CN109446697B - ELM-based mine wind speed fault branch diagnosis method - Google Patents
ELM-based mine wind speed fault branch diagnosis method Download PDFInfo
- Publication number
- CN109446697B CN109446697B CN201811326508.5A CN201811326508A CN109446697B CN 109446697 B CN109446697 B CN 109446697B CN 201811326508 A CN201811326508 A CN 201811326508A CN 109446697 B CN109446697 B CN 109446697B
- Authority
- CN
- China
- Prior art keywords
- roadway
- fault
- elm
- samples
- wind
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Emergency Alarm Devices (AREA)
- Ventilation (AREA)
Abstract
The invention provides an ELM-based mine wind speed fault branch diagnosis method, and relates to the technical field of mine ventilation fault diagnosis. The method comprises the following specific steps: step 1: determining a fault roadway range according to a fault roadway determination method based on a wind resistance-wind flow change influence relation matrix; step 2: building an ELM fault branch diagnosis model; generating a sample set according to the roadway set W (j) to be detected, compiling a program and establishing an ELM fault branch diagnosis model; when an air speed sensor of the monitoring system alarms in an overrun mode, an ELM fault branch diagnosis model is started immediately, a branch number is used as an output value Y, and a monitoring value of the air speed sensor is used as an input value X; and 3, step 3: verifying a prediction result; the method is used for verifying the feasibility of the ELM fault branch diagnosis model for conducting mine wind speed fault branch diagnosis. The method avoids the process of training all branches, and greatly reduces the data scale and the calculated amount.
Description
Technical Field
The invention relates to the technical field of mine ventilation fault diagnosis, in particular to an ELM-based mine wind speed fault branch diagnosis method.
Background
Coal mine safety monitoring systems are installed in large, medium and small mines in China in succession, however, the application of the monitoring systems basically stays in the measurement of single data mainly comprising methane. The methane monitoring value is easy to judge the gas overrun position, the wind speed overrun value is difficult to judge the fault source of the ventilation system, and the current research is rare, and related research mainly refers to fault diagnosis of ventilation system equipment, including ventilators, elevators and the like. The mine ventilation system is a complex large system, and factors causing mine wind current change are many, and the factors are the change of tunnel wind resistance. Therefore, the failure of the underground roadway ventilation system can be attributed to the change of the roadway wind resistance. The underground roadway ventilation network is an organically-connected whole, the wind resistance of any roadway in the network changes, and the wind volume of other roadways also changes correspondingly. Thousands of tunnels in the ventilation network of the complex underground tunnel are possible, and the influence of the wind resistance change of each tunnel on the ventilation system cannot be specifically analyzed, namely, a wind speed sensor is arranged on each tunnel. And the change data given by the monitoring system can only reflect that the air volume of the roadway changes, and as for the change, whether the change is caused by the fault of the roadway, other roadway faults or fan faults, whether the fault source is a single point or multiple points can not be judged. Therefore, the failure of the roadway where the sensor is located can be judged without depending on the alarm sent by a certain node in the monitoring system; in the prior art, a method for determining the range of a fault roadway exists, manual detection is performed by a sensitivity sorting method, the calculated amount is large, and the specific branch with the fault cannot be determined.
An Extreme Learning Machine (ELM) is a training algorithm that has a fast training speed, obtains a globally optimal solution, and has good bloom performance. ELM randomly generates the connection weight between the input layer and the hidden layer and the threshold value of the hidden layer neuron, and only sets the number of the hidden layer neuron without adjustment in the training process to obtain the unique optimal solution.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a mine wind speed fault branch diagnosis method based on ELM aiming at the defects of the prior art, the method avoids the process of training all branches, and greatly reduces the data scale and the calculated amount.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows:
a mine wind speed fault branch diagnosis method based on ELM comprises the following steps:
step 1: determining a fault roadway range according to a fault roadway determination method based on a wind resistance-wind flow change influence relation matrix;
and 2, step: establishing an ELM fault branch diagnosis model;
and step 3: verifying a prediction result;
setting alpha wind speed sensor overrun alarms, counting s samples in training samples, and carrying out ELM fault branch diagnosis model training according to a branch sequence; the method comprises the following steps of (1) randomly mixing z samples of test samples, and carrying out ELM fault branch diagnosis model training on the sample sequence; comparing the obtained training sample prediction result and the prediction result of the test sample with a preset result; and verifying the feasibility of the ELM fault branch diagnosis model for performing the mine wind speed fault branch diagnosis.
The specific steps of step 1 are as follows:
step 1.1: constructing a wind resistance-wind flow change influence relation matrix of the roadway to be detected with the fault:
wherein i, j = {1,2,3, …, n }, and n is the total number of the roadways in the ventilation network of the roadway to be detected with the fault; a is a ij The value is obtained by the following method: sequentially changing the wind resistance of the ith roadway in the ventilation network of the roadway with the fault to be detected, keeping the wind resistance of other roadways unchanged, obtaining the wind volume values of all the roadways through network calculation, converting the wind volume values into wind speed values, comparing the wind speed values with the allowable wind speed value range of the corresponding roadway in coal mine safety regulations, and judging whether the wind speed value of the jth roadway exceeds the allowable range, if so, a ij =1, otherwise, a ij =0;
Step 1.2: determining an initial set W '(j) = { W' 1j ,w' 2j ,…,w' ij ,…,w' nj W 'in the formula' ij =i*a ij Deleting the element of 0 in W' (j) to obtain a roadway set W (j), and storing the roadway set W (j) in a roadway ventilation network fault roadway range library;
step 1.3: and obtaining a roadway set W (j) to be detected according to the roadway set corresponding to the roadway where the wind speed sensor is in the overrun alarm: when only one tunnel wind speed sensor gives an alarm in an overrun mode, acquiring a tunnel set W (j) corresponding to a tunnel j where the alarm sensor is located from a tunnel ventilation network fault tunnel range library as a tunnel set W (j) to be detected; when two or more roadway wind speed sensors alarm in an overrun mode, acquiring a roadway set corresponding to a roadway where each alarm sensor is located from a roadway ventilation network fault roadway range library, and calculating the intersection of all the roadway sets to be a to-be-detected roadway set W (j).
The specific steps of step 2 are as follows:
step 2.1: generating a sample set; the sample set comprises training samples and prediction samples;
generating a sample set without considering the total entrance and the total return of the tunnel in the ventilation network of the tunnel with the fault to be detected, so that the residual branch number is k;
training samples and prediction samples were generated as follows:
generating a training sample: increasing the resistance of k to generate samples, increasing the wind resistance by 1 step length, increasing by 6 steps, and obtaining s training samples through network calculation;
generation of prediction samples: increasing the resistance of k to generate samples, increasing the wind resistance by 1 step length by 3 steps, and resolving through a network to obtain z prediction samples;
step 2.2: building an ELM fault branch diagnosis model;
step 2.2.1: programming a program; utilizing elmtractin and elmpress functions; the Elmtrain function is used for creating and training the ELM, and the elmredict function is used for carrying out ELM simulation test; the calling formats are respectively as follows:
[IW,B,LW,TF,TYPE]=elmtrain(P,T,N,TF,TYPE)
Y=elmpredict(P,IW,B,LW,TF,TYPE)
wherein, P is an input matrix of a training set; t is an output matrix of the training set; n is the number of samples of the training set; TF is an activation function of hidden layer neurons, and the values of TF can be 'sig', 'sin', 'hardlim'; TYPE is the application TYPE of ELM, the values of the TYPE are 0 and 1,0 represent regression and fitting, and 1 represents classification; IW is the connection weight between the input layer and the hidden layer; b is the breadth of the hidden layer neuron; LW is the connection weight of the hidden layer and the output layer;
step 2.2.2: establishing an ELM fault branch diagnosis model, taking the air volume value of the tunnel where the air velocity sensor is positioned as an input value X, and changing the branch number e of the wind resistance i As an output value Y;
step 2.3: when an air speed sensor of the monitoring system gives an alarm in an overrun mode, an ELM fault branch diagnosis model is started immediately, a branch number is used as an output value Y, and a monitoring value of the air speed sensor is used as an input value X;
adopt the produced beneficial effect of above-mentioned technical scheme to lie in: according to the ELM-based mine wind speed fault branch diagnosis method, the fault roadway determination method and the extreme learning machine are combined to establish the ELM fault branch diagnosis model, so that the process of training all branches is avoided, the data scale and the calculated amount are greatly reduced, the specific fault position can be judged according to the alarm position of a wind speed sensor in an underground monitoring system, and the specific branch number can be obtained.
Drawings
FIG. 1 is a simplified ventilation network diagram of a first embodiment of a ventilation system for a mine;
FIG. 2 is a flowchart of a mine wind speed fault branch diagnosis method based on ELM according to an embodiment of the present invention;
FIG. 3 is a diagram of the predicted effect of training samples according to the first embodiment of the present invention;
FIG. 4 is a diagram illustrating predicted sample prediction effects provided by the first embodiment of the present invention;
fig. 5 is a diagram illustrating the performance evaluation effect provided by the first embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention, but are not intended to limit the scope of the invention.
The method of this example is as follows.
The first embodiment:
as shown in fig. 1, wind speed fault branch diagnosis is performed on a certain mine ventilation system;
an ELM-based mine wind speed fault branch diagnosis method is shown in FIG. 2 and comprises the following steps:
step 1: determining a fault roadway range according to a fault roadway determination method based on a wind resistance-wind flow change influence relation matrix;
step 1.1: constructing a wind resistance-wind flow change influence relation matrix of the roadway to be detected with the fault:
wherein i, j = {1,2,3, …, n }, and n is the total number of the roadways in the ventilation network of the roadway to be detected with the fault; a is ij Take values asThe following method comprises the following steps: sequentially changing the wind resistance of the ith roadway in the ventilation network of the roadway with the fault to be detected, keeping the wind resistance of other roadways unchanged, obtaining the wind volume values of all the roadways through network calculation, converting the wind volume values into wind speed values, comparing the wind speed values with the allowable wind speed value range of the corresponding roadway in coal mine safety regulations, and judging whether the wind speed value of the jth roadway exceeds the allowable range, if so, a ij =1, otherwise, a ij =0;
In the embodiment, the wind resistance values of all the roadways are sequentially changed in the mine ventilation simulation system, and the air volume change of the roadways is calculated to obtain a wind resistance-wind current change influence relation matrix:
step 1.2: determining an initial set W '(j) = { W' 1j ,w' 2j ,…,w' ij ,…,w' nj W 'in the formula' ij =i*a ij Deleting the element of 0 in W' (j) to obtain a roadway set W (j), and storing the roadway set W (j) in a roadway ventilation network fault roadway range library;
step 1.3: obtaining a roadway set W (j) to be detected according to a roadway set corresponding to the roadway where the alarm sensor is located: when only one tunnel wind speed sensor gives an alarm in an overrun mode, acquiring a tunnel set W (j) corresponding to a tunnel j where the alarm sensor is located from a tunnel ventilation network fault tunnel range library as a tunnel set W (j) to be detected; when two or more roadway wind speed sensors alarm in an overrun mode, acquiring a roadway set corresponding to a roadway where each alarm sensor is located from a roadway ventilation network fault roadway range library, and calculating the intersection of all the roadway sets to be a to-be-detected roadway set W (j).
And 2, step: establishing an ELM fault branch diagnosis model;
step 2.1: generating a sample set; the sample set comprises training samples and prediction samples;
generating a sample set without considering the total entrance and the total return of the tunnel in the ventilation network of the tunnel with the fault to be detected, so that the residual branch number is k;
training samples and prediction samples were generated as follows:
generating a training sample: increasing the resistance of k to generate samples, increasing the wind resistance by 1 step length, increasing by 6 steps, and obtaining s training samples through network calculation;
generation of prediction samples: increasing the resistance of k to generate samples, increasing the wind resistance by 1 step length by 3 steps, and obtaining z prediction samples through network calculation;
step 2.2: building an ELM fault branch diagnosis model;
step 2.2.1: programming a program; utilizing elmtran and elmpress functions; the Elmtrain function is used for creating and training an ELM, and the Elmpredict function is used for carrying out an ELM simulation test; the calling formats are respectively as follows:
[IW,B,LW,TF,TYPE]=elmtrain(P,T,N,TF,TYPE)
Y=elmpredict(P,IW,B,LW,TF,TYPE)
wherein, P is an input matrix of the training set; t is an output matrix of the training set; n is the number of samples of the training set; TF is an activation function of hidden layer neurons, and the values of TF can be 'sig', 'sin', 'hardlim'; TYPE is the application TYPE of ELM, the value of which is 0 and 1,0 represents regression and fitting, and 1 represents classification; IW is the connection weight between the input layer and the hidden layer; b is the breadth of hidden layer neurons; LW is the connection weight of the hidden layer and the output layer;
step 2.2.2: establishing an ELM fault branch diagnosis model, taking the air volume value of the tunnel where the air velocity sensor is positioned as an input value X, and changing the branch number e of the wind resistance i As an output value Y;
in this embodiment, the input value X and the output value Y are set as follows:
step 2.3: when an air speed sensor of the monitoring system gives an alarm in an overrun mode, an ELM fault branch diagnosis model is started immediately, a branch number is used as a category serial number to serve as an output value Y, and a monitoring value of the air speed sensor is used as an input value X;
and 3, step 3: verifying a prediction result;
setting alpha wind speed sensor overrun alarm, performing ELM fault branch diagnosis model training on training samples in total of s samples according to a branch sequence, and obtaining a prediction result of the training samples which is the same as a preset result, wherein as shown in FIG. 3, original data in the graph is a known fault branch number, and prediction data is a fault branch number predicted by the ELM fault branch diagnosis model; the method comprises the following steps of carrying out ELM fault branch diagnosis model training on z samples in total of test samples in a random mixing sample sequence to obtain a prediction result of the test samples, as shown in FIG. 4, predicting correct 48 samples with the accuracy rate of 94.1176%, calculating the mean square error between a predicted value and a true value of a test set to be 4.7451 and the square correlation coefficient to be 0.8969, as shown in FIG. 5; the feasibility of the ELM fault branch diagnosis model for performing mine wind speed fault branch diagnosis is verified.
The second embodiment:
taking an iron-method damming mine as a test mine, carrying out wind speed fault branch diagnosis on the damming mine, firstly carrying out ventilation resistance test on the damming mine, and carrying out network solution by virtue of mine ventilation simulation software MVSS; the daming mine is a 4-in 2-pass system, the total number of the roadway branches is 330, and the errors of the air quantity values calculated by the network and the air quantity values actually tested are within 5%. The total grade hole of the mine is 2.5509m2, the Mingming mine is an easily ventilated mine, and the system is provided with a tunnel of a wind speed sensor as shown in table 1;
TABLE 1 air volume change of the tunnel where the air velocity sensor is located
Step 1: determining a fault roadway range according to a fault roadway determination method based on a wind resistance-wind flow change influence relation matrix;
step 1.1: constructing a wind resistance-wind flow change influence relation matrix of the roadway to be detected with the fault:
wherein i, j = {1,2,3, …, n }, and n is the total number of roadways in the roadway ventilation network to be detected; a is ij The value is obtained by the following method: sequentially changing the wind resistance of the ith roadway in the ventilation network of the roadway to be detected with the fault, keeping the wind resistance of other roadways unchanged, obtaining the wind volume values of all the roadways through network calculation, converting the wind volume values into wind speed values, comparing the wind speed values with the allowable wind speed value range of the corresponding roadway in coal mine safety regulations, and judging whether the wind speed value of the jth roadway exceeds the allowable range, if so, a ij =1, otherwise, a ij =0;
Step 1.2: determining an initial set W '(j) = { W' 1j ,w' 2j ,…,w' ij ,…,w' nj W 'in the formula' ij =i*a ij Deleting the element of 0 in W' (j) to obtain a roadway set W (j), and storing the roadway set W (j) in a roadway ventilation network fault roadway range library;
in this embodiment, the sets of fault branches of the roadway where the damming mine wind speed sensor is located are respectively:
W(303)={132,134,138,155,160,180,181,192,196,217,239,240,267,303,310,313,324};
W(159)={132,134,138,155,159,164,180,181,192,196,217,239,240,264,267,313,324};
W(240)={15,16,18,19,20,21,45,132,134,138,147,155,164,179,180,181,182,190,192,196,199,200,202,203,204,205,206,217,237,239,240,242,264,265,266,267,313,324};
W(199)={179,181,182,190,192,196,199,200,202,203,204,205,206,209,210,211,217,245,250,264,267,313,324};
W(21)={2,3,4,5,15,16,18,19,20,23,24,25,30,41,44,45,51,65,67,69,70,80,81,82,83,84,85,89,90,91,92,93,94,103,113,115,121,124,126,129,130,131,132,134,138,142,152,155,166,167,180,181,192,194,196,217,225,227,230,232,233,239,240,264,265,266,267,271,308,309,311,313,314,316,321,322,324};
W(46)={2,3,4,15,16,18,19,20,21,23,25,30,41,44,45,46,47,48,51,54,55,60,65,67,69,70,73,80,81,82,83,84,85,89,90,91,92,93,94,95,96,103,113,115,119,121,123,124,126,128,129,130,132,134,196,232,233,264,267,309,311,316};
W(69)={73,81,82,83,84,89,90,91,92,93,94,95,132,134,160,180,181,192,196,217,239,240,264,267,303,313,324};
step 1.3: obtaining a roadway set W (j) to be detected according to a roadway set corresponding to the roadway where the alarm sensor is located: when only one tunnel wind speed sensor gives an alarm in an overrun mode, acquiring a tunnel set W (j) corresponding to a tunnel j where the alarm sensor is located from a tunnel ventilation network fault tunnel range library as a tunnel set W (j) to be detected; and when two or more roadway wind speed sensors alarm in an overrun mode, acquiring a roadway set corresponding to the roadway where each alarm sensor is located from a roadway ventilation network fault roadway range library, and calculating the intersection of all the roadway sets to be a roadway set W (j) to be detected.
In this embodiment, a set W × (j) of lanes to be detected is = {2,3,4,5,15,16,18,19,20,21,23,24,25,30,41,44,45,46,47,48,51,54,55,60,65,67,69,70,73,80, 81,82,83,81,85,89,90,91,92,93,94,95,96,103,113,115,119,121,123,124,126,128,129,130,131,132,134,138,142,152,155,159,160,164,166,167,170,179,180,181,182,190,192,194,196,199,200,202,203,204,205,206,209,210,211,217,225, 227,230,232,233,237,239,240,242,245,250,264,265,266,267, 271,303,308,309,310,311,313,314,316,321,322,324} for a total of 113 branches, and the branch number is a branch number in the set W × (j);
and 2, step: building an ELM fault branch diagnosis model;
step 2.1: generating a sample set; the sample set comprises training samples and prediction samples;
generating a sample set without considering the total entrance and the total return of the roadway in the roadway ventilation network to be detected with the fault, wherein the number of the remaining branches is k;
training samples and prediction samples were generated as follows:
generating a training sample: increasing the resistance of k to generate samples, increasing the wind resistance by 1 step by 6 steps, and resolving through a network to obtain s training samples;
generation of prediction samples: increasing the resistance of k to generate samples, increasing the wind resistance by 1 step length by 3 steps, and obtaining z prediction samples through network calculation;
step 2.2: building an ELM fault branch diagnosis model;
step 2.2.1: programming a program; utilizing elmtractin and elmpress functions; the Elmtrain function is used for creating and training an ELM, and the Elmpredict function is used for carrying out an ELM simulation test; the calling formats are respectively as follows:
[IW,B,LW,TF,TYPE]=elmtrain(P,T,N,TF,TYPE)
Y=elmpredict(P,IW,B,LW,TF,TYPE)
wherein, P is an input matrix of the training set; t is an output matrix of the training set; n is the number of samples of the training set; TF is an activation function of hidden layer neurons, and the values of TF can be 'sig', 'sin', 'hardlim'; TYPE is the application TYPE of ELM, the value of which is 0 and 1,0 represents regression and fitting, and 1 represents classification; IW is the connection weight between the input layer and the hidden layer; b is the breadth of hidden layer neurons; LW is the connection weight of the hidden layer and the output layer;
step 2.2.2: building an ELM fault branch diagnosis model, taking the air volume value of the roadway where the air velocity sensor is as an input value X, and changing the branch number e of the wind resistance i As an output value Y;
in this embodiment, the input value X and the output value Y are set as follows:
step 2.3: when an air speed sensor of the monitoring system alarms in an overrun mode, an ELM fault branch diagnosis model is started immediately, a branch number is used as an output value Y, and a monitoring value of the air speed sensor is used as an input value X;
and step 3: verifying a prediction result;
setting alpha wind speed sensor overrun alarm, counting s samples of training samples, and carrying out ELM fault branch diagnosis model training according to a branch sequence; the method comprises the following steps of (1) randomly mixing z samples of test samples, and carrying out ELM fault branch diagnosis model training on the sample sequence; comparing the obtained training sample prediction result and the prediction result of the test sample with a preset result; and verifying the feasibility of the ELM fault branch diagnosis model for performing the mine wind speed fault branch diagnosis.
In this embodiment, under the condition of ensuring the safety of the ventilation system of the daming mine, the air door of-410 northwest main return airway (branch 179) is opened on site, at this time, the air speed sensors of the return airway (branch 199) of the mining area of the west wing and the return airway (branch 240) special for the west wing overrun the alarm, the air volume results measured by each air speed sensor are shown in table 1,
and (3) taking the variable air volume values in the table 1 as input values, diagnosing the wind speed fault branch by using an ELM fault branch diagnosis model, wherein the prediction result is the No. 179 branch, and the fault branch diagnosis result is consistent with the industrial test result.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit of the corresponding technical solutions and scope of the present invention as defined in the appended claims.
Claims (1)
1. A mine wind speed fault branch diagnosis method based on ELM is characterized in that: the method comprises the following steps:
step 1: determining a fault roadway range according to a fault roadway determination method based on a wind resistance-wind flow change influence relation matrix;
step 1.1: constructing a wind resistance-wind flow change influence relation matrix of the roadway to be detected with the fault:
wherein i, j = {1,2,3, … n }, and n is the total number of the roadways in the ventilation network of the roadway to be detected with the fault; a is ij The value is obtained by the following method: sequentially changing the wind resistance of the ith roadway in the ventilation network of the roadway to be detected with the fault, keeping the wind resistance of other roadways unchanged, obtaining the wind volume values of all the roadways through network calculation, converting the wind volume values into wind speed values, comparing the wind speed values with the allowable wind speed value range of the corresponding roadway in coal mine safety regulations, and judging whether the wind speed value of the jth roadway exceeds the allowable range, if so, a ij =1, otherwise, a ij =0;
Step 1.2: determining an initial set W '(j) = { W' 1j ,w′ 2j ,…,w′ ij ,…,w′ nj W 'in the formula' ij =i*a ij Erasure and deleteExcept the element of 0 in W' (j), obtaining a roadway set W (j), and storing the roadway set W (j) into a roadway ventilation network fault roadway range library;
step 1.3: and obtaining a roadway set W (j) to be detected according to the roadway set corresponding to the roadway where the wind speed sensor is in the overrun alarm: when only one tunnel wind speed sensor gives an alarm in an overrun mode, acquiring a tunnel set W (j) corresponding to a tunnel j where the alarm sensor is located from a tunnel ventilation network fault tunnel range library as a tunnel set W (j) to be detected; when two or more roadway wind speed sensors alarm in an overrun mode, acquiring a roadway set corresponding to a roadway where each alarm sensor is located from a roadway ventilation network fault roadway range library, and calculating the intersection of all the roadway sets to be a roadway set W (j) to be detected;
step 2: establishing an ELM fault branch diagnosis model;
step 2.1: generating a sample set; the sample set comprises training samples and prediction samples;
generating a sample set without considering the total entrance and the total return of the tunnel in the ventilation network of the tunnel with the fault to be detected, so that the residual branch number is k;
training samples and prediction samples were generated as follows:
generating a training sample: increasing the resistance of k to generate samples, increasing the wind resistance by 1 step by 6 steps, and resolving through a network to obtain s training samples;
generation of prediction samples: increasing the resistance of k to generate samples, increasing the wind resistance by 1 step length by 3 steps, and obtaining z prediction samples through network calculation;
step 2.2: establishing an ELM fault branch diagnosis model;
step 2.2.1: programming a program; utilizing elmtran and elmpress functions; the Elmtrain function is used for creating and training an ELM, and the Elmpredict function is used for carrying out an ELM simulation test; the calling formats are respectively as follows:
[IW,B,LW,TF,TYPE]=elmtrain(P,T,N,TF,TYPE)
Y=elmpredict(P,IW,B,LW,TF,TYPE)
wherein, P is an input matrix of the training set; t is an output matrix of the training set; n is the number of samples of the training set; TF is an activation function of hidden layer neurons, and the values of TF can be 'sig', 'sin', 'hardlim'; TYPE is the application TYPE of ELM, the value of which is 0 and 1,0 represents regression and fitting, and 1 represents classification; IW is the connection weight between the input layer and the hidden layer; b is the breadth of the hidden layer neuron; LW is the connection weight of the hidden layer and the output layer;
step 2.2.2: establishing an ELM fault branch diagnosis model, taking the air volume value of the tunnel where the air velocity sensor is positioned as an input value X, and changing the branch number e of the wind resistance i As an output value Y;
step 2.3: when an air speed sensor of the monitoring system alarms in an overrun mode, an ELM fault branch diagnosis model is started immediately, a branch number is used as an output value Y, and a monitoring value of the air speed sensor is used as an input value X;
and 3, step 3: verifying a prediction result;
setting alpha wind speed sensor overrun alarm, counting s samples of training samples, and carrying out ELM fault branch diagnosis model training according to a branch sequence; the method comprises the following steps of (1) randomly mixing z samples of test samples, and carrying out ELM fault branch diagnosis model training on the random mixed sample sequence; comparing the obtained training sample prediction result and the prediction result of the test sample with a preset result; and verifying the feasibility of the ELM fault branch diagnosis model for performing the mine wind speed fault branch diagnosis.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811326508.5A CN109446697B (en) | 2018-11-08 | 2018-11-08 | ELM-based mine wind speed fault branch diagnosis method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811326508.5A CN109446697B (en) | 2018-11-08 | 2018-11-08 | ELM-based mine wind speed fault branch diagnosis method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109446697A CN109446697A (en) | 2019-03-08 |
CN109446697B true CN109446697B (en) | 2023-04-18 |
Family
ID=65552398
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811326508.5A Active CN109446697B (en) | 2018-11-08 | 2018-11-08 | ELM-based mine wind speed fault branch diagnosis method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109446697B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111693726B (en) * | 2019-03-14 | 2021-11-23 | 辽宁工程技术大学 | Ventilation system fault diagnosis wind speed sensor arrangement method based on neighborhood rough set |
CN110566259B (en) * | 2019-09-26 | 2021-04-13 | 辽宁工程技术大学 | Ventilation system resistance variation type fault diagnosis method based on air volume and air pressure monitoring value |
CN110705114B (en) * | 2019-10-10 | 2023-04-07 | 辽宁工程技术大学 | Ventilation fault diagnosis method without training sample |
CN112257927B (en) * | 2020-10-22 | 2023-12-15 | 上海基蓝软件有限公司 | SF-based 6 Method for accurately testing air quantity of full-air net |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102999786B (en) * | 2012-10-17 | 2016-08-31 | 浙江埃菲生能源科技有限公司 | Photovoltaic generation power short-term earthquake prediction method based on similar day tagsort Yu extreme learning machine |
CN103400210A (en) * | 2013-08-13 | 2013-11-20 | 广西电网公司电力科学研究院 | Short-term wind-speed combination forecasting method |
CN104123476B (en) * | 2014-08-12 | 2017-03-08 | 大连海事大学 | Forecasting of Gas Concentration method based on extreme learning machine and its device |
CN105138782A (en) * | 2015-09-02 | 2015-12-09 | 上海大学 | Non-stationary pulse wind speed high-precision prediction method based on EEMD-ELM |
CN105205495A (en) * | 2015-09-02 | 2015-12-30 | 上海大学 | Non-stationary fluctuating wind speed forecasting method based on EMD-ELM |
US10057367B2 (en) * | 2016-03-02 | 2018-08-21 | Huawei Technologies Canada Co., Ltd. | Systems and methods for data caching in a communications network |
CN108090235A (en) * | 2016-11-21 | 2018-05-29 | 辽宁工程技术大学 | The failure tunnel for influencing relational matrix based on windage-distinguished and admirable variation determines method |
CN107122861B (en) * | 2017-04-28 | 2020-02-11 | 辽宁工程技术大学 | Gas emission quantity prediction method based on PCA-PSO-ELM |
-
2018
- 2018-11-08 CN CN201811326508.5A patent/CN109446697B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN109446697A (en) | 2019-03-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109446697B (en) | ELM-based mine wind speed fault branch diagnosis method | |
CN108507117A (en) | A kind of Air-conditioning system sensor method for diagnosing faults based on wavelet neural network | |
Hourdakis et al. | Practical procedure for calibrating microscopic traffic simulation models | |
CN110319982B (en) | Buried gas pipeline leakage judgment method based on machine learning | |
CN112036075A (en) | Abnormal data judgment method based on environmental monitoring data association relation | |
CN108831121B (en) | Early warning method and device for mine safety production | |
CN115081963B (en) | Underground water quality risk analysis method and system | |
CN108266219A (en) | Mine ventilation system resistive-switching single fault source diagnostic method based on air quantity feature | |
CN111191855B (en) | Water quality abnormal event identification and early warning method based on pipe network multi-element water quality time sequence data | |
CN111505064A (en) | Catalytic combustion type methane sensor service state evaluation method | |
US11416643B2 (en) | Smoke detection system layout design | |
CN115310361A (en) | Method and system for predicting underground dust concentration of coal mine based on WGAN-CNN | |
CN117114206B (en) | Calculation method for coal mine water damage index data trend | |
CN109086186A (en) | log detection method and device | |
CN108090235A (en) | The failure tunnel for influencing relational matrix based on windage-distinguished and admirable variation determines method | |
JP4164556B2 (en) | Dynamic simulation device for road tunnel ventilation control and its program | |
CN111477012A (en) | Tracing method and device based on road condition state prediction model and computer equipment | |
CN106874627B (en) | Detection method for detecting construction quality and working state of mine anchor rod | |
CN116164241A (en) | Intelligent detection method for leakage faults of gas extraction pipe network | |
CN114922612A (en) | Coal gas permeability prediction method based on LVQ-CPSO-BP algorithm | |
CN112836990B (en) | Tunnel monitoring equipment fault judging method and device and electronic equipment | |
CN113255151A (en) | Tunnel construction deformation prediction method and system based on composite neural network | |
CN107957269B (en) | Inertial navigation system fault characteristic judgment and testability prediction method | |
JP7251048B2 (en) | Congestion prediction device | |
Xie et al. | Secondary Modeling of Air Quality Based on LSTM Cycle Neural Network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |