Gas-liquid linkage valve fault diagnosis method and system and computer readable medium
By constructing a feature matrix and a fault diagnosis model, the problem of relying on disassembly inspection for fault diagnosis of traditional pneumatic-hydraulic linkage valves is solved, achieving efficient and accurate fault diagnosis and avoiding reliance on human experience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- PETROCHINA CO LTD
- Filing Date
- 2024-11-15
- Publication Date
- 2026-05-15
AI Technical Summary
Traditional methods for diagnosing gas-liquid linkage valve faults require frequent disassembly and inspection, rely on manual experience, make it difficult to accurately locate faults, and affect the safe operation of gas pipelines.
Based on the operating information of the gas-liquid linkage valve, a feature matrix is constructed. The pre-constructed fault diagnosis model is then used for fault diagnosis. Through data acquisition, preprocessing, feature extraction, and feature matrix construction, disassembly inspection is avoided, thereby improving diagnostic efficiency and accuracy.
It enables efficient fault diagnosis without disassembly, improving the accuracy and efficiency of fault diagnosis and reducing reliance on human experience.
Smart Images
Figure CN122045908A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of fault diagnosis technology, and specifically to a fault diagnosis method, system, and computer-readable medium for a pneumatic-hydraulic linkage valve. Background Technology
[0002] As a key automatic shut-off valve in gas transmission stations, the gas-liquid linkage ball valve plays an important role in the safe operation and protection of gas transmission pipelines.
[0003] The pneumatic-hydraulic ball valve mainly realizes the functions of shutting off and opening the ball valve, which are controlled by four solenoid valves respectively. The solenoid valves for pipe rupture, remote opening, and remote closing are normally closed solenoid valves, which are not energized under normal circumstances, but are energized in special or emergency situations; the ESD solenoid valve is a normally open solenoid valve, whose coil is continuously energized under normal circumstances, and is de-energized and closes the valve in an emergency.
[0004] As a core component of the pipe rupture system and emergency shut-off system in large-diameter long-distance natural gas pipelines, the gas-hydraulic ball valve has the highest safety level among all gas pipeline valves and has a significant impact on safe production operations, naturally making it a key focus of maintenance. Traditional gas pipeline valve maintenance primarily relies on "regular inspections and reactive maintenance." This approach lacks real-time capability, leading to delayed valve fault diagnosis. This can result in minor issues like equipment malfunctions or failures, or even serious safety accidents. Furthermore, due to the complex fault modes and causes of gas-hydraulic ball valves, routine maintenance often fails to detect early, minor faults, missing the optimal period for low-cost intervention. Summary of the Invention
[0005] The technical problem to be solved by this invention is that traditional methods for diagnosing faults in gas-hydraulic linkage valves require frequent disassembly and inspection, which not only affects the safe operation of gas pipelines but also relies too heavily on manual experience, making it difficult to accurately locate faults. The purpose of this invention is to provide a method, system, and computer-readable medium for diagnosing faults in gas-hydraulic linkage valves. Based on the operating information of the gas-hydraulic linkage valve, a feature matrix is constructed and input into a pre-constructed fault diagnosis model for fault diagnosis, avoiding disassembly and inspection and improving fault diagnosis efficiency.
[0006] This invention is achieved through the following technical solution:
[0007] This solution provides a method for diagnosing faults in a pneumatic-hydraulic linkage valve, including:
[0008] Step 1: Collect the operating information of the pneumatic-hydraulic linkage valve and preprocess the operating information;
[0009] Step 2: Extract features based on the preprocessed runtime information, and construct a feature matrix based on the extracted features;
[0010] Step 3: Input the feature matrix into the constructed fault diagnosis model for fault diagnosis;
[0011] Step 4: Output the fault diagnosis results.
[0012] The working principle of this solution is as follows: Traditional methods for diagnosing faults in gas-hydraulic linkage valves require frequent disassembly and inspection, which not only affects the safe operation of gas pipelines but also relies too heavily on manual experience, making it difficult to accurately locate faults. The purpose of this invention is to provide a method, system, and computer-readable medium for diagnosing faults in gas-hydraulic linkage valves. Based on the operating information of the gas-hydraulic linkage valve, a feature matrix is constructed and input into a pre-constructed fault diagnosis model for fault diagnosis, avoiding disassembly and inspection and improving fault diagnosis efficiency.
[0013] A further optimized solution is that the operating information includes: gas tank pressure, gas-liquid tank top pressure, valve position, oil cylinder pressure, gas-liquid tank liquid level, solenoid valve current, and solenoid valve voltage.
[0014] A further optimization scheme is that the preprocessing method includes: cleaning abnormal data in the operation information and fitting data points in the abnormal data domain to eliminate the abnormal data.
[0015] Step two includes the following process:
[0016] Time-domain feature parameters are extracted from the preprocessed running information. The spectral features include: maximum value, minimum value, average value, median, peak-to-peak value, absolute average value, variance, standard deviation, kurtosis, skewness, root mean square, mean square value, root square amplitude, waveform factor, peak factor, impulse factor, and margin factor.
[0017] Spectral feature parameters are extracted from the preprocessed running information. The spectral features include: centroid frequency, mean square frequency, root mean square frequency, frequency variance, and frequency standard deviation.
[0018] Entropy-related indices are calculated based on the preprocessed operational information. These entropy-related indices include: power spectral entropy, energy entropy, and envelope entropy.
[0019] To optimize the spectral feature parameters, a data space feature is constructed based on the time-domain feature parameters, the optimized spectral feature parameters, and entropy-related indicators.
[0020] A further optimization scheme is to optimize the spectral feature parameters in the data space of the constructed spectral feature parameters, including the following methods:
[0021] Fourier transforms were performed using the centroid frequency, mean square frequency, and root mean square frequency as the transform function f(t), respectively:
[0022]
[0023] In the formula, F(n) represents the frequency domain representation of the transform function f(t); e -int Represents a complex exponential function;
[0024] Based on the Fourier transform results, the frequency spectral densities P of the centroid frequency, mean square frequency, and root mean square frequency were calculated respectively. i i = centroid frequency, mean square frequency, or root mean square frequency;
[0025] Spatial characteristics of the centroid frequency, mean square frequency, and root mean square frequency calculated based on frequency spectral density:
[0026]
[0027] The original centroid frequency, mean square frequency, and root mean square frequency are replaced with the calculated data spatial characteristics.
[0028] A further optimized scheme involves calculating the frequency spectral density of the centroid frequency, mean square frequency, and root mean square frequency based on the Fourier transform results; including the following method:
[0029] The frequency spectral density is calculated based on the following formula:
[0030]
[0031] Among them, F * (n) is the conjugate of F(n), and N represents the signal length of F(n).
[0032] A further optimized solution includes the following step before inputting the feature matrix into the pre-constructed fault diagnosis model for fault diagnosis:
[0033] T1, normalize the feature matrix so that each element in the feature matrix is mapped to the range of -1 to 1;
[0034] In step T2, the n-dimensional feature matrix obtained in step T1 is mapped to a linearly separable high-dimensional feature space, and then principal component analysis is performed to reduce the dimensionality of the high-dimensional feature space. During the principal component analysis dimensionality reduction process in the high-dimensional feature space, the calculation is simplified based on the kernel trick.
[0035] The further optimized solution is that step T2 includes the following process:
[0036] The dataset X = {x1, x2, ..., xn} corresponds to the n-dimensional feature matrix. n} contains n samples, where x n ∈R N ;
[0037] A nonlinear mapping function Φ is introduced to project the dataset X onto a high-dimensional feature space Y;
[0038] In a high-dimensional feature space Y, the data mappings in dataset X need to satisfy the centrality requirement, that is:
[0039] The covariance matrix C of the high-dimensional feature space Y is:
[0040] The eigenvalue expression for the covariance matrix C is: λν=Cν; where λ represents the eigenvalue and ν represents the eigenvector.
[0041] The projection of sample Φ(x) into the high-dimensional feature space is described as follows:
[0042]
[0043] In the formula h is the k-th eigenvector. k (x) is the k-th nonlinear principal component, K(x) i (x) represents the kernel function, and x represents the variable.
[0044] A further optimization scheme involves obtaining the eigenvector ν and eigenvalue λ using the following methods:
[0045] The eigenvector ν is derived from Φ(x) i Linear expressions include: In the formula α i It is a constant factor;
[0046] Then λΦ(x) i )ν=Φ(x i )Cν; Establish kernel function K=Φ(x i )·Φ(x j ),have:
[0047] nλα=Kα
[0048] Solving for nλα=Kα yields the eigenvector ν and eigenvalue λ.
[0049] This solution also provides a fault diagnosis system for a pneumatic-hydraulic linkage valve, used to implement the aforementioned fault diagnosis method for a pneumatic-hydraulic linkage valve; the system includes:
[0050] The data acquisition module is used to acquire the operating information of the gas-liquid linkage valve and to preprocess the operating information.
[0051] The feature extraction module is used to extract features based on the preprocessed runtime information and construct a feature matrix based on the extracted features;
[0052] The diagnostic module is used to input the feature matrix into the pre-constructed fault diagnosis model for fault diagnosis.
[0053] The output module is used to output the fault diagnosis results.
[0054] This solution also provides a computer-readable medium storing a computer program, which, when executed by a processor, can implement a method for determining oil circuit blockage in a pneumatic-hydraulic linkage valve as described above.
[0055] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0056] This invention provides a method, system, and computer-readable medium for diagnosing faults in a pneumatic-hydraulic linkage valve. A feature matrix is constructed based on the valve's operating information, and the feature matrix is input into a pre-constructed fault diagnosis model for fault diagnosis, avoiding disassembly inspection and improving fault diagnosis efficiency.
[0057] This invention provides a method, system, and computer-readable medium for diagnosing faults in a pneumatic-hydraulic linkage valve. The constructed feature matrix includes time-domain feature parameters, spectral feature parameters, and entropy-related indicators. Feature parameters of the acquired signals are extracted from multiple dimensions to form the feature matrix, thereby improving the accuracy of fault diagnosis.
[0058] The present invention provides a method, system, and computer-readable medium for diagnosing faults in a pneumatic-hydraulic linkage valve. A feature matrix is constructed based on the operating information of the pneumatic-hydraulic linkage valve and input into a pre-constructed fault diagnosis model for fault diagnosis, which can avoid disassembly inspection and improve fault diagnosis efficiency. The constructed feature matrix includes time-domain feature parameters, spectral feature parameters, and entropy-related indicators. Feature parameters of the acquired signals are extracted from multiple dimensions to form the feature matrix, which can effectively improve the accuracy of fault diagnosis.
[0059] The present invention provides a method, system and computer-readable medium for diagnosing faults in a gas-liquid linkage valve. Before performing principal component analysis to reduce the dimensionality of the data, this solution constructs data space features of spectral feature parameters to optimize the spectral feature parameters. After optimizing the spectral feature parameters, the principal component dimensionality reduction method is used to remove redundant features, reducing the data to a low dimension. This effectively solves the problem of class overlap and redundant features, and speeds up the calculation. Attached Figure Description
[0060] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:
[0061] Figure 1 A schematic diagram of the fault diagnosis method for a pneumatic-hydraulic linkage valve;
[0062] Figure 2 This is a schematic diagram of a fault diagnosis system for a pneumatic-hydraulic linkage valve.
[0063] Figure 3 This is a schematic diagram illustrating the fault diagnosis effect of the gas-liquid linkage valve in Example 3. Detailed Implementation
[0064] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.
[0065] As a critical automatic shut-off valve in gas transmission stations, the pneumatic-hydraulic linkage ball valve plays a vital role in the safe operation and protection of gas pipelines. Traditional methods for diagnosing pneumatic-hydraulic linkage valve faults require frequent disassembly and inspection, which not only affects the safe operation of gas pipelines but also relies too heavily on manual experience, making it difficult to accurately locate faults. Therefore, this solution provides the following embodiments to address the problems in the existing technology:
[0066] Example 1
[0067] This embodiment provides a method for diagnosing faults in a pneumatic-hydraulic linkage valve, such as... Figure 1 As shown, it includes:
[0068] Step 1: Collect the operating information of the pneumatic-hydraulic linkage valve and preprocess the operating information;
[0069] Step 2: Extract features based on the preprocessed runtime information, and construct a feature matrix based on the extracted features;
[0070] Step 3: Input the feature matrix into the constructed fault diagnosis model for fault diagnosis;
[0071] Step 4: Output the fault diagnosis results.
[0072] The operational information includes: gas tank pressure, gas-liquid tank top pressure, valve position, oil cylinder pressure, gas-liquid tank liquid level, solenoid valve current, and solenoid valve voltage.
[0073] The preprocessing method includes: cleaning abnormal data in the operation information and fitting data points in the abnormal data domain to eliminate the abnormal data.
[0074] Step two includes the following process:
[0075] Time-domain feature parameters are extracted from the preprocessed running information. The spectral features include: maximum value, minimum value, average value, median, peak-to-peak value, absolute average value, variance, standard deviation, kurtosis, skewness, root mean square, mean square value, root square amplitude, waveform factor, peak factor, impulse factor, and margin factor.
[0076] Spectral feature parameters are extracted from the preprocessed running information. The spectral features include: centroid frequency, mean square frequency, root mean square frequency, frequency variance, and frequency standard deviation.
[0077] Entropy-related indices are calculated based on the preprocessed operational information. These entropy-related indices include: power spectral entropy, energy entropy, and envelope entropy.
[0078] To optimize the spectral feature parameters, a data space feature is constructed based on the time-domain feature parameters, the optimized spectral feature parameters, and entropy-related indicators.
[0079] A further optimization scheme is to optimize the spectral feature parameters in the data space of the constructed spectral feature parameters, including the following methods:
[0080] Fourier transforms were performed using the centroid frequency, mean square frequency, and root mean square frequency as the transform function f(t), respectively:
[0081]
[0082] In the formula, F(n) represents the frequency domain representation of the transform function f(t); e -int Represents a complex exponential function;
[0083] Based on the Fourier transform results, the frequency spectral densities P of the centroid frequency, mean square frequency, and root mean square frequency were calculated respectively. i i = centroid frequency, mean square frequency, or root mean square frequency;
[0084] Spatial characteristics of the centroid frequency, mean square frequency, and root mean square frequency calculated based on frequency spectral density:
[0085]
[0086] The original centroid frequency, mean square frequency, and root mean square frequency are replaced with the calculated data spatial characteristics.
[0087] A further optimized scheme involves calculating the frequency spectral density of the centroid frequency, mean square frequency, and root mean square frequency based on the Fourier transform results; including the following method:
[0088] The frequency spectral density is calculated based on the following formula:
[0089]
[0090] Among them, F * (n) is the conjugate of F(n), and N represents the signal length of F(n).
[0091] Since the differences in the distribution of data across different categories are not significant, the correlation between the data features obtained after principal component analysis (PCA) dimensionality reduction is extremely high, easily leading to feature redundancy. Therefore, this approach constructs data space features of spectral feature parameters to optimize these parameters before performing PCA dimensionality reduction. After optimizing the spectral feature parameters, PCA dimensionality reduction is then used to remove redundant features, reducing the data to a lower dimension. This effectively solves the problems of class overlap and redundant features, and also speeds up the computation.
[0092] In the specific implementation process, the feature matrix constructed in this embodiment includes time-domain feature parameters, spectral feature parameters, and entropy-related indicators. It extracts feature parameters of the acquired signal from multiple dimensions to form a feature matrix, thereby improving the accuracy of fault diagnosis.
[0093] Before inputting the feature matrix into the constructed fault diagnosis model for fault diagnosis, the following steps are also included:
[0094] T1, normalize the feature matrix so that each element in the feature matrix is mapped to the range of -1 to 1; considering the influence of dimensions, this embodiment normalizes the feature matrix so that the data in the feature matrix is mapped to the range of -1 to 1.
[0095] T2 involves dimensionality reduction of the feature matrix obtained in step T1. Each acquired operational information signal corresponds to a 29-dimensional feature matrix. Under multiple sensors, this will result in a high dimensionality of the feature matrix. Therefore, this scheme uses the kpca (kernel principal component analysis) method to reduce the dimensionality of the feature matrix.
[0096] Specific step T2 includes the following process: mapping the n-dimensional feature matrix to a linearly separable high-dimensional feature space, and then performing principal component analysis (PCA) to reduce the dimensionality of the high-dimensional feature space; simplifying the calculation based on the kernel trick during the PCA dimensionality reduction process in the high-dimensional feature space: the dataset X corresponding to the n-dimensional feature matrix is X = {x1, x2, ..., x...}. n} contains n samples, where x n ∈R N ;
[0097] A nonlinear mapping function Φ is introduced to project the dataset X onto a high-dimensional feature space Y;
[0098] In a high-dimensional feature space Y, the data mappings in dataset X need to satisfy the centrality requirement, that is:
[0099] The covariance matrix C of the high-dimensional feature space Y is:
[0100] The eigenvalue expression for the covariance matrix C is: λν=Cν; where λ represents the eigenvalue and ν represents the eigenvector.
[0101] The projection of sample Φ(x) into the high-dimensional feature space is described as follows:
[0102]
[0103] In the formula h is the k-th eigenvector. k (x) is the k-th nonlinear principal component, K(x) i (x) represents the kernel function, and x represents the variable.
[0104] Methods for obtaining eigenvectors ν and eigenvalues λ include:
[0105] The eigenvector ν is derived from Φ(x) i Linear expressions include: In the formula α i It is a constant factor;
[0106] Then λΦ(x) i )ν=Φ(x i )Cν; Establish kernel function K=Φ(x i )·Φ(x j ),have:
[0107] nλα=Kα
[0108] nλα=Kα
[0109] Solving for nλα=Kα yields the eigenvector ν and eigenvalue λ.
[0110] This embodiment provides a fault diagnosis method for a pneumatic-hydraulic linkage valve. A feature matrix is constructed based on the operating information of the pneumatic-hydraulic linkage valve, and the feature matrix is input into a pre-constructed fault diagnosis model for fault diagnosis, avoiding disassembly inspection and improving fault diagnosis efficiency.
[0111] Example 2
[0112] This embodiment provides a fault diagnosis system for a pneumatic-hydraulic linkage valve, used to implement the fault diagnosis method for a pneumatic-hydraulic linkage valve described in Embodiment 1; as follows: Figure 2 As shown, the system includes:
[0113] The data acquisition module is used to acquire the operating information of the gas-liquid linkage valve and to preprocess the operating information.
[0114] The feature extraction module is used to extract features based on the preprocessed runtime information and construct a feature matrix based on the extracted features;
[0115] The diagnostic module is used to input the feature matrix into the pre-constructed fault diagnosis model for fault diagnosis.
[0116] The output module is used to output the fault diagnosis results.
[0117] Example 3
[0118] This embodiment provides a computer-readable medium storing a computer program. The computer program, when executed by a processor, can implement a method for determining oil circuit blockage in a pneumatic-hydraulic linkage valve as described in Embodiment 1. The computer program performs the following steps:
[0119] Step 1: Collect the operating information of the pneumatic-hydraulic linkage valve and preprocess the operating information;
[0120] Step 2: Extract features based on the preprocessed runtime information, and construct a feature matrix based on the extracted features;
[0121] Step 3: Input the feature matrix into the constructed fault diagnosis model for fault diagnosis;
[0122] Step 4: Output the fault diagnosis results.
[0123] In specific implementation, this embodiment uses LSTM (Long Short-Term Memory) network to classify faults in each sample composed of feature matrices and labels. The samples are divided into a test set and a training set in a 3:1 ratio. The test set is used for neural network training, and the training set is used to verify the model's prediction accuracy. The fault diagnosis effect of the pneumatic-hydraulic actuator is as follows: Figure 3 As shown, the diagnostic accuracy is as high as 99.28%. Specifically, the Long Short-Term Memory (LSTM) network in this embodiment outputs the following results (including fault type and fault cause):
[0124] Solenoid valve electrical fault, causes of fault: ① coil open circuit, ② coil short circuit, ③ voltage too low / too high;
[0125] Mechanical failure of solenoid valve, causes of failure: ① broken spring, ② valve core stuck by foreign object, ③ damaged sealing gasket;
[0126] Oil circuit blockage, possible causes: ① oil line blockage, ② hydraulic filter blockage, ③ flow control valve blockage.
[0127] Gas line blockage, causes of failure: ① gas filter blockage, ② solenoid valve, pilot valve and / or shuttle valve inlet and outlet blockage, ③ gas line blockage.
[0128] Gas leak, causes of failure: ① Loose gas pipeline and connector connections; ② Failure of solenoid valve, pilot valve, shuttle valve seals; ③ Solenoid valve, pilot valve, shuttle valve cores stuck by foreign objects; ④ Broken solenoid valve spring.
[0129] Oil leakage, the causes of the malfunction are: ① loose connection between oil line and joint, ② failure of piston seal;
[0130] Abnormal torque, possible causes: ① valve body jamming, ② damage to the actuator transmission device.
[0131] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for diagnosing faults in a pneumatic-hydraulic linkage valve, characterized in that, include: Step 1: Collect the operating information of the pneumatic-hydraulic linkage valve and preprocess the operating information; Step 2: Extract features based on the preprocessed runtime information, and construct a feature matrix based on the extracted features; Step 3: Input the feature matrix into the constructed fault diagnosis model for fault diagnosis; Step 4: Output the fault diagnosis results.
2. The method for diagnosing faults in a pneumatic-hydraulic linkage valve according to claim 1, characterized in that, The operational information includes: gas tank pressure, gas-liquid tank top pressure, valve position, oil cylinder pressure, gas-liquid tank liquid level, solenoid valve current, and solenoid valve voltage.
3. The method for diagnosing faults in a pneumatic-hydraulic linkage valve according to claim 2, characterized in that, The preprocessing method includes: cleaning abnormal data in the operation information and fitting data points in the abnormal data domain to eliminate the abnormal data.
4. The method for diagnosing faults in a pneumatic-hydraulic linkage valve according to claim 1, characterized in that, Step two includes the following process: Time-domain feature parameters are extracted from the preprocessed running information. The time-domain features include: maximum value, minimum value, average value, median, peak-to-peak value, absolute average value, variance, standard deviation, kurtosis, skewness, root mean square, mean square value, root square amplitude, waveform factor, peak factor, impulse factor, and margin factor. Spectral feature parameters are extracted from the preprocessed running information. The spectral features include: centroid frequency, mean square frequency, root mean square frequency, frequency variance, and frequency standard deviation. Entropy-related indices are calculated based on the preprocessed operational information. These entropy-related indices include: power spectral entropy, energy entropy, and envelope entropy. To optimize the spectral feature parameters, a data space feature is constructed based on the time-domain feature parameters, the optimized spectral feature parameters, and entropy-related indicators.
5. The method for diagnosing faults in a pneumatic-hydraulic linkage valve according to claim 4, characterized in that, The method for constructing spectral feature parameters by optimizing the data space of the spectral feature parameters includes: Fourier transforms were performed using the centroid frequency, mean square frequency, and root mean square frequency as the transform function f(t), respectively: In the formula, F(n) represents the frequency domain representation of the transform function f(t); e -int Represents a complex exponential function; Based on the Fourier transform results, the frequency spectral densities P of the centroid frequency, mean square frequency, and root mean square frequency were calculated respectively. i i = centroid frequency, mean square frequency, or root mean square frequency; Spatial characteristics of the centroid frequency, mean square frequency, and root mean square frequency calculated based on frequency spectral density: The original centroid frequency, mean square frequency, and root mean square frequency are replaced with the calculated data spatial characteristics.
6. The method for diagnosing faults in a pneumatic-hydraulic linkage valve according to claim 5, characterized in that, The frequency spectral density of the centroid frequency, mean square frequency, and root mean square frequency is calculated based on the Fourier transform results, including the following method: The frequency spectral density is calculated based on the following formula: Among them, F * (n) is the conjugate of F(n), and N represents the signal length of F(n).
7. The method for diagnosing faults in a pneumatic-hydraulic linkage valve according to claim 4, characterized in that, Before inputting the feature matrix into the constructed fault diagnosis model for fault diagnosis, the following steps are also included: T1, normalize the feature matrix so that each element in the feature matrix is mapped to the range of -1 to 1; T2, map the n-dimensional feature matrix obtained in step T1 to a linearly separable high-dimensional feature space, and then perform principal component analysis (PCA) to reduce the dimensionality of the high-dimensional feature space; in the PCA dimensionality reduction process of the high-dimensional feature space, the kernel trick is used to simplify the calculation: the dataset X corresponding to the n-dimensional feature matrix is X = {x1, x2, ..., x...}. n } contains n samples, where x n ∈R N ; A nonlinear mapping function Φ is introduced to project the dataset X onto a high-dimensional feature space Y; In a high-dimensional feature space Y, the data mappings in dataset X need to satisfy the centrality requirement, that is: The covariance matrix C of the high-dimensional feature space Y is: The eigenvalue expression for the covariance matrix C is: λν=Cν; where λ represents the eigenvalue and ν represents the eigenvector. The projection of sample Φ(x) into the high-dimensional feature space is described as follows: In the formula h is the k-th eigenvector. k (x) is the k-th nonlinear principal component, K(x) i (x) represents the kernel function, and x represents the variable.
8. The method for diagnosing faults in a pneumatic-hydraulic linkage valve according to claim 7, characterized in that, Methods for obtaining eigenvectors ν and eigenvalues λ include: The eigenvector ν is derived from Φ(x) i Linear expressions include: In the formula α i It is a constant factor; Then λΦ(x) i )ν=Φ(x i )Cν; Establish kernel function K=Φ(x i )·Φ(x j ),have: nλα=Kα Solving for nλα=Kα yields the eigenvector ν and eigenvalue λ.
9. A fault diagnosis system for a pneumatic-hydraulic linkage valve, characterized in that, A method for diagnosing faults in a pneumatic-hydraulic linkage valve as described in any one of claims 1-8; the system comprises: The data acquisition module is used to acquire the operating information of the gas-liquid linkage valve and to preprocess the operating information. The feature extraction module is used to extract features based on the preprocessed runtime information and construct a feature matrix based on the extracted features; The diagnostic module is used to input the feature matrix into the pre-constructed fault diagnosis model for fault diagnosis. The output module is used to output the fault diagnosis results.
10. A computer-readable medium having a computer program stored thereon, characterized in that, The computer program, when executed by a processor, can implement a method for determining oil circuit blockage in a pneumatic-hydraulic linkage valve as described in any one of claims 1-8.