CN111274736A - Water flowing fractured zone prediction method based on supervised learning neural network algorithm - Google Patents
Water flowing fractured zone prediction method based on supervised learning neural network algorithm Download PDFInfo
- Publication number
- CN111274736A CN111274736A CN202010068784.7A CN202010068784A CN111274736A CN 111274736 A CN111274736 A CN 111274736A CN 202010068784 A CN202010068784 A CN 202010068784A CN 111274736 A CN111274736 A CN 111274736A
- Authority
- CN
- China
- Prior art keywords
- water flowing
- fractured zone
- flowing fractured
- height
- prediction model
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Animal Husbandry (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Agronomy & Crop Science (AREA)
- Development Economics (AREA)
- Marine Sciences & Fisheries (AREA)
- Mining & Mineral Resources (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a water flowing fractured zone prediction method based on a mentor learning neural network algorithm. The steps are S1: collecting influence factor indexes of the height of the mine water flowing fractured zone and the corresponding height of the water flowing fractured zone to form a sample data set; s2: normalizing the collected influence factor data; s3: dividing the normalized data set into a training sample and a test sample; s4: selecting training sample data, and establishing a water flowing fractured zone height prediction model by using a MATLAB RBF neural network tool; s5: substituting the test sample data into the water flowing fractured zone height prediction model to obtain a corresponding water flowing fractured zone height prediction value; s6: and calculating the relative error and the decision coefficient of the predicted value and the actual value of the test sample, judging whether the prediction model is effective, if the prediction model is ineffective, reselecting the influence factor index, and repeating the steps S1-S6 until an effective water flowing fractured zone height prediction model is found.
Description
Technical Field
The invention relates to a water flowing fractured zone prediction method based on a supervised learning neural network algorithm, and belongs to the technical field of coal resource safety exploitation.
Background
The method has the advantages that the height of the water-flowing fractured zone is predicted with high precision, and mine flood is effectively prevented and controlled, so that the method becomes a key problem concerned by scholars at home and abroad. The prediction of the height of the water flowing fractured zone (referred to as 'height guiding' for short) is a key parameter for determining whether the underground coal mining can be safely carried out in the water body. However, due to the characteristics of complexity, diversity, ambiguity, uncertainty and the like of rock mass media, the height prediction of the water-flowing fractured zone in China still stays at the stage of using by combining experience and theory for many years, the experience formula provided by the coal pillar setting and coal pressing mining procedures (called as 'three procedures' for short) of buildings, water bodies, railways and main roadways and the experience formula provided by the geological exploration specification of hydrological and geological engineering of mine areas (GB12719-91) have certain difference with the height of the water-flowing fractured zone actually measured in the existing mine, and the requirement of efficient and safe production of the mine cannot be completely met. Therefore, the method for predicting the height of the water flowing fractured zone more effectively and accurately has important practical significance.
At present, methods for predicting the height of a water flowing fractured zone of a mine can be roughly divided into four types: firstly, field measurement methods (including borehole imaging, acoustic detection, etc.); second, engineering analogy and empirical methods; thirdly, simulating an experimental method by using physically similar materials; fourthly, a computer software (algorithm) assists a simulation prediction method; the methods have the defects of high prediction cost, complex operation, large error and the like.
Chinese patent CN103544548A discloses a method for predicting the height of a mine water flowing fractured zone, Chinese patent CN103778480A discloses a method for predicting the height of a fractured zone based on sensitivity analysis, Chinese patent CN104200292A discloses a method for predicting the height of a water flowing fractured zone, Chinese patent CN104732304A discloses a method for predicting the height of a water flowing fractured zone based on a gray artificial neural network combined model, the above patent applications respectively provide a method for predicting the height of a water flowing fractured zone, but the accuracy of numerical values predicted by each method is closely related to parameter selection and data characteristics, the optimization problem of parameters is not well solved at present, meanwhile, correlation exists among multiple indexes of data, and the above methods cannot be fully considered, so that the prediction precision is limited.
Disclosure of Invention
The invention aims to provide a water-flowing fractured zone prediction method based on a neural network algorithm learned by a mentor.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a method for predicting a water flowing fractured zone based on a supervised learning neural network algorithm is characterized by comprising the following steps: determining factor indexes influencing the height of a water flowing fractured zone by combining a specific coal mining method of a coal mine and the structure and mechanical characteristics of an overlying rock stratum of the coal seam, and collecting sample data under each factor index which is relatively perfect in arrangement; processing the collected sample data by a normalization method, and eliminating the influence of each index dimension on the prediction model; in order to test the accuracy of the model, the normalized sample data is distributed into a training sample and a test sample in advance; selecting all training sample data, obtaining the relation between each factor index and the height of the water flowing fractured zone by using a MATLABRBF neural network tool, and establishing a water flowing fractured zone height prediction model; substituting the test sample data into the water flowing fractured zone height prediction model to obtain the predicted value of the water flowing fractured zone height corresponding to each test data; comparing the predicted value and the actual value of the test sample, and judging whether the prediction model is effective or not through the relative error of data and a decision coefficient; if the water flowing fractured zone height prediction model is effective, the water flowing fractured zone height can be accurately predicted by the established water flowing fractured zone height prediction model; and if the prediction model is invalid, reselecting the influence factor index, and repeating the steps until a high prediction accuracy water flowing fractured zone height prediction model is found.
The method specifically comprises the following steps:
s1: collecting data: determining factor indexes influencing the height of the water flowing fractured zone by combining a specific coal mining method of a coal mine and the structure and mechanical characteristics of an overlying rock layer of the coal seam, and collecting influence factor index data of the height of the water flowing fractured zone of a single mine or a plurality of mines and the corresponding height of the water flowing fractured zone to form a sample data set;
s2: data normalization processing: processing the collected sample data set by adopting a normalization method, and eliminating the influence of each index dimension on a prediction model;
s3: data classification: randomly distributing the normalized sample data into training samples and inspection samples, wherein the training sample data accounts for 2/3-9/10 of all the data, and the rest are inspection samples;
s4: establishing a prediction model: selecting all training sample data, utilizing a MATLABRBF neural network tool to program the training sample data to carry out training with instructor learning, obtaining the relation between each factor index and the height of the water flowing fractured zone, and establishing a water flowing fractured zone height prediction model;
s5: and (3) calculating a predicted value: substituting the test sample data into the water flowing fractured zone height prediction model to obtain the predicted value of the water flowing fractured zone height corresponding to each test data;
s6: evaluation and analysis: comparing the predicted value and the actual value of the test sample, and calculating the relative error and the decision coefficient between the predicted value and the actual value so as to judge whether the prediction model is effective or not; if the relative error between the predicted value and the actual value is less than or equal to 5% and the determination coefficient is more than or equal to 95% effective, the model is effective, and the established water flowing fractured zone height prediction model can accurately predict the height of the water flowing fractured zone; and if the relative error between the predicted value and the actual value is more than 5% or the decision coefficient is less than 95%, the model is invalid, the established water flowing fractured zone height prediction model cannot accurately predict the height of the water flowing fractured zone, the influence factor indexes are required to be selected again, and the steps from S1 to S6 are repeated until an effective water flowing fractured zone height prediction model is found.
MATLAB is programmed as follows:
clc,clear
load ssgs. txt% save the original data in a plain text file ssgs. txt%
% normalized data
n＝size(ssgs,1)；
x＝ssgs(:,[1:4])；octane＝ssgs(:,5)；
vec ═ sum (x); % normalized column sum of 1
[M,D]＝size(x)；
B＝repmat(vec,M,1)；
score x./B; % training sample
% establishing RBF prediction model
P_train＝score(1:51,:)'；
T_train＝octane(1:51,:)'；
P _ test ═ score (52: end:)'; % test sample
T_test＝octane(52:end,:)'；
N＝size(P_test,2)；
% creation of a network
net＝newrbe(P_train,T_train,0.287)；
% simulation test
T_sim_rbf＝sim(net,P_test)；
% Performance evaluation
% relative error
error_rbf＝abs(T_sim_rbf-T_test)./T_test；
% coefficient of determination R2
R2_rbf＝(N*sum(T_sim_rbf.*T_test)-sum(T_sim_rbf)*sum(T_test))^2/((N*sum((T_sim_rbf).^2)-(sum(T_sim_rbf))^2)*(N*sum((T_test).^2)-(sum(T_test))^2))；
Comparison of% results
result_rbf＝[T_test'T_sim_rbf'error_rbf']
figure
plot(1:N,T_test,'k:*',1:N,T_sim_rbf,'k-.^')；
Legend ('actual value', 'RBF predicted value')
xlabel ('test sample')
ylabel ('Water-channelling height of Water-carrying fissure')
Testing the prediction result' of the height of the water-flowing fractured zone of the sample; [ ' R ^2 ═ num2str (R2_ RBF) ' (RBF) ' ] };
title(string)
as an improvement on the technical scheme, the influence factor indexes of the water flowing fractured zone in the step S1 include mining height, hard rock lithology proportionality coefficient, working face slant length, mining depth and propulsion speed.
As an improvement to the above technical solution, the normalization processing formula adopted in step S2 is:
in formula (II), X'ijIs normalized data, X'ijThe value range is between (0, 1); xijFor the collected original sample data, i represents the ith group of sample data, j represents the jth influence factor index, and n represents the number of the collected sample groups.
As an improvement to the above technical solution, in step S4, training for leading a teacher to learn is to change an average relative error of data by adjusting a spread value (a spread value range is often [0,1]) in a neural network to characterize a complex relationship between data input and output, so as to automatically correct the model by using the average relative error of data, and when the average relative error takes a minimum value, the stability of the prediction model is the best.
As an improvement to the above technical solution, the two evaluation indexes adopted in step S6 are relative errors EiAnd determining the coefficient R2The calculation formulas are respectively as follows:
wherein the content of the first and second substances,
is the predicted value of the ith group of samples; y isi(i ═ 1,2, …, n) is the actual value of the ith set of samples; n is the number of groups collected for the sample.
The invention has the beneficial effects that:
in the process of establishing the water flowing fractured zone height prediction model, the data normalization method is used to eliminate the influence of dimensions on the prediction result, and the RBF neural network is used for regression prediction, so that the traditional neural network model can be improved, and the model is corrected by using the error of the model, so that the water flowing fractured zone height is predicted better, and the accuracy is higher. The method has the advantages of high calculation speed and small relative error, the feasibility, the reliability and the stability of the method are verified in practice, and the method has great theoretical significance and application value for predicting the height of the mine water flowing fractured zone.
Drawings
FIG. 1 is a flow chart of the predictive model steps of the present invention.
FIG. 2 is a comparison graph of actual values and predicted values of test samples in the examples.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, which are provided so that the disclosure will be fully clear and will fully convey the scope of the invention to those skilled in the art. It is to be understood that the embodiments described are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived from the embodiments of the present invention by a person skilled in the art without any creative effort, should be included in the protection scope of the present invention.
Example one
As shown in fig. 1, in the present embodiment, a method for predicting a water-flowing fractured zone based on a neural network algorithm learned by a mentor includes: determining factor indexes influencing the height of a water flowing fractured zone by combining a specific coal mining method of a coal mine and the structure and mechanical characteristics of an overlying rock stratum of the coal seam, and collecting sample data under each factor index which is relatively perfect in arrangement; processing the collected sample data by a normalization method, and eliminating the influence of each index dimension on the prediction model; distributing the normalized sample data into a training sample and a test sample; selecting all training sample data, obtaining the relation between each factor index and the height of the water flowing fractured zone by using a MATLABRBF neural network tool, and establishing a water flowing fractured zone height prediction model; substituting the test sample data into the water flowing fractured zone height prediction model to obtain the predicted value of the water flowing fractured zone height corresponding to each test data; comparing the predicted value and the actual value of the test sample, and judging whether the prediction model is effective or not through the relative error of data and a decision coefficient; if the water flowing fractured zone height prediction model is effective, the water flowing fractured zone height can be accurately predicted by the established water flowing fractured zone height prediction model; and if the prediction model is invalid, reselecting the influence factor index, and repeating the steps until a high prediction accuracy water flowing fractured zone height prediction model is found.
As shown in fig. 1, a method for predicting a water-flowing fractured zone based on a neural network algorithm learned by a mentor includes the following steps:
s1: collecting data: the method comprises the steps of determining factor indexes influencing the height of the water flowing fractured zone by combining a specific coal mining method of a coal mine and the structure and mechanical characteristics of an overlying rock layer of the coal mine, collecting influence factor index data of the height of the water flowing fractured zone of a single mine or a plurality of mines and the corresponding height of the water flowing fractured zone to form a sample data set, and collecting 56 groups of sample data in the table shown in table 1.
TABLE 1 sample data table of height of water flowing fractured zone of different mines across the country
S2: data normalization processing: processing 56 groups of collected sample data by adopting a normalization method, and eliminating the influence of each index dimension on a prediction model; the normalization processing formula is as follows:
in formula (II), X'ijIs normalized data, X'ijThe value range is between (0, 1); xijFor the collected original sample data, i represents the ith group of sample data, j represents the jth influence factor index, and n represents the number of the collected sample groups.
The data in table 1 are substituted into the above formula for normalization processing to obtain a normalized data table, see table 2.
Table 2 table 1 data after normalization
S3: data classification: and randomly distributing the normalized sample data into training samples and testing samples. If the training samples are distributed more, the trained model is closer to the information containing most samples, but the evaluation result is not stable and accurate enough due to less test samples; if the distribution of the test samples is more, the training samples are fewer, the model obtained by training cannot contain most information of the samples, and the evaluation result obtained by substituting the test samples into the model may have a larger difference from the actual value, so that the fidelity of the evaluation result is reduced. The common method is that training sample data accounts for 2/3-9/10 of all data, and the rest are inspection samples; therefore, in this example, the 1 st to 51 th groups of data among 56 groups of samples are used as training samples, and the remaining 52 nd to 56 th groups of data are used as test samples.
S4: establishing a prediction model: selecting 1 st to 51 th groups of training sample data, utilizing a MATLABRBF neural network tool to program the training sample data to carry out training with instructor learning, obtaining the relation between each factor index and the height of the water flowing fractured zone, and establishing a water flowing fractured zone height prediction model. Training for leading teacher learning is to represent the complex relation between data input and output by regulating the average relative error of the spread value (the spread value range is often [0,1]) in the neural network to change the average relative error of data, so as to realize automatic correction of the model by using the average relative error of data, and when the average relative error is the minimum value, the stability of the prediction model is the best. In general, the larger the spread value, the smoother the fit of the function. However, an excessively large spread would require a very large number of neurons to accommodate the rapid changes in the function; conversely, if the spread value is too small, it means that many neurons are required to adapt to slow changes in the function, resulting in poor network performance. In the program, the step length is set to be 0.001 from 0 to 1 through traversal search calculation, the minimum value is calculated for the average relative error of the test sample data, and when the spread value is 0.287, the minimum value is calculated for the average relative error of the test sample data. And continuously adjusting the value of the spread, and finally determining that the spread is 0.287, wherein the average relative error of the model is minimum, the decision coefficient is maximum, namely the predicted value is closest to the actual value, the stability of the model is best, and a relatively reliable water flowing fractured zone height prediction model is obtained.
S5: and (3) calculating a predicted value: and substituting the 52 th to 56 th groups of test sample data into the water flowing fractured zone height prediction model to obtain the predicted value of the water flowing fractured zone height corresponding to each test data, which is shown in table 3.
TABLE 3 actual, predicted and relative error of test samples
| Serial number | Actual value | Prediction value | Relative errorDifference (D) |
| 1 | 44.34 | 43.9138 | 0.96％ |
| 2 | 62.17 | 62.3943 | 1.02％ |
| 3 | 40.14 | 41.3162 | 1.96％ |
| 4 | 63.60 | 61.3162 | 0.84％ |
| 5 | 42.99 | 44.0818 | 0.10％ |
S6: evaluation and analysis: comparing the predicted value and the actual value of the test sample in the table 3, and calculating the relative error and the decision coefficient between the predicted value and the actual value so as to judge whether the prediction model is effective or not; if the relative error between the predicted value and the actual value is less than or equal to 5% and the determination coefficient is more than or equal to 95% effective, the model is effective, and the established water flowing fractured zone height prediction model can accurately predict the height of the water flowing fractured zone; and if the relative error between the predicted value and the actual value is more than 5% or the decision coefficient is less than 95%, the model is invalid, the established water flowing fractured zone height prediction model cannot accurately predict the height of the water flowing fractured zone, the influence factor indexes are required to be selected again, and the steps from S1 to S6 are repeated until an effective water flowing fractured zone height prediction model is found. In this example, a comparison graph of the actual value and the predicted value is drawn using MATLAB, and a decision coefficient is calculated, as shown in fig. 2.
Through comparison, the relative error between the actual value and the predicted value is less than 5 percent, and R is2And if the prediction model is more than 99 percent, the prediction model has very high precision and meets the requirement of high-precision prediction.
MATLAB is programmed as follows:
clc,clear
load ssgs. txt% save the original data in a plain text file ssgs. txt%
% normalized data
n＝size(ssgs,1)；
x＝ssgs(:,[1:4])；octane＝ssgs(:,5)；
vec ═ sum (x); % normalized column sum of 1
[M,D]＝size(x)；
B＝repmat(vec,M,1)；
score x./B; % training sample
% establishing RBF prediction model
P_train＝score(1:51,:)'；
T_train＝octane(1:51,:)'；
P _ test ═ score (52: end:)'; % test sample
T_test＝octane(52:end,:)'；
N＝size(P_test,2)；
% creation of a network
net＝newrbe(P_train,T_train,0.287)；
% simulation test
T_sim_rbf＝sim(net,P_test)；
% Performance evaluation
% relative error
error_rbf＝abs(T_sim_rbf-T_test)./T_test；
% coefficient of determination R2
R2_rbf＝(N*sum(T_sim_rbf.*T_test)-sum(T_sim_rbf)*sum(T_test))^2/((N*sum((T_sim_rbf).^2)-(sum(T_sim_rbf))^2)*(N*sum((T_test).^2)-(sum(T_test))^2))；
Comparison of% results
result_rbf＝[T_test'T_sim_rbf'error_rbf']
figure
plot(1:N,T_test,'k:*',1:N,T_sim_rbf,'k-.^')；
Legend ('actual value', 'RBF predicted value')
xlabel ('test sample')
ylabel ('Water-channelling height of Water-carrying fissure')
Testing the prediction result' of the height of the water-flowing fractured zone of the sample; [ ' R ^2 ═ num2str (R2_ RBF) ' (RBF) ' ] }; title (string).
Claims (6)
1. A method for predicting a water flowing fractured zone based on a supervised learning neural network algorithm is characterized by comprising the following steps: determining factor indexes influencing the height of a water flowing fractured zone by combining a specific coal mining method of a coal mine and the structure and mechanical characteristics of an overlying rock stratum of the coal seam, and collecting sample data under each factor index which is relatively perfect in arrangement; processing the collected sample data by a normalization method, and eliminating the influence of each index dimension on the prediction model; distributing the normalized sample data into a training sample and a test sample; selecting all training sample data, obtaining the relation between each factor index and the height of the water flowing fractured zone by using a MATLAB RBF neural network tool, and establishing a water flowing fractured zone height prediction model; substituting the test sample data into the water flowing fractured zone height prediction model to obtain the predicted value of the water flowing fractured zone height corresponding to each test data; comparing the predicted value and the actual value of the test sample, and judging whether the prediction model is effective or not through the relative error of data and a decision coefficient; if the water flowing fractured zone height prediction model is effective, the water flowing fractured zone height can be accurately predicted by the established water flowing fractured zone height prediction model; and if the prediction model is invalid, reselecting the influence factor index, and repeating the steps until a high prediction accuracy water flowing fractured zone height prediction model is found.
2. The method for predicting the water flowing fractured zone based on the mentoring neural network algorithm according to claim 1, wherein the method comprises the following steps: the method specifically comprises the following steps:
s1: collecting data: determining factor indexes influencing the height of the water flowing fractured zone by combining a specific coal mining method of a coal mine and the structure and mechanical characteristics of an overlying rock layer of the coal seam, and collecting influence factor index data of the height of the water flowing fractured zone of a single mine or a plurality of mines and the corresponding height of the water flowing fractured zone to form a sample data set;
s2: data normalization processing: processing the collected sample data set by adopting a normalization method, and eliminating the influence of each index dimension on a prediction model;
s3: data classification: distributing the normalized sample data into training samples and inspection samples, wherein the training sample data accounts for 2/3-9/10 of all data, and the rest are inspection samples;
s4: establishing a prediction model: selecting all training sample data, utilizing a MATLAB RBF neural network tool to program the training sample data to carry out training with instructor learning, obtaining the relation between each factor index and the height of the water flowing fractured zone, and establishing a water flowing fractured zone height prediction model;
s5: and (3) calculating a predicted value: substituting the test sample data into the water flowing fractured zone height prediction model to obtain the predicted value of the water flowing fractured zone height corresponding to each test data;
s6: evaluation and analysis: comparing the predicted value and the actual value of the test sample, and calculating the relative error and the decision coefficient between the predicted value and the actual value so as to judge whether the prediction model is effective or not; if the relative error between the predicted value and the actual value is less than or equal to 5% and the determination coefficient is more than or equal to 95% effective, the model is effective, and the established water flowing fractured zone height prediction model can accurately predict the height of the water flowing fractured zone; and if the relative error between the predicted value and the actual value is more than 5% or the decision coefficient is less than 95%, the model is invalid, the established water flowing fractured zone height prediction model cannot accurately predict the height of the water flowing fractured zone, the influence factor indexes are required to be selected again, and the steps S1-S6 are repeated until an effective water flowing fractured zone height prediction model is found.
3. The method for predicting the water flowing fractured zone based on the mentoring neural network algorithm according to claim 2, wherein the method comprises the following steps: and the influence factor indexes of the water flowing fractured zone in the step S1 comprise mining height, hard rock lithology proportionality coefficient, working face slant length, mining depth and propulsion speed.
4. The method for predicting the water flowing fractured zone based on the mentoring neural network algorithm according to claim 2, wherein the method comprises the following steps: the normalization processing formula adopted in step S2 is:
in the formula (I), the compound is shown in the specification,
in order to normalize the processed data, the data is normalized,
the value range is between (0, 1);
in order to collect the original sample data of the sample,
is shown as
The data of the group of samples is,
is shown as
An influence factor index, n, represents the number of sets of samples collected.
5. The method for predicting the water flowing fractured zone based on the mentoring neural network algorithm according to claim 2, wherein the method comprises the following steps: step S4, training for instructor learning is to represent the complex relation between data input and output by adjusting the average relative error of the spread value change data in the neural network, so as to realize automatic correction of the model by using the average relative error of the data, and when the average relative error is the minimum value, the stability of the prediction model is the best; the span value interval is usually [0,1 ].
6. The method for predicting the water flowing fractured zone based on the mentoring neural network algorithm according to claim 2, wherein the method comprises the following steps: the two evaluation indexes adopted in step S6 are relative errors
And determining the coefficient
The calculation formulas are respectively as follows:
wherein the content of the first and second substances,
is as followsiA prediction value of the group sample;
is as followsiActual values of the group samples;nthe number of groups collected for the sample.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010068784.7A CN111274736A (en) | 2020-01-21 | 2020-01-21 | Water flowing fractured zone prediction method based on supervised learning neural network algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010068784.7A CN111274736A (en) | 2020-01-21 | 2020-01-21 | Water flowing fractured zone prediction method based on supervised learning neural network algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111274736A true CN111274736A (en) | 2020-06-12 |
Family
ID=70999059
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010068784.7A Pending CN111274736A (en) | 2020-01-21 | 2020-01-21 | Water flowing fractured zone prediction method based on supervised learning neural network algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111274736A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112948924A (en) * | 2021-02-06 | 2021-06-11 | 合肥工业大学 | Near unconsolidated formation mining water flowing fractured zone height determination method based on overlying strata structure |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102221651A (en) * | 2011-03-11 | 2011-10-19 | 太原理工大学 | Fault on-line diagnosis and early warning method of flameproof dry-type transformer for mine |
| CN103544548A (en) * | 2013-10-31 | 2014-01-29 | 辽宁工程技术大学 | Method for predicting height of mine water flowing fractured zone |
| CN103778480A (en) * | 2014-01-14 | 2014-05-07 | 辽宁工程技术大学 | Fissure-zone height prediction method based on sensitivity analysis |
| CN104200292A (en) * | 2014-09-28 | 2014-12-10 | 山东科技大学 | Forecasting method for height of water-flowing fractured zone |
| CN104732304A (en) * | 2015-04-15 | 2015-06-24 | 河南理工大学 | Grey artificial neural network combination model based method for predicting height of water-flowing fractured zone |
| CN105719025A (en) * | 2016-01-26 | 2016-06-29 | 华北电力大学（保定） | Prediction method for corrosion rate of Q235 galvanized steel grounding net of transformer station |
| CN108694023A (en) * | 2018-02-22 | 2018-10-23 | 长安大学 | A kind of test method of marshal piece stability and flow valuve |
| CN110188922A (en) * | 2019-05-05 | 2019-08-30 | 中国长江电力股份有限公司 | A kind of long-term Runoff Forecast method in the RBF neural based on runoff mechanism |
-
2020
- 2020-01-21 CN CN202010068784.7A patent/CN111274736A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102221651A (en) * | 2011-03-11 | 2011-10-19 | 太原理工大学 | Fault on-line diagnosis and early warning method of flameproof dry-type transformer for mine |
| CN103544548A (en) * | 2013-10-31 | 2014-01-29 | 辽宁工程技术大学 | Method for predicting height of mine water flowing fractured zone |
| CN103778480A (en) * | 2014-01-14 | 2014-05-07 | 辽宁工程技术大学 | Fissure-zone height prediction method based on sensitivity analysis |
| CN104200292A (en) * | 2014-09-28 | 2014-12-10 | 山东科技大学 | Forecasting method for height of water-flowing fractured zone |
| CN104732304A (en) * | 2015-04-15 | 2015-06-24 | 河南理工大学 | Grey artificial neural network combination model based method for predicting height of water-flowing fractured zone |
| CN105719025A (en) * | 2016-01-26 | 2016-06-29 | 华北电力大学（保定） | Prediction method for corrosion rate of Q235 galvanized steel grounding net of transformer station |
| CN108694023A (en) * | 2018-02-22 | 2018-10-23 | 长安大学 | A kind of test method of marshal piece stability and flow valuve |
| CN110188922A (en) * | 2019-05-05 | 2019-08-30 | 中国长江电力股份有限公司 | A kind of long-term Runoff Forecast method in the RBF neural based on runoff mechanism |
Non-Patent Citations (1)
| Title |
|---|
| 张新盈: "K-means和QGA优化RBF神经网络模型在导水裂缝带高度预测方面的应用", 中国矿业, vol. 27, no. 8, pages 164 - 167 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112948924A (en) * | 2021-02-06 | 2021-06-11 | 合肥工业大学 | Near unconsolidated formation mining water flowing fractured zone height determination method based on overlying strata structure |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104533400B (en) | Method for reconstructing logging curve | |
| CN108984886B (en) | Method for inverting dynamic connectivity among wells of multi-layer oil reservoir | |
| CN107122860B (en) | Rock burst danger level prediction method based on grid search and extreme learning machine | |
| CN110619353B (en) | Multi-scale logging curve automatic identification method based on deep learning | |
| CN109214026A (en) | A kind of shale gas horizontal well initial productivity prediction technique | |
| CN103617147A (en) | Method for identifying mine water-inrush source | |
| CN113570226A (en) | Method for evaluating occurrence probability grade of tunnel water inrush disaster in fault fracture zone | |
| CN106991509A (en) | Log Forecasting Methodology based on radial basis function neural network model | |
| CN112529341A (en) | Drilling well leakage probability prediction method based on naive Bayesian algorithm | |
| CN111767674A (en) | Well logging lithology identification method based on active domain adaptation | |
| CN114723095A (en) | Missing well logging curve prediction method and device | |
| CN108280289B (en) | Rock burst danger level prediction method based on local weighted C4.5 algorithm | |
| CN115481565A (en) | Earth pressure balance shield tunneling parameter prediction method based on LSTM and ant colony algorithm | |
| CN114357750A (en) | Goaf water filling state evaluation method | |
| CN107942383A (en) | Roof sandstone watery grade prediction technique | |
| CN116427915A (en) | Conventional logging curve crack density prediction method and system based on random forest | |
| CN117251802B (en) | Heterogeneous reservoir parameter prediction method and system based on transfer learning | |
| CN115577645A (en) | Construction method and prediction method of combustion and explosion fracturing fracture range prediction model | |
| CN111274736A (en) | Water flowing fractured zone prediction method based on supervised learning neural network algorithm | |
| CN117272841B (en) | Shale gas dessert prediction method based on hybrid neural network | |
| CN107818217B (en) | Seismic guiding method based on horizontal well exploration type visual analysis | |
| Ifrene et al. | New Insights in Fracture Porosity Estimation Using Machine Learning and Advanced Logging Tools | |
| CN113946790A (en) | Method, system, equipment and terminal for predicting height of water flowing fractured zone | |
| CN114117898A (en) | While-drilling gamma logging forward modeling method based on machine learning algorithm | |
| CN113554311A (en) | Method for evaluating engineering quality of Ordovician limestone water damage under ground directional hole grouting treatment push-coated body |
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 |