13 CN CN201711111513.XA patent/CN107942383A/en active Pending
-
2018
- 2018-07-13 WO PCT/CN2018/095623 patent/WO2019091134A1/en active Application Filing
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2005052639A1 (en) * | 2003-10-28 | 2005-06-09 | Western Geco, Llc | A method for estimating porosity and saturation in a subsurface reservoir |
CN102194056A (en) * | 2011-05-05 | 2011-09-21 | 中国矿业大学(北京) | BN-GIS (Bayesian Network-Geographic Information System) method for evaluating and predicting water inrush danger of coal-seam roof and floor |
CN105005712A (en) * | 2015-08-21 | 2015-10-28 | 山东科技大学 | Method for evaluating water yield property of limestone aquifer |
CN106250681A (en) * | 2016-07-26 | 2016-12-21 | 山东科技大学 | A kind of Forecasting Methodology for immediate roof stability |
Non-Patent Citations (1)
Title |
---|
LIU, DEIMIN ET AL.: "Study on Water Enrichment Prediction of Coal Roof Sandstone Aquifer Based on PNN", COAL TECHNOLOGY, vol. 33, no. 09, 10 September 2014 (2014-09-10), pages 336 - 338, ISSN: 1008-8725 * |
Also Published As
Publication number | Publication date |
---|---|
CN107942383A (en) | 2018-04-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2019091134A1 (en) | Method for predicting water-rich levels of sandstone in coal seam roof | |
CN108009716B (en) | Multi-level evaluation method for influence factors of horizontal well volume fracturing effect | |
CN109241627B (en) | Probability grading dynamic support method and device for automatically designing support scheme | |
CN107387051B (en) | Repeated fracturing well selection method for multi-stage fractured horizontal well with low-permeability heterogeneous oil reservoir | |
WO2019042483A2 (en) | Tbm boring rock state real-time sensing system and method | |
CN109740119B (en) | Rapid estimation method for uniaxial compressive strength of surrounding rock of TBM tunneling tunnel | |
CN104217124A (en) | TBM (Tunnel Boring Machine) construction surrounding rock classification method depending on engineering sample data | |
CN103729798B (en) | Coal mine safety evaluation system based on improved k-means clustering | |
Li et al. | Risk assessment of floor water inrush using entropy weight and variation coefficient model | |
Yang et al. | The fuzzy comprehensive evaluation of water and sand inrush risk during underground mining | |
CN107578152B (en) | Evaluation method of multi-branch horizontal well gas extraction scheme | |
Qu et al. | Multi-model fusion for assessing risk of inrush of limestone karst water through the mine floor | |
CN105005712A (en) | Method for evaluating water yield property of limestone aquifer | |
CN105043905A (en) | Method for determining long-term strength parameter of rock based on steady-flow variable-rate tangent line | |
CN108280289B (en) | Rock burst danger level prediction method based on local weighted C4.5 algorithm | |
Zhang et al. | Geological Type Recognition by Machine Learning on In‐Situ Data of EPB Tunnel Boring Machines | |
CN102323998B (en) | Method for checking wind-friction resistance value of mine roadway | |
Daohong et al. | Application of GA-SVM in classification of surrounding rock based on model reliability examination | |
CN113962003A (en) | Tunnel surrounding rock large deformation assessment method | |
CN113640380A (en) | Multi-stage classification method and system for rail damage detection | |
CN107605474B (en) | Method and device for predicting gas formation yield while drilling | |
CN104463686A (en) | Method for identifying shale gas reservoir while drilling by utilizing discriminant analysis method | |
Fei et al. | Projection pursuit model based on particle swarm optimization for rock burst prediction | |
CN111476406B (en) | Oil-water well casing damage early warning method and device and storage medium | |
CN112329255A (en) | Rock burst prediction method based on tendency degree and uncertain measure |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 18875649 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 18875649 Country of ref document: EP Kind code of ref document: A1 |