A real-time early warning method for early failure of mine hydraulic support system

Abstract

The present invention discloses a real-time early warning method for a mine hydraulic support system, which comprises two stages, namely, offline training and online monitoring. The offline training determines the threshold value of the health index of the hydraulic support system. In the online monitoring stage, the working parameters of the hydraulic support system are collected in real time, the Hotelling's statistic of the working parameters and the statistic are calculated and compared with the threshold value of the health index to determine whether the hydraulic support system has a fault. The abnormal change of the correlation between the working parameters of the hydraulic support system is used as the early fault feature of the hydraulic support system. The principal component analysis method is used to project the standardized sample data of the working parameters of the hydraulic support system into the principal component subspace and the residual subspace. The Hotelling's statistic and the statistic are respectively constructed on the two subspaces to quantitatively describe the degree of change of the correlation between the working parameters, which can effectively overcome the deficiency of the existing single variable threshold detection method that is insensitive to early faults.

Description

Technical Field

[0001] The invention belongs to the technical field of hydraulic support system fault detection, and in particular relates to a real-time early warning method for early faults of a mine hydraulic support system. Background Art

[0002] Support is necessary during mining. The hydraulic support system is the core equipment for comprehensive mechanized mining in mines. Once a failure occurs, it will cause the roof rock to fall, causing casualties and equipment damage, which is extremely harmful. To ensure the normal and continuous mining of mines, it is necessary to first ensure the safe and reliable operation of the hydraulic support system. However, the hydraulic support system has a complex structure and has been subjected to high loads and alternating stresses in mines for a long time. Its working safety faces serious threats. These reasons lead to a high failure rate of the hydraulic support system in actual use. For example, in a random inspection, Bailong Coal Mine found that 61 of the 96 hydraulic supports on a certain working face had emulsion leakage problems to varying degrees, with a failure rate of 64%. Timely diagnosis of faults in the hydraulic support system and ensuring the normal operation of the hydraulic support system are urgent needs for safe mining in mines.

[0003] The fault detection technology of hydraulic support system is a key technology in the field of hydraulic support fault diagnosis. The fault detection technology of hydraulic support system refers to the use of state observers, signal processing, or engineering experience to analyze and process the collected hydraulic support system state information to determine whether the current equipment operation has a fault. This technology is the prerequisite for implementing the fault type identification of hydraulic support system. The existing hydraulic support system fault detection technology in my country is essentially a single variable threshold detection method. This method compares the collected hydraulic support system working parameters with the pre-set threshold. When the pre-set threshold is exceeded, it is determined that the hydraulic support system has a fault, and then a fault alarm is issued. This method can detect mid- and late-stage faults in the hydraulic support system, but in the early stage of the fault, its characteristic information is weak, and the working parameters of the hydraulic support system often do not show a large deviation from the normal range, but still fluctuate within the normal range. The existing technology is difficult to deal with this situation, so it is impossible to detect and warn of early faults in the hydraulic support system.

[0004] After searching the literature on the prior art, it was found that the public document "A method for early warning of hydraulic support working conditions" (Chinese invention patent, publication number: CN112145231A, publication date: 2020.12.29) proposed a method for early warning of hydraulic support system fault detection. The public document describes itself as: "This application discloses an early warning method for hydraulic support working conditions, which relates to the technical field of safe mining of coal mines, wherein the method includes: real-time monitoring of the working resistance data of all hydraulic supports on the coal mine working face; determining the index values ​​of multiple evaluation indicators for evaluating the working conditions of hydraulic supports according to the working resistance change data in each coal mining cycle; generating early warning information according to the index values ​​of the evaluation indicators. Its technical solution is to monitor the working resistance data of all hydraulic supports on the coal mine working face in real time, and to determine multiple evaluation indicators for comprehensively evaluating the working conditions of hydraulic supports through analysis of the working resistance data of hydraulic supports during the mining process of the fully-mechanized mining working face of the coal mine, thereby realizing real-time monitoring, analysis and early warning of the working conditions of hydraulic supports, and timely locating problems existing in support support and roof management, which plays an important role in ensuring the safety of the roof of the fully-mechanized mining working face." Its shortcomings are: the "evaluation index" that describes the working status of the hydraulic support is obtained by comparing the measured values ​​of working parameters such as initial support force and average working resistance with preset standard values. It belongs to the single variable threshold detection method and can only realize mid- and late-stage fault detection. It cannot make fault warnings in the early stages of faults, so it is difficult to effectively avoid the harm caused by faults. Summary of the invention

[0005] To solve the above problems, the present invention proposes a real-time early warning method for early failures of mine hydraulic support systems, which is a real-time detection method for early failures of hydraulic support systems based on working parameter correlation monitoring; different from the existing univariate threshold detection method, this method takes the abnormal change of the correlation between working parameters as the early failure feature of the hydraulic support system, and realizes the detection and warning of early failures by real-time monitoring of the changes in the correlation between working parameters.

[0006] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0007] A real-time early warning method for early faults of a mine hydraulic support system further includes two stages: offline training and online monitoring, wherein the offline training determines the threshold value of the health index of the hydraulic support system, and the online monitoring stage collects the working parameters of the hydraulic support system in real time and calculates the Hotelling's T of the working parameters. 2 The statistics and Q statistics are compared with the threshold value of the health index to determine whether the hydraulic support system fails; including the following steps:

[0008] Offline training phase:

[0009] Step 1-1: Collect sample data of working parameters of the hydraulic support system in a healthy state;

[0010] Step 1-2: preprocessing the sample data of the working parameters of the hydraulic support system in the healthy state, removing the unit restriction of the sample data, converting it into a dimensionless pure value, and recording it as the working parameter in the healthy state after standardization;

[0011] Step 1-3: Calculate Hotelling's T of the working parameters under the standardized health state 2 The threshold values ​​of the statistic and the Q statistic are denoted as and Q UCL ; and Q UCL Determining a threshold value of a health indicator of the hydraulic support system;

[0012] Online monitoring stage:

[0013] Step 2-1: Real-time detection of working parameter data of the hydraulic support system;

[0014] Step 2-2: preprocessing the data of the real-time detection of the working parameters of the hydraulic support system, removing the unit restriction of the data of the working parameters, converting them into dimensionless pure values, and recording them as standardized real-time detection working parameters;

[0015] Step 2-3: projecting the standardized real-time detection working parameters into the principal component subspace and the residual subspace using principal component analysis;

[0016] Step 2-4: Construct Hotelling's T in the principal component subspace and the residual subspace respectively. 2 Statistics and Q statistics, denoted as T 2 and Q;

[0017] Step 2-5: Determine whether the hydraulic support system is faulty. or (Q>Q UCL ), it is determined that the hydraulic support system has a fault and an early warning signal is issued; otherwise, it is determined that the hydraulic support system has no fault.

[0018] Furthermore, in step 1-2, the pre-processing process uses the following formula:

[0019]

[0020] in, is the working parameter under normalized healthy state, X i It is the working parameter of the hydraulic support system in a healthy state; For Xi The average value of S Xi For X i The standard deviation of; different i values ​​represent different types of working parameters, i = 1, 2, 3, 4, 5, 6, 7...n; n is the total number of types of working parameters.

[0021] Furthermore, in steps 1-3, the Hotelling's T of the working parameters under the standardized health state is calculated. 2 The threshold values ​​of the statistic and the Q statistic are denoted as and Q LCL ; including the following steps:

[0022] Step 1-31: Calculate the covariance matrix S of the working parameters in the normalized healthy state. The calculation method is as follows:

[0023]

[0024] Where m is the number of sampling groups of working parameters under standardized healthy state; X * It is the sample data matrix of working parameters under the standardized healthy state;

[0025] Step 1-32: Calculate the eigenvalues ​​and eigenvectors of the covariance matrix S, and arrange them in descending order according to the size of the eigenvalues; the eigenvalue is denoted by λ i ; The eigenvector is denoted by p i ;i=1, 2, 3, 4, 5, 6, 7...n;

[0026] Step 1-33: Calculate the eigenvalue λ i Variance contribution rate CONT i And the cumulative variance contribution rate CPV(l) of the first l eigenvalues ​​is calculated as follows:

[0027]

[0028]

[0029] in, is the sum of n eigenvalues, j = 1, 2, 3, 4, 5, 6, 7...n; l ≤ n;

[0030] Step 1-34: Set the cumulative variance contribution rate threshold CPV according to the detection requirements UCL , through the formula CPV(k)>CPV UCL Determine the cutoff value k;

[0031] Step 1-35: Determine Hotelling's T 2 Thresholds for the statistic and Q statistic and Q UCL , the calculation formula is:

[0032]

[0033]

[0034] Among them, F α (k, mk) represents the critical value of the F distribution with two degrees of freedom, k and mk, at the significance level α; c α is the critical value of the standard normal distribution at the significance level α; the value of the significance level α is selected according to the detection requirements;

[0035] Among them, θ i The calculation method of and h0 is:

[0036]

[0037] Among them, θ i i=1, 2, 3, It is the i-th power of the j-th eigenvalue, i = 1, 2, 3, j = 1, 2, 3, 4, 5, 6, 7...n.

[0038] Furthermore, in step 2-2, the pre-processing process uses the following formula:

[0039]

[0040] in, To detect working parameters in real time after standardization; X 2i To detect the working parameters of the hydraulic support system in real time; For X i The average value of S Xi For X i The standard deviation of X i It is the working parameter of the hydraulic support system in a healthy state.

[0041] Furthermore, in step 2-3, the principal component analysis method is used to project the data of the standardized real-time detection working parameters into the principal component subspace and the residual subspace, including the following steps:

[0042] Step 2-31: According to the k value obtained in step 1-34, the feature vector p i Divided into and

[0043] Step 2-32: Wherein, using The normalized real-time detection working parameters are projected onto the principal element subspace using the basis vector; After the basis vector is normalized, the real-time detection working parameters are projected onto the residual subspace; the projection method is as follows:

[0044]

[0045]

[0046] in, It is the projection of the normalized real-time detection working parameters on the principal element subspace; It is the projection of the normalized real-time detection working parameters on the residual subspace; It is a column vector consisting of standardized real-time detection working parameters.

[0047] Furthermore, in steps 2-4, T 2 The calculation formula of Q is:

[0048]

[0049]

[0050] Among them, Λ is the diagonal matrix composed of the first k eigenvalues ​​of the covariance matrix S.

[0051] Furthermore, in step 1-1, the working parameters of the hydraulic support system in the healthy state include the pump outlet emulsion pressure, the pump outlet emulsion temperature, the emulsion pressure after the filter, the front pillar emulsion pressure, the front pillar emulsion temperature, the rear pillar emulsion pressure, the front extension beam emulsion pressure and the push emulsion pressure.

[0052] Furthermore, in step 2-1, the types of working parameters of the hydraulic support system detected in real time are the same as those in step 1-1, including pump outlet emulsion pressure, pump outlet emulsion temperature, emulsion pressure after the filter, front pillar emulsion pressure, front pillar emulsion temperature, rear pillar emulsion pressure, front extension beam emulsion pressure and push emulsion pressure.

[0053] The beneficial effects of the present invention are as follows:

[0054] Through the implementation of the above scheme, the abnormal change of the correlation between the working parameters of the hydraulic support system is taken as the early fault feature of the hydraulic support system. The principal component analysis method is used to project the data of the working parameters of the standardized hydraulic support system into the principal component subspace and the residual subspace. Among them, the principal component subspace reflects the main statistical characteristics of the working parameters, and the residual subspace reflects the statistical characteristics of the working parameters that are not described by the principal component subspace; Hotelling's T 2Statistics and Q statistics are used to quantitatively describe the degree of change of the correlation between working parameters. They can effectively overcome the deficiency of the existing single variable threshold detection method that is insensitive to early faults, and realize timely and effective detection of early faults of the hydraulic support system and issue early warning signals; in the early stage of the fault, the existence of the fault is detected in time, thereby guiding the maintenance personnel to implement preventive maintenance on the hydraulic support system and ensure that the hydraulic support system operates in the best working condition. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The present invention is a flow chart of a method for real-time early warning of early failure of a mine hydraulic support system.

[0056] Figure 2 The present invention is implemented in a pump-pipe-valve hydraulic system.

[0057] Figure 3 The effect of implementing the existing univariate threshold detection method on a pump-pipe-valve hydraulic system is shown. DETAILED DESCRIPTION

[0058] The specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings, but it should be understood that the protection scope of the present invention is not limited by the specific implementation modes.

[0059] like Figure 1 As shown, a real-time early warning method for early faults of a mine hydraulic support system includes two stages: offline training and online monitoring. The offline training determines the threshold value of the health index of the hydraulic support system. The online monitoring stage collects the working parameters of the hydraulic support system in real time and calculates the Hotelling's T of the working parameters. 2 The statistics and Q statistics are compared with the threshold value of the health index to determine whether the hydraulic support system has a fault; the steps include:

[0060] Offline training phase:

[0061] Step 1-1: Collect sample data of the working parameters of the hydraulic support system in a healthy state, denoted as X i Different i values ​​represent different types of working parameters, i=1, 2, 3, 4, 5, 6, 7...n; n is the total number of types of working parameters; preferably, the following 8 types of working parameters of the hydraulic support system are selected as the collection objects, that is, n=8, namely, pump outlet emulsion pressure X1, pump outlet emulsion temperature X2, filter rear emulsion pressure X3, front pillar emulsion pressure X4, front pillar emulsion temperature X5, rear pillar emulsion pressure X6, front extension beam emulsion pressure X7 and push emulsion pressure X8.

[0062] Step 1-2: Preprocess the sample data of the working parameters of the hydraulic support system in the healthy state so that it falls into a small specific interval to eliminate the influence of different dimensions of various working parameters on the fault detection results; remove the unit restriction of the sample data and convert it into a dimensionless pure value, which is recorded as the working parameter in the healthy state after standardization; the use formula is as follows:

[0063]

[0064] in, X is the working parameter under normalized healthy state; i It is the working parameter of the hydraulic support system in a healthy state; For X i The average values ​​of S Xi For X i The standard deviations are {S X1 , S X2 , S X3 , S X4 , S X5 , S X6 , S X7 , S X8 |.

[0065] Step 1-3: Calculate Hotelling's T for normalized health-related work parameters 2 The threshold values ​​of the statistic and the Q statistic are denoted as and Q UCL ; and Q UCL Determining the threshold value of the health indicator of the hydraulic support system; comprising the following steps:

[0066] Step 1-31: Calculate the covariance matrix S of the working parameters in the normalized healthy state. The calculation method is as follows:

[0067]

[0068] Where m is the number of sampling groups of working parameters under standardized healthy state; X * is the sample data matrix of the working parameters under the standardized healthy state, its rows represent the values ​​of the working parameters under the standardized healthy state at the same time, and its columns represent the values ​​of the working parameters at m different times;

[0069] Step 1-32: Calculate the eigenvalues ​​and eigenvectors of the covariance matrix S, and arrange them in descending order according to the size of the eigenvalues; the eigenvalue is denoted by λ i ; The eigenvector is denoted by p i; i=1, 2, 3, 4, 5, 6, 7, 8.

[0070] Step 1-33: Calculate the eigenvalue λ i Variance contribution rate CONT i And the cumulative variance contribution rate CPV(l) of the first l eigenvalues ​​is calculated as follows:

[0071]

[0072]

[0073] in, is the sum of 8 eigenvalues; l≤8;

[0074] Step 1-34: Set the cumulative variance contribution rate threshold CPV according to the detection requirements UCL , through the formula CPV(k)>CPV UCL Determine the cutoff value k;

[0075] Step 1-35: Determine Hotelling's T 2 Thresholds for the statistic and Q statistic and Q UCL , the calculation formula is:

[0076]

[0077]

[0078] Among them, F α (k, mk) represents the critical value of the F distribution with two degrees of freedom, k and mk, at the significance level α, and its value can be obtained by looking up the table; c α is the critical value of the standard normal distribution at the significance level α, and its value can be obtained by looking up the table; the value of the significance level α is selected according to the detection requirements;

[0079] Among them, θ i The calculation method of and h0 is:

[0080]

[0081] Among them, among them, θ i i=1, 2, 3; It is the i-th power of the j-th eigenvalue, i=1,2,3,j=1,2,3,4,5,6,7,8.

[0082] Online monitoring stage:

[0083] Step 2-1: Real-time detection of the working parameters of the hydraulic support system, denoted as X 2iDifferent i values ​​represent different types of working parameters, i=1, 2, 3, 4, 5, 6, 7...n; n is the total number of types of working parameters; the types of working parameters of the real-time detection of the hydraulic support system are the same as those in step 1-1. Preferably, the following 8 types of working parameters of the hydraulic support system are selected as the collection objects, that is, n=8, which are respectively, the pump outlet emulsion pressure X 21 , pump outlet emulsion temperature X 22 , emulsion pressure after filter X 23 , front support emulsion pressure X 24 , front pillar emulsion temperature X 25 , rear support emulsion pressure X 26 , emulsion pressure of the front beam X 27 and push emulsion pressure X 28 ; Store the collected working parameters of the hydraulic support system in the form of a column vector.

[0084] Step 2-2: Use the same preprocessing formula as in step 1-2, bring in the data of the working parameters of the hydraulic support system detected in real time in step 2-1, and preprocess the data, remove the unit restriction of the working parameter data, and convert it into a dimensionless pure value; the formula used is as follows:

[0085]

[0086] in, To detect working parameters in real time after standardization; X 2i To detect the working parameters of the hydraulic support system in real time; For X i The average value of S Xi For X i The standard deviation of .

[0087] Step 2-3: Projecting the standardized data of the working parameters of the hydraulic support system into the principal component subspace and the residual subspace using the principal component analysis method, including the following steps:

[0088] Step 2-31: According to the k value obtained in step 1-34, the feature vector p i Divided into and

[0089] Step 2-32: Wherein, using The normalized real-time detection working parameters are projected onto the principal element subspace using the basis vector; After the basis vector is normalized, the real-time detection working parameters are projected onto the residual subspace; the projection method is as follows:

[0090]

[0091]

[0092] in, It is the projection of the normalized real-time detection working parameters on the principal element subspace; It is the projection of the normalized real-time detection working parameters on the residual subspace; It is a column vector consisting of standardized real-time detection working parameters.

[0093] Step 2-4: Construct Hotelling's T in the principal element subspace and the residual subspace respectively 2 Statistics and Q statistics, denoted as T 2 and Q, are used to quantitatively describe the correlation between the working parameters of the hydraulic support system, and thus serve as an indicator for detecting early failures of the hydraulic support system; T 2 The calculation formula of Q is:

[0094]

[0095]

[0096] Among them, A is the diagonal matrix composed of the first k eigenvalues ​​of the covariance matrix S, and its calculation formula is:

[0097]

[0098] Step 2-5: Determine whether the hydraulic support system has a fault. If the following conditions are met, or (Q>Q UCL ) is detected, the hydraulic support system is judged to be faulty and a warning signal is issued; otherwise, it is determined that the hydraulic support system is not faulty.

[0099] like Figure 2 As shown in FIG. 1 , the implementation effect of the present invention on a pump-pipe-valve hydraulic system is shown, and the fault type diagnosed is pipeline leakage. Figure 3 As shown in the figure, the detection result of the fault using the single variable threshold detection method is shown in the figure. The dotted line indicates the alarm threshold. It can be seen that using the traditional threshold detection method, a considerable number of fault samples cannot be detected, and the fault detection rate is only 53.75%, which has a serious problem of missed detection. After using the correlation detection method, by constructing Hotelling's T 2 Statistics and Q statistics describe the changes in the correlation of working parameters, which increases the detection rate to 89%, a significant improvement.

[0100] The above embodiments are only for explaining the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the technology can modify or change the above embodiments without violating the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A real-time early warning method for early failure of a mine hydraulic support system, characterized in that: It includes two stages: offline training and online monitoring. The offline training determines the threshold value of the health index of the hydraulic support system. The online monitoring stage collects the working parameters of the hydraulic support system in real time and calculates the Hotelling's T of the working parameters. 2 The statistics and Q statistics are compared with the threshold value of the health index to determine whether the hydraulic support system fails; including the following steps: Offline training phase: Step 1-1: Collect sample data of working parameters of the hydraulic support system in a healthy state; Step 1-2: preprocessing the sample data of the working parameters of the hydraulic support system in the healthy state, removing the unit restriction of the sample data, converting it into a dimensionless pure value, and recording it as the working parameter in the healthy state after standardization; Step 1-3: Calculate Hotelling's T of the working parameters under the standardized health state 2 The threshold values ​​of the statistic and the Q statistic are denoted as and Q UCL ; and Q UCL Determining a threshold value of a health indicator of the hydraulic support system; Online monitoring stage: Step 2-1: Real-time detection of working parameter data of the hydraulic support system; Step 2-2: preprocessing the data of the real-time detection of the working parameters of the hydraulic support system, removing the unit restriction of the data of the working parameters, converting them into dimensionless pure values, and recording them as standardized real-time detection working parameters; Step 2-3: projecting the standardized real-time detection working parameters into the principal component subspace and the residual subspace using principal component analysis; Step 2-4: Construct Hotelling's T in the principal component subspace and the residual subspace respectively. 2 Statistics and Q statistics, denoted as T 2 and Q; Step 2-5: Determine whether the hydraulic support system is faulty. or Q>Q UCL When the hydraulic support system fails, it is determined that the hydraulic support system fails and a warning signal is issued; otherwise, it is determined that the hydraulic support system does not fail; In step 1-1, the working parameters of the hydraulic support system in the healthy state include the pump outlet emulsion pressure, the pump outlet emulsion temperature, the emulsion pressure after the filter, the front support emulsion pressure, the front support emulsion temperature, the rear support emulsion pressure, the front extension beam emulsion pressure and the push emulsion pressure; In step 2-1, the types of working parameters of the hydraulic support system detected in real time are the same as those in step 1-1, including pump outlet emulsion pressure, pump outlet emulsion temperature, emulsion pressure after the filter, front pillar emulsion pressure, front pillar emulsion temperature, rear pillar emulsion pressure, front extension beam emulsion pressure and push emulsion pressure.

2. The method for early warning of hydraulic support system failure in mines according to claim 1 is characterized in that: In step 1-2, the pre-processing process uses the following formula: in, is the working parameter under normalized healthy state, x i It is the working parameter of the hydraulic support system in a healthy state; For X i The average value of S Xi For X i The standard deviation of the working parameters; different i values ​​represent different types of working parameters, i = 1, 2, 3, 4, 5, 6, 7…n; n is the total number of types of working parameters.

3. The method for early warning of hydraulic support system failure in mines according to claim 2 is characterized in that: In steps 1-3, calculate Hotelling's T of the working parameters under the standardized health state 2 The threshold values ​​of the statistic and the Q statistic are denoted as and Q UCL ; The following steps are involved: Step 1-31: Calculate the covariance matrix S of the working parameters in the normalized healthy state. The calculation method is as follows: Wherein, m is the number of sampling groups of working parameters under the standardized healthy state; X* is the sample data matrix of working parameters under the standardized healthy state; Step 1-32: Calculate the eigenvalues ​​and eigenvectors of the covariance matrix S, and arrange them in descending order according to the size of the eigenvalues; the eigenvalue is denoted by λ i ; The eigenvector is denoted by p i ; i = 1, 2, 3, 4, 5, 6, 7 ... n; Step 1-33: Calculate the eigenvalue λ i Variance contribution rate CONT i And the cumulative variance contribution rate CPV(l) of the first l eigenvalues ​​is calculated as follows: in, is the sum of n eigenvalues, j = 1, 2, 3, 4, 5, 6, 7…n; l ≤ n; Step 1-34: Set the cumulative variance contribution rate threshold CPV according to the detection requirements UCL , through the formula CPV(k)>CPV UCL Determine the cutoff value k; Step 1-35: Determine Hotelling's T 2 Thresholds for the statistic and Q-statistic and Q UCL , the calculation formula is: Among them, F α (k,mk) represents the critical value of the F distribution with two degrees of freedom, k and mk, at the significance level α; c α is the critical value of the standard normal distribution at the significance level α; the value of the significance level α is selected according to the detection requirements; Among them, θ i The calculation method of and h0 is: Among them, θ i where i=1,2,3, It is the i-th power of the j-th eigenvalue, i=1,2,3, j=1,2,3,4,5,6,7…n.

4. The method for early warning of hydraulic support system failure in mines according to claim 3 is characterized in that: In step 2-2, the pre-processing process uses the following formula: in, To detect working parameters in real time after standardization; X 2i To detect the working parameters of the hydraulic support system in real time; For X i The average value of S Xi For X i The standard deviation of X i It is the working parameter of the hydraulic support system in a healthy state.

5. The real-time early warning method for early failure of a mine hydraulic support system according to claim 4 is characterized in that: In step 2-3, the principal component analysis method is used to project the data of the normalized real-time detection working parameters into the principal component subspace and the residual subspace, including the following steps: Step 2-31: According to the k value obtained in step 1-34, the feature vector p i Divided into and Step 2-32: Wherein, using The normalized real-time detection working parameters are projected onto the principal element subspace using the basis vector; After the basis vector is normalized, the real-time detection working parameters are projected onto the residual subspace; the projection method is as follows: in, It is the projection of the normalized real-time detection working parameters on the principal element subspace; It is the projection of the normalized real-time detection working parameters on the residual subspace; It is a column vector consisting of standardized real-time detection working parameters.

6. The real-time early warning method for early failure of a mine hydraulic support system according to claim 5 is characterized in that: In steps 2-4, T 2 The calculation formula of and Q is: Among them, Λ is the diagonal matrix composed of the first k eigenvalues ​​of the covariance matrix S.

Images