CN110566259B - Ventilation system resistance variation type fault diagnosis method based on air volume and air pressure monitoring value - Google Patents
Ventilation system resistance variation type fault diagnosis method based on air volume and air pressure monitoring value Download PDFInfo
- Publication number
- CN110566259B CN110566259B CN201910916393.3A CN201910916393A CN110566259B CN 110566259 B CN110566259 B CN 110566259B CN 201910916393 A CN201910916393 A CN 201910916393A CN 110566259 B CN110566259 B CN 110566259B
- Authority
- CN
- China
- Prior art keywords
- fault
- branch
- mine ventilation
- ventilation network
- 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
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH DRILLING; MINING
- E21F—SAFETY DEVICES, TRANSPORT, FILLING-UP, RESCUE, VENTILATION, OR DRAINING IN OR OF MINES OR TUNNELS
- E21F1/00—Ventilation of mines or tunnels; Distribution of ventilating currents
- E21F1/006—Ventilation at the working face of galleries or tunnels
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH DRILLING; MINING
- E21F—SAFETY DEVICES, TRANSPORT, FILLING-UP, RESCUE, VENTILATION, OR DRAINING IN OR OF MINES OR TUNNELS
- E21F17/00—Methods or devices for use in mines or tunnels, not covered elsewhere
- E21F17/18—Special adaptations of signalling or alarm devices
Abstract
The invention provides a ventilation system resistance type fault diagnosis method based on an air volume and air pressure monitoring value. The method comprises the steps of generating a simulation data fault sample set about resistance change fault, namely air volume and air pressure through mine ventilation network resolving, then respectively constructing a support vector machine classification model and a support vector machine regression model, using air volume and air pressure monitoring values as input of a support vector machine, using fault branch numbers as output of the classification model, and using fault quantity as output of the regression model, so as to realize diagnosis of fault positions and fault quantities.
Description
Technical Field
The invention relates to the technical field of mine ventilation, in particular to a ventilation system resistance variation type fault diagnosis method based on an air volume and air pressure monitoring value.
Background
The wind resistance of the mine ventilation system can be changed under the condition that the section area of the tunnel is changed due to the air door switch or damage, the tunnel falling or deformation, the mine car running, the cage lifting and the like in the mine ventilation system, the condition that the wind resistance is durably changed due to the air door switch or damage, the tunnel falling or deformation and the like is called as the resistance change type fault of the mine ventilation system, the place where the fault occurs is called as a fault position, and the size of the wind resistance change is called as a fault amount. Under the condition of normal operation of a mine, all parameters of a roadway should be kept in a relatively stable state, and once a fault occurs underground, all parameters of the mine can be changed, so that serious threat is generated to safe production. Therefore, timely and accurate diagnosis of underground faults has great significance in reducing loss caused by faults and ensuring normal operation of a mine ventilation system.
However, the current diagnosis of the downhole fault only uses a single characteristic of air volume, the fault diagnosis based on the single characteristic of air volume is an inappropriate problem due to the lack of n-m +1 linear independent equations, and the accuracy of the current sensor cannot reach high enough monitoring accuracy, so that the accuracy of the fault diagnosis is low.
The diagnosis process of the resistance variation type fault of the ventilation system is to make decisive judgment on the operation state of the ventilation system by using the monitoring values of main parameters of the mine ventilation system, such as roadway air pressure, dust concentration, carbon dioxide concentration and the like. Because the underground wind pressure is easy to monitor and sensitive to the deformation resistance type fault, the invention combines the wind pressure characteristic to form the wind volume and wind pressure composite characteristic and uses the Support Vector Machine (SVM) algorithm to monitor the underground fault so as to improve the accuracy of the fault diagnosis of the deformation resistance type of the mine ventilation system.
Disclosure of Invention
The method takes the common monitoring value of the air volume and the air pressure as the input of a support vector machine, avoids the discomfort of the single characteristic of the air volume, reduces the limitation of the single characteristic, provides technical support for the fault diagnosis of the ventilation system, and has great significance for improving the accuracy of the fault diagnosis.
In order to solve the technical problem, the invention provides a ventilation system resistance variation type fault diagnosis method based on an air volume and air pressure monitoring value, which comprises the following steps:
step 1: preparing topological relation data of a mine ventilation network, wherein the topological relation data comprise original wind resistance and wind volume of each branch of the mine ventilation network and a fan characteristic curve of a fan, and the fan characteristic curve equation is specifically expressed as follows:
Hτ＝t0+t1Qτ+t2Qτ 2 (1)
in the formula, HτIndicating wind pressure, Q, of the ventilatorτIndicating the air volume, t, of the ventilator0,t1,t2Representing the wind pressure characteristic coefficient of the ventilator;
step 2: resolving through a mine ventilation network to obtain the wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network after the fault occurs, and generating a simulation data fault sample set about resistance change fault-wind volume and wind pressure, wherein the fault sample set comprises the fault number and the fault volume of each branch of the mine ventilation network, the original wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network after the fault occurs, and the wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network after the fault occurs;
and step 3: and respectively constructing a support vector machine classification model taking a monitoring fault position as a target and a support vector machine regression model taking a monitoring fault amount as a target according to the fault sample set, taking the air volume and the air pressure of each branch of the mine ventilation network after the fault occurs in the fault sample set as the input of the support vector machine, taking the number of each branch of the mine ventilation network in the fault sample set as the output of the support vector machine classification model, and taking the fault amount of each branch of the mine ventilation network in the fault sample set as the output of the support vector machine regression model, so that the diagnosis of the fault position and the fault amount is realized.
Step 2, resolving through a mine ventilation network to obtain the wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network after the fault, and generating a simulation data fault sample set about resistance change fault-wind volume and wind pressure, wherein the fault sample set comprises the fault number and the fault volume of each branch of the mine ventilation network, the original wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network, and the wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network after the fault occurs, and the specific expression is as follows:
step 2.1: defining an original air volume matrix of all branches of a mine ventilation network as Q, an original wind resistance matrix as R, an original wind pressure matrix as H, and recording as:
and satisfies the following conditions:
in the formula, qnRepresenting the original air volume, r, of the nth network branchnRepresenting the original wind resistance, h, of the nth network branchnIndicates the original wind pressure of the nth network branch, and B ═ Bij)m×nA complete correlation matrix, q, representing the mine ventilation networkjRepresenting the original air volume of the jth network branch, bijThe ith row and the jth column of the complete correlation matrix of the mine ventilation network are represented, m represents the number of nodes of the mine ventilation network, and C ═ Cij)s×nA loop matrix representing a mine ventilation network, cijThe ith row and the jth column of the loop matrix representing the mine ventilation network, s representing the total number of loops of the mine ventilation network, n representing the number of branches of the mine ventilation network, HTTranspose of the original wind pressure matrix H, HηAdditional resistance matrix, H, representing all loops of the mine ventilation networkη TAdditional resistance matrix H representing all loops of mine ventilation networkηTranspose of hjJ column, H, representing the original wind pressure matrix H of the mine ventilation networkηiLoop additive resistance matrix H representing mine ventilation networkηRow i of (1);
step 2.2: hypothesis branch eiA resistive type fault occurs with a fault amount of delta riCalculating branch e using equation (5)iThe wind resistance matrix R ' consisting of the wind resistances of all branches of the mine ventilation network after the fault is expressed as R ' ═ R '1,r'2,…,r'n}，
r'i＝ri±△ri，(i＝1,2,...,n) (5)
In the formula, ri' denotes the branch eiTake place ofBranch windage after failure, riDenotes the branch eiOriginal wind resistance, Δ riDenotes the branch eiThe number of faults occurring, n representing the number of all branches of the mine ventilation network;
step 2.3: and solving the air volume of the mine ventilation network after the fault occurs through network calculation, and expressing an air volume matrix Q ' consisting of the air volumes of all branches of the mine ventilation network after the fault occurs as Q ' ═ Q '1,q'2,…,q'nThe network solution is expressed as:
Q'＝f(R') (6)
in the formula, Q 'represents the air volume of the mine ventilation network after the fault, R' represents the wind resistance of the mine ventilation network after the fault, and f (#) represents a function related to the wind resistance and the air volume;
step 2.4: solving the wind pressure of the mine ventilation network after the fault occurs by using the formula (7), and then a wind pressure matrix H ' consisting of the wind pressures of all branches of the mine ventilation network after the fault occurs is represented as H ' ═ H '1,h'2,…,h'n}，
Wherein h'iIndicating a post-fault branch eiWind pressure of ri' denotes the branch eiBranch windage after failure, q'iIndicating a post-fault branch eiThe flow rate of the mine ventilation network is determined by the flow state factor, x is 1 in laminar flow and 2 in complete turbulent flow, the transition state takes an intermediate value of 1-2, and n represents the number of all branches of the mine ventilation network;
step 2.5: calculating the wind resistance matrix R ', the wind volume matrix Q ' and the wind pressure matrix H ' of the mine ventilation network after the fault, and numbering the fault branch e when the fault occurs each timeiΔ r amount of failureiAnd recording the original wind resistance value, wind quantity and wind pressure of each branch of the mine ventilation network to a fault sample set.
step 3.1: constructing a support vector machine classification model aiming at monitoring a fault position by using the fault sample set, taking the air volume and the air pressure of each branch of the mine ventilation network after the fault occurs in the fault sample set as the input of the support vector machine classification model, taking the serial number of each branch of the mine ventilation network in the fault sample set as the output of the support vector machine classification model, introducing a symbolic function as shown in a formula (9), classifying the samples with the symbolic function value of 1 into one class, classifying the samples with the symbolic function value of-1 into another class, realizing the correct classification of the fault samples, and then realizing the classification of the branch e by using a formula (8)iDiagnosing the fault position;
where x denotes the input sample, sgn denotes the sign function, xiThe ith support vector of the support vector machine classification model is shown, l represents the number of the support vector machines,
representing the Lagrangian multiplier, yiRepresenting a category label, b*Threshold vector, f (x), representing the support vector machine kerneli) Watch (A)For support vector xiThe sign function of (a);
step 3.2: constructing a regression model of a support vector machine with the aim of monitoring fault amount by using the fault sample set, taking the air volume and the air pressure of each branch of the mine ventilation network after the fault occurs in the fault sample set as the input of the regression model of the support vector machine, taking the fault amount of each branch of the mine ventilation network in the fault sample set as the output of the regression model of the support vector machine, adopting an epsilon-insensitive loss function with sparsity shown in a formula (10) to realize the diagnosis of the fault amount, namely realizing the regression of the samples in the fault sample set,
c(y,f*(x))＝max{0,|y-f*(x)-ε}；ε>0 (10)
where y denotes the sample in the fault sample set, f*(x) Representing a hyperplane in the feature space, and epsilon represents regression precision;
step 3.3: and checking the generated fault position diagnosis classification model and the fault quantity diagnosis regression model, comparing the fault branch number of the fault position diagnosis classification model with the fault number of each branch in the fault sample set, comparing the wind resistance value of each branch after the fault output by the fault quantity diagnosis regression model with the wind resistance value of each branch after the fault in the fault sample set, and respectively drawing a scatter plot comparison chart to verify the accuracy of the output diagnosis of the fault diagnosis classification model and the regression model.
The invention has the beneficial effects that:
the invention provides a ventilation system resistance variation type fault diagnosis method based on an air volume and air pressure monitoring value, which is based on a ventilation network, the network topological relation of the ventilation system is complex, the change of the air resistance of one branch can cause the change of the air volume and the air pressure of the branch or even the whole ventilation network, and the faults of different branches in different degrees can cause the different numerical values of the change of the air volume and the air pressure, so that the serial number of the fault branch, the fault volume, the branch air volume and the air pressure are in one-to-one correspondence, the characteristics of the air volume and the air pressure can be used as the attributes of fault diagnosis, and the information complementation of multidimensional characteristics is realized by the composite characteristics, the unsuitability of a single air volume characteristic is avoided, and the accuracy of the.
Drawings
Fig. 1 is a flowchart of a method for diagnosing a fault of a resistance variation type of a ventilation system based on an air volume and air pressure monitoring value in an embodiment of the invention.
FIG. 2 is a schematic diagram of support vector machine regression.
Fig. 3 is a diagram of a mine ventilation network in an embodiment of the invention.
Detailed Description
The invention will be further described with reference to the accompanying drawings and specific examples: a ventilation system resistance variation type fault diagnosis method based on an air volume and air pressure monitoring value is provided, the flow of which is shown in figure 1, and the method specifically comprises the following steps:
step 1: preparing topological relation data of a mine ventilation network, wherein the topological relation data comprise original wind resistance and wind volume of each branch of the mine ventilation network and a fan characteristic curve of a fan, and the fan characteristic curve equation is specifically expressed as follows:
Hτ＝t0+t1Qτ+t2Qτ 2 (1)
in the formula, HτIndicating wind pressure, Q, of the ventilatorτIndicating the air volume, t, of the ventilator0,t1,t2Representing the wind pressure characteristic coefficient of the ventilator;
as shown in fig. 3, the number of branches n and the number of nodes m of the mine ventilation network are 100 and 71, and the branches provided with air volume adjusting facilities such as air doors and air windows are shown in fig. 3, the original wind resistance of each branch is shown in table 1, and the mine ventilation network is provided with three ventilators respectively installed on e9、e39And e78The characteristic curves of the three ventilators are respectively as follows:
H1(q)＝1932.25+44.84q-0.64q2；
H2(q)＝1828.13+19.2q-0.08q2；
H3(q)＝3054.94+8.64q-0.05q2，
under the condition of original wind resistance parameters, the ventilation systemThe total wind resistance of the system network is 0.012 Ns2/m8The operating point is (474.678 m)3/s，2644.5Pa)。
TABLE 1 mine Ventilation network Branch parameters
Step 2: resolving through a mine ventilation network to obtain the wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network after the fault, and generating a simulation data fault sample set about resistance change fault-wind volume and wind pressure, wherein the fault sample set comprises the fault serial number and the fault volume of each branch of the mine ventilation network, the original wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network, and the wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network after the fault occurs, and the specific expression is as follows:
step 2.1: defining an original air volume matrix of all branches of a mine ventilation network as Q, an original wind resistance matrix as R, an original wind pressure matrix as H, and recording as:
and satisfies the following conditions:
in the formula, qnRepresenting the original air volume, r, of the nth network branchnRepresenting the original wind resistance, h, of the nth network branchnIndicates the original wind pressure of the nth network branch, and B ═ Bij)m×nA complete correlation matrix, q, representing the mine ventilation networkjRepresenting the original air volume of the jth network branch, bijIndicating mine ventilationThe ith row and the jth column of the network complete correlation matrix are arranged, m represents the number of nodes of the mine ventilation network, and C is (C)ij)s×nA loop matrix representing a mine ventilation network, cijThe ith row and the jth column of the loop matrix representing the mine ventilation network, s representing the total number of loops of the mine ventilation network, n representing the number of branches of the mine ventilation network, HTTranspose of the original wind pressure matrix H, HηAdditional resistance matrix, H, representing all loops of the mine ventilation networkη TAdditional resistance matrix H representing all loops of mine ventilation networkηTranspose of hjJ column, H, representing the original wind pressure matrix H of the mine ventilation networkηiLoop additive resistance matrix H representing mine ventilation networkηRow i of (1);
step 2.2: hypothesis branch eiA resistive type fault occurs with a fault amount of delta riCalculating branch e using equation (5)iThe wind resistance matrix R ' consisting of the wind resistances of all branches of the mine ventilation network after the fault is expressed as R ' ═ R '1,r'2,…,r'n}，
r'i＝ri±△ri，(i＝1,2,...,n) (5)
In the formula (II), r'iDenotes the branch eiBranch windage after fault, riDenotes the branch eiOriginal wind resistance, Δ riDenotes the branch eiThe number of faults occurring, n representing the number of all branches of the mine ventilation network;
step 2.3: and solving the air volume of the mine ventilation network after the fault occurs through network calculation, and expressing an air volume matrix Q ' consisting of the air volumes of all branches of the mine ventilation network after the fault occurs as Q ' ═ Q '1,q'2,…,q'nThe network solution is expressed as:
Q'＝f(R') (6)
in the formula, Q 'represents the air volume of the mine ventilation network after the fault, R' represents the wind resistance of the mine ventilation network after the fault, and f (#) represents a function related to the wind resistance and the air volume;
step 2.4: solving the wind pressure of the mine ventilation network after the fault occurs by using the formula (7), and then a wind pressure matrix H ' consisting of the wind pressures of all branches of the mine ventilation network after the fault occurs is represented as H ' ═ H '1,h'2,…,h'n}，
Wherein h'iIndicating a post-fault branch eiWind pressure of r'iDenotes the branch eiBranch windage after failure, q'iIndicating a post-fault branch eiThe flow rate of the mine ventilation network is determined by the flow state factor, x is 1 in laminar flow and 2 in complete turbulent flow, the transition state takes an intermediate value of 1-2, and n represents the number of all branches of the mine ventilation network;
step 2.5: calculating the wind resistance matrix R ', the wind volume matrix Q ' and the wind pressure matrix H ' of the mine ventilation network after the fault, and numbering the fault branch e when the fault occurs each timeiΔ r amount of failureiAnd recording the original wind resistance value, wind quantity and wind pressure of each branch of the mine ventilation network to a fault sample set.
And step 3: respectively constructing a support vector machine classification model taking a monitoring fault position as a target and a support vector machine regression model taking a monitoring fault amount as a target according to the fault sample set, taking the air volume and the air pressure of each branch of the mine ventilation network after the fault occurs in the fault sample set as the input of the support vector machine, taking the number of each branch of the mine ventilation network in the fault sample set as the output of the support vector machine classification model, and taking the fault amount of each branch of the mine ventilation network in the fault sample set as the output of the support vector machine regression model, so as to realize the diagnosis of the fault position and the fault amount, wherein the specific expression is as follows:
step 3.1: constructing a support vector machine classification model taking the monitoring fault position as a target by using the fault sample set, and taking the air volume of each branch of the mine ventilation network after the fault occurs in the fault sample setAnd wind pressure is used as the input of a classification model of the support vector machine, the serial numbers of all branches of the mine ventilation network in the fault sample set are used as the output of the classification model of the support vector machine, the sign function is introduced as the formula (9), the samples with the sign function value of 1 are divided into one class, the samples with the sign function value of-1 are divided into another class, the correct classification of the fault samples is realized, and then the formula (8) is used for realizing the classification of the branch eiDiagnosing the fault position;
where x denotes the input sample, sgn denotes the sign function, xiThe ith support vector of the support vector machine classification model is shown, l represents the number of the support vector machines,
representing the Lagrangian multiplier, yiRepresenting a category label, b*Threshold vector, f (x), representing the support vector machine kerneli) Representation for support vector xiThe sign function of (a);
step 3.2: constructing a regression model of a support vector machine with the aim of monitoring fault amount by using the fault sample set, taking the air volume and the air pressure of each branch of the mine ventilation network after the fault occurs in the fault sample set as the input of the regression model of the support vector machine, taking the fault amount of each branch of the mine ventilation network in the fault sample set as the output of the regression model of the support vector machine, adopting an epsilon-insensitive loss function with sparsity shown in a formula (10) to realize the diagnosis of the fault amount, namely realizing the regression of the samples in the fault sample set,
c(y,f*(x))＝max{0,|y-f*(x)|-ε}；ε>0 (10)
in the formula, y represents a failure sampleConcentrated sample, f*(x) Representing a hyperplane in the feature space, and epsilon represents regression precision;
the essence of equation (10) is to determine a hyperplane f (x) such that the predicted value f is*(xi) Both sides of the hyperplane f (x) are in a space determined by taking epsilon as the width, if the predicted value exists in the space, the prediction effect is good, and the original data is not lost, as shown in fig. 2.
In the training process, the branched air volume and the air pressure of the ventilation system are used as input characteristics, all the branched air volumes of the characteristic vector ventilation system are input, the fault branch number is used as output characteristics, and the training process is established through an SVM (support vector machine) one-to-one algorithm
Two-classifier SVMabHere, a and b respectively represent two different branch numbers, which respectively correspond to the branch air volume samples of the ventilation system at the time of the failure, and the branch number values are taken for both a and b. Such as SVM28Is denoted by e2Monitoring air volume sample under branch fault condition and e8And (3) monitoring a binary model constructed by the air volume sample under the condition of branch fault, and analogizing the binary models between every two other fault branches. After training is finished, the SVM fault position diagnosis classifier is finally formed by the training data and the training data.
Step 3.3: and checking the generated fault position diagnosis classification model and the fault quantity diagnosis regression model, comparing the fault branch number of the fault position diagnosis classification model with the fault number of each branch in the fault sample set, comparing the wind resistance value of each branch after the fault output by the fault quantity diagnosis regression model with the wind resistance value of each branch after the fault in the fault sample set, and respectively drawing a scatter plot comparison chart to verify the accuracy of the output diagnosis of the fault diagnosis classification model and the regression model.
Assuming that each branch in a ventilation system in a simulation experiment has multiple resistance change faults, for each wind resistance change of any branch, wind volume calculation is carried out on the ventilation network, and the fault branch number e when the fault occurs is numberediAnd equivalent wind resistance r 'of fault branch'iResistance, resistanceVariable Δ riRecording the branch air quantity Q and the air pressure H of the ventilation system as a fault sample set;
network resolving is carried out by utilizing the specific steps given in the step 2, 47100 fault samples generated by the ventilation system are obtained, and all the fault samples are recorded to form a fault sample data space;
respectively generating training and testing sample data sets in an interlaced sample extraction mode, namely generating the training sample set by using odd-numbered samples of an original sample set, generating the testing sample set by using even-numbered samples, 23550 training samples and 23550 testing samples, wherein the fault position diagnosis accuracy reaches 98.2293%, the relative error of fault amount diagnosis is shown in table 2, wherein the relative error sigma is:
TABLE 2 relative error of diagnosis of failure amount of test sample in percentage of each
The fault position diagnosis is carried out by utilizing the composite characteristics of air volume and air pressure, and the accuracy reaches 98.2293%; in the fault quantity diagnosis result, the number of the test samples with the relative error between the diagnosed fault quantity and the real fault quantity smaller than 5% reaches 14598, the total number of the whole test samples reaches 61.99%, the accuracy is greatly improved compared with the fault diagnosis by utilizing the single characteristic of air quantity, the actual requirement on the site can be met, and the intelligent management method for the mine ventilation safety is significant.
Claims (2)
1. A ventilation system resistance variation type fault diagnosis method based on an air volume and air pressure monitoring value is characterized by comprising the following steps:
step 1: preparing topological relation data of a mine ventilation network, wherein the topological relation data comprise original wind resistance and wind volume of each branch of the mine ventilation network and a fan characteristic curve of a fan, and a fan characteristic curve equation is specifically expressed as follows:
Hτ＝t0+t1Qτ+t2Qτ 2 (1)
in the formula, HτIndicating wind pressure, Q, of the ventilatorτIndicating the air volume, t, of the ventilator0,t1,t2Representing the wind pressure characteristic coefficient of the ventilator;
step 2: resolving through a mine ventilation network to obtain the wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network after the fault, and generating a simulation data fault sample set about resistance change fault-wind volume and wind pressure, wherein the fault sample set comprises the fault serial number and the fault volume of each branch of the mine ventilation network, the original wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network, and the wind resistance value, the wind volume and the wind pressure of each branch of the mine ventilation network after the fault occurs, and the specific expression is as follows:
step 2.1: defining an original air volume matrix of all branches of a mine ventilation network as Q, an original wind resistance matrix as R, an original wind pressure matrix as H, and recording as:
and satisfies the following conditions:
in the formula, qnRepresenting the original air volume, r, of the nth network branchnRepresenting the original wind resistance, h, of the nth network branchnIndicates the original wind pressure of the nth network branch, and B ═ Bij)m×nA complete correlation matrix, q, representing the mine ventilation networkjRepresenting the original air volume of the jth network branch, bijThe ith row and the jth column of the complete correlation matrix of the mine ventilation network are represented, m represents the number of nodes of the mine ventilation network, and C ═ Cij)s×nA loop matrix representing a mine ventilation network, cijThe ith row and the jth column of the loop matrix representing the mine ventilation network, s representing the total number of loops of the mine ventilation network, n representing the number of branches of the mine ventilation network, HTTranspose of the original wind pressure matrix H, HηAdditional resistance matrix, H, representing all loops of the mine ventilation networkη TAdditional resistance matrix H representing all loops of mine ventilation networkηTranspose of hjJ column, H, representing the original wind pressure matrix H of the mine ventilation networkηiLoop additive resistance matrix H representing mine ventilation networkηRow i of (1);
step 2.2: hypothesis branch eiA resistive type fault occurs with a fault amount of delta riCalculating branch e using equation (5)iThe wind resistance matrix R ' consisting of the wind resistances of all branches of the mine ventilation network after the fault is expressed as R ' ═ R '1,r′2,…,r′n}，
ri'＝ri±△ri，(i＝1,2,...,n) (5)
In the formula, ri' denotes the branch eiBranch windage after fault, riDenotes the branch eiOriginal wind resistance, Δ riDenotes the branch eiThe number of faults occurring, n representing the number of all branches of the mine ventilation network;
step 2.3: and solving the air volume of the mine ventilation network after the fault occurs through network calculation, and expressing an air volume matrix Q ' consisting of the air volumes of all branches of the mine ventilation network after the fault occurs as Q ' ═ Q '1,q′2,…,q′nThe network solution is expressed as:
Q'＝f(R') (6)
in the formula, Q 'represents the air volume of the mine ventilation network after the fault, R' represents the wind resistance of the mine ventilation network after the fault, and f (#) represents a function related to the wind resistance and the air volume;
step 2.4: solving the wind pressure of the mine ventilation network after the fault occurs by using the formula (7), and then a wind pressure matrix H ' consisting of the wind pressures of all branches of the mine ventilation network after the fault occurs is represented as H ' ═ H '1,h′2,…,h′n}，
Wherein h'iIndicating a post-fault branch eiWind pressure of ri' denotes the branch eiBranch windage after failure, q'iIndicating a post-fault branch eiThe flow rate of the mine ventilation network is determined by the flow state factor, x is 1 in laminar flow and 2 in complete turbulent flow, the transition state takes an intermediate value of 1-2, and n represents the number of all branches of the mine ventilation network;
step 2.5: calculating the wind resistance matrix R ', the wind volume matrix Q ' and the wind pressure matrix H ' of the mine ventilation network after the fault, and numbering the fault branch e when the fault occurs each timeiΔ r amount of failureiRecording original wind resistance values, wind volume and wind pressure of all branches of the mine ventilation network to a fault sample set;
and step 3: and respectively constructing a support vector machine classification model taking a monitoring fault position as a target and a support vector machine regression model taking a monitoring fault amount as a target according to the fault sample set, taking the air volume and the air pressure of each branch of the mine ventilation network after the fault occurs in the fault sample set as the input of the support vector machine, taking the number of each branch of the mine ventilation network in the fault sample set as the output of the support vector machine classification model, and taking the fault amount of each branch of the mine ventilation network in the fault sample set as the output of the support vector machine regression model, so that the diagnosis of the fault position and the fault amount is realized.
2. The method for diagnosing the fault of the resistance variation type of the ventilation system based on the wind volume and the wind pressure monitoring value as claimed in claim 1, wherein the step 3 is specifically expressed as:
step 3.1: constructing a support vector machine classification model aiming at monitoring a fault position by using the fault sample set, taking the air volume and the air pressure of each branch of the mine ventilation network after the fault occurs in the fault sample set as the input of the support vector machine classification model, taking the serial number of each branch of the mine ventilation network in the fault sample set as the output of the support vector machine classification model, introducing a symbolic function such as a formula (9), classifying the samples with the symbolic function value of 1 into one class, classifying the samples with the symbolic function value of-1 into another class, realizing the correct classification of the fault samples, and then realizing the classification of the branch e by using the formula (8)iDiagnosing the fault position;
where x denotes the input sample, sgn denotes the sign function, xiThe ith support vector of the support vector machine classification model is shown, l represents the number of the support vector machines,
representing the Lagrangian multiplier, yiRepresenting a category label, b*Threshold vector, f (x), representing the support vector machine kerneli) Representation for support vector xiThe sign function of (a);
step 3.2: constructing a regression model of a support vector machine with the aim of monitoring fault amount by using the fault sample set, taking the air volume and the air pressure of each branch of the mine ventilation network after the fault occurs in the fault sample set as the input of the regression model of the support vector machine, taking the fault amount of each branch of the mine ventilation network in the fault sample set as the output of the regression model of the support vector machine, adopting an epsilon-insensitive loss function with sparsity shown in a formula (10) to realize the diagnosis of the fault amount, namely realizing the regression of the samples in the fault sample set,
c(y,f*(x))＝max{0,|y-f*(x)|-ε}；ε>0 (10)
where y denotes the sample in the fault sample set, f*(x) Representing a hyperplane in the feature space, and epsilon represents regression precision;
step 3.3: and checking the generated fault position diagnosis classification model and the fault quantity diagnosis regression model, comparing the fault branch number of the fault position diagnosis classification model with the fault number of each branch in the fault sample set, comparing the wind resistance value of each branch after the fault output by the fault quantity diagnosis regression model with the wind resistance value of each branch after the fault in the fault sample set, and respectively drawing a scatter plot comparison chart to verify the accuracy of the output diagnosis of the fault diagnosis classification model and the regression model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910916393.3A CN110566259B (en) | 2019-09-26 | 2019-09-26 | Ventilation system resistance variation type fault diagnosis method based on air volume and air pressure monitoring value |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910916393.3A CN110566259B (en) | 2019-09-26 | 2019-09-26 | Ventilation system resistance variation type fault diagnosis method based on air volume and air pressure monitoring value |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110566259A CN110566259A (en) | 2019-12-13 |
| CN110566259B true CN110566259B (en) | 2021-04-13 |
Family
ID=68782419
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910916393.3A Active CN110566259B (en) | 2019-09-26 | 2019-09-26 | Ventilation system resistance variation type fault diagnosis method based on air volume and air pressure monitoring value |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110566259B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102650214A (en) * | 2012-05-31 | 2012-08-29 | 中煤科工集团重庆研究院 | Early warning method and system for on-line monitoring and analysis of mine ventilation system |
| CN102865093A (en) * | 2012-09-17 | 2013-01-09 | 太原理工大学 | Mine ventilation automatic detection and operation system |
| CN103266906A (en) * | 2013-05-03 | 2013-08-28 | 中国矿业大学 | Roadway wind resistance parameter self-adjusting verification method for mine ventilation system statuses |
| CN103603691A (en) * | 2013-11-28 | 2014-02-26 | 煤炭科学研究总院 | Mine ventilation dynamic resolving and analysis early warning method and system |
| JP2014221993A (en) * | 2013-05-14 | 2014-11-27 | 株式会社谷沢製作所 | Damper gear of mobile tunnel sound-insulating wall |
| CN105137233A (en) * | 2015-08-26 | 2015-12-09 | 西安科技大学 | Mine main fan safety hazard monitoring and early warning system and method |
| 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 |
| CN108266219A (en) * | 2018-01-23 | 2018-07-10 | 辽宁工程技术大学 | Mine ventilation system resistive-switching single fault source diagnostic method based on air quantity feature |
| CN109377810A (en) * | 2018-11-30 | 2019-02-22 | 中国矿业大学 | The monitoring of stope airflow parameter and automatic adjustment experiment porch and system |
| CN109446697A (en) * | 2018-11-08 | 2019-03-08 | 辽宁工程技术大学 | A kind of mine wind speed fault branch diagnostic method based on ELM |
-
2019
- 2019-09-26 CN CN201910916393.3A patent/CN110566259B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102650214A (en) * | 2012-05-31 | 2012-08-29 | 中煤科工集团重庆研究院 | Early warning method and system for on-line monitoring and analysis of mine ventilation system |
| CN102865093A (en) * | 2012-09-17 | 2013-01-09 | 太原理工大学 | Mine ventilation automatic detection and operation system |
| CN103266906A (en) * | 2013-05-03 | 2013-08-28 | 中国矿业大学 | Roadway wind resistance parameter self-adjusting verification method for mine ventilation system statuses |
| JP2014221993A (en) * | 2013-05-14 | 2014-11-27 | 株式会社谷沢製作所 | Damper gear of mobile tunnel sound-insulating wall |
| CN103603691A (en) * | 2013-11-28 | 2014-02-26 | 煤炭科学研究总院 | Mine ventilation dynamic resolving and analysis early warning method and system |
| CN105137233A (en) * | 2015-08-26 | 2015-12-09 | 西安科技大学 | Mine main fan safety hazard monitoring and early warning system and method |
| 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 |
| CN108266219A (en) * | 2018-01-23 | 2018-07-10 | 辽宁工程技术大学 | Mine ventilation system resistive-switching single fault source diagnostic method based on air quantity feature |
| CN109446697A (en) * | 2018-11-08 | 2019-03-08 | 辽宁工程技术大学 | A kind of mine wind speed fault branch diagnostic method based on ELM |
| CN109377810A (en) * | 2018-11-30 | 2019-02-22 | 中国矿业大学 | The monitoring of stope airflow parameter and automatic adjustment experiment porch and system |
Non-Patent Citations (1)
| Title |
|---|
| 基于SVM的通风系统故障诊断惩罚系数与核函数系数优化研究;周启超等;《中国安全生产科学技术》;20190430;第15卷(第4期);第45-51页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110566259A (en) | 2019-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108960303B (en) | Unmanned aerial vehicle flight data anomaly detection method based on LSTM | |
| CN112131760A (en) | CBAM model-based prediction method for residual life of aircraft engine | |
| CN105279365B (en) | For the method for the sample for learning abnormality detection | |
| CN109408552A (en) | The monitoring of the civil aircraft system failure and recognition methods based on LSTM-AE deep learning frame | |
| CN107725283A (en) | A kind of fan trouble detection method based on depth belief network model | |
| CN103983453A (en) | Differentiating method of fault diagnosis of executing mechanism of aeroengine and sensor | |
| CN107993012B (en) | Time-adaptive online transient stability evaluation method for power system | |
| CN105574587B (en) | A kind of online operating mode course monitoring method of plastic injection molding process | |
| CN113723010B (en) | Bridge damage early warning method based on LSTM temperature-displacement correlation model | |
| CN105606914A (en) | IWO-ELM-based Aviation power converter fault diagnosis method | |
| CN109491338A (en) | A kind of relevant method for diagnosing faults of multimode procedure quality based on sparse GMM | |
| CN105223241A (en) | A kind of compensation method of humidity sensor | |
| CN108390380B (en) | Method and system for predicting state parameter trend of transformer | |
| Bayer et al. | Beta regression control chart for monitoring fractions and proportions | |
| CN107688687A (en) | One kind considers time-length interrelation and the probabilistic life-span prediction method of part | |
| CN108266219A (en) | Mine ventilation system resistive-switching single fault source diagnostic method based on air quantity feature | |
| CN110566259B (en) | Ventilation system resistance variation type fault diagnosis method based on air volume and air pressure monitoring value | |
| CN114266278A (en) | Dual-attention-network-based method for predicting residual service life of equipment | |
| Aqlan et al. | Integrating artificial neural networks and cluster analysis to assess energy efficiency of buildings | |
| CN103559414B (en) | A kind of envirment factor the point estimation method based on Arrhenius relationship | |
| CN116894180B (en) | Product manufacturing quality prediction method based on different composition attention network | |
| CN109523082A (en) | A method of based on CNN-LSTM flight, normally clearance rate is predicted | |
| CN110705114B (en) | Ventilation fault diagnosis method without training sample | |
| CN105046089A (en) | Method for predicting strong rainfall and flood disasters | |
| CN107545112A (en) | Complex equipment Performance Evaluation and Forecasting Methodology of the multi-source without label data machine learning |
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 |