Method for predicting service life of sealing structure of aviation hydraulic pipeline system

Abstract

The invention discloses a service life prediction method for a sealing structure of an aviation hydraulic pipeline system, and relates to the technical field of service life prediction of sealing structures, and the method comprises the following steps: constructing different service life test data sets, and carrying out clustering analysis according to environmental parameters; pre-training one machine learning model based on each cluster of life test clustering data set; environment parameter data and state parameter data are collected; calculating the distance between the environment parameter data and the environment parameter clustering center, and selecting at least two life prediction models for life prediction; and calculating the weight of each residual life prediction result according to the distance, and carrying out weighted summation to obtain a final life prediction result. According to the method, the service life test data sets under different environmental parameter test data are constructed, the clustering analysis and the machine learning model are combined, the weight is calculated according to the distance, and the accuracy and reliability of service life prediction of the sealing structure are improved.

Description

Technical Field

[0001] The present application relates to the technical field of sealing structure life prediction, and in particular to a method for predicting the life of a sealing structure of an aviation hydraulic pipeline system. Background Art

[0002] The aviation hydraulic pipeline system is the main component of the aviation hydraulic system. This system completes mechanical movement through the flow of liquid and pressure transmission, and is the basis for the normal operation of multiple key aviation systems. The reliability and life of the sealing structure of the aviation hydraulic pipeline system directly affect the stability and efficiency of the hydraulic system. By predicting the life of the sealing structure of the aviation hydraulic pipeline system, it is possible to timely discover the potential hidden dangers of the sealing structure and prevent hydraulic system failures caused by sealing failure, thereby ensuring aviation flight safety.

[0003] In the prior art, Chinese patent CN117034640A discloses a pipeline life prediction method, system, terminal device and storage medium, the method comprising: if a single performance parameter analysis item corresponds to multiple operating parameters, then a multivariate analysis table of the corresponding performance parameter analysis item is generated in combination with each operating parameter; the multivariate analysis table is parsed to obtain the periodic impact trend of the corresponding performance parameter analysis item; a life prediction model corresponding to the target pipeline is established in combination with the periodic impact trend and the performance analysis index of the target pipeline; the historical maintenance data of the target pipeline is imported into the life prediction model to generate life prediction data corresponding to the target pipeline; if there are multiple target pipelines, the life prediction data and the associated data between the target pipelines are combined to generate a life prediction distribution map of the pipeline system.

[0004] Chinese patent CN114528662A discloses a method, device and operating machine for predicting the remaining life of a structural part. The method comprises: inputting target data of the operating machine into a remaining life prediction model so that the remaining life prediction model outputs the remaining life of the structural part to be predicted in the operating machine.

[0005] However, the above-mentioned existing technologies do not take into account that different environmental combinations (such as high pressure, high temperature and high vibration, high pressure, low temperature and high vibration, low pressure, low temperature and low vibration, etc.) have different nonlinear effects on the aging of the sealing structure, and the correlation between the environmental parameters and the distribution of the prediction model training data is poor, and the accuracy and reliability of the sealing structure life prediction are poor. Summary of the invention

[0006] The present application provides a life prediction method for the sealing structure of an aviation hydraulic pipeline system, which is used to solve the problems that the prior art does not take into account the different nonlinear effects of different environmental combinations on the aging of the sealing structure, the correlation between environmental parameters and the distribution of prediction model training data is poor, and the accuracy and reliability of the sealing structure life prediction are poor.

[0007] On the one hand, the present application provides a life prediction method for a sealing structure of an aviation hydraulic pipeline system, comprising the following steps: Step 1: construct a life test data set under different environmental parameter test data, perform cluster analysis according to the environmental parameters, and obtain several cluster life test cluster data sets.

[0008] Step 2: Pre-train a machine learning model based on each cluster of life test clustering data sets to obtain several life prediction models.

[0009] Step three: Collect environmental parameter data and status parameter data of the sealing structure of the aviation hydraulic pipeline system.

[0010] Step 4: Calculate the distance between the environmental parameter data and the environmental parameter cluster center corresponding to each life prediction model, and select at least two life prediction models according to the distance from near to far.

[0011] Step five, respectively using each selected life prediction model to perform life prediction according to the environmental parameter data and the state parameter data to obtain at least two remaining life prediction results.

[0012] Step six, calculating the weight of each remaining life prediction result according to the distance between the environmental parameter data and the environmental parameter cluster center of each selected life prediction model, and performing weighted summation to obtain the final life prediction result.

[0013] In a possible implementation, step one includes: The same type of aviation hydraulic pipeline system sealing structure is subjected to life tests under different environmental parameter test data in advance, and a life test data set under different environmental parameter test data is constructed. Each set of data in the life test data set includes environmental parameter test data, state parameter test data at a certain moment and the corresponding test remaining life.

[0014] The life test data sets under different environmental parameter test data are clustered according to the environmental parameters to obtain several cluster life test data sets.

[0015] In a possible implementation, in step 1, the cluster analysis adopts one of a K-means clustering algorithm, a hierarchical clustering algorithm, and a spectral clustering algorithm.

[0016] In a possible implementation, in step 2, each life prediction model uses the same model architecture.

[0017] The machine learning model adopts one of a feedforward neural network model, a convolutional neural network model, and a support vector machine model.

[0018] In one possible implementation, in step 2, when the machine learning model adopts a feedforward neural network model or a convolutional neural network model, before pre-training each machine learning model, the initial weight of each type of data in the corresponding life test clustering data set is set.

[0019] In a possible implementation, the initial weight of each type of data in the life test clustering data set corresponding to the setting adopts one of an entropy weight method and a coefficient of variation method.

[0020] In a possible implementation, in step three, the environmental parameter data includes: pressure data, temperature data, vibration data, flow data, and the state parameter data includes: pressure difference data and contact stress data.

[0021] In a possible implementation, in step 4, the distance between the environmental parameter data and the environmental parameter cluster center corresponding to each life prediction model is calculated using one of the methods selected from Euclidean distance, Mahalanobis distance, cosine similarity, and Manhattan distance.

[0022] In a possible implementation, in step six, a Gaussian kernel function is used to calculate the weight of each remaining life prediction result according to the distance between the environmental parameter data and the environmental parameter cluster center of each selected life prediction model.

[0023] In a possible implementation, step six also includes: step seven, obtaining several final life prediction results within a preset time period and performing linear fitting to generate a life function line, modifying the abnormal values ​​in the several final life prediction results to corresponding values ​​in the life function line, and optimizing and training the corresponding life prediction model based on the modified final life prediction results and the corresponding environmental parameter data and state parameter data.

[0024] The linear fitting adopts the least square method.

[0025] The life prediction method of a sealing structure of an aviation hydraulic pipeline system in this application has the following advantages: By constructing a life test data set under different environmental parameter test data, combining cluster analysis, machine learning models, and calculating weights based on distance, the nonlinear influence of different environmental combinations on the aging of the sealing structure is considered, the correlation between the environmental parameters and the distribution of the prediction model training data is improved, and the accuracy and reliability of the sealing structure life prediction are improved.

[0026] By pre-training a machine learning model based on each cluster of life test clustering data sets, several life prediction models are obtained, which improves the targetedness of the training and thus improves the reliability of the life prediction model.

[0027] The initial weight of each type of data in the life test clustering data set corresponding to the proposed setting adopts one of the entropy weight method and the coefficient of variation method. The initial weight is set for each cluster of life test clustering data sets through the weighting method, which improves the rationality of the initial weight, enables the training process of the life prediction model to converge quickly, and improves the training efficiency.

[0028] The proposed method uses Gaussian kernel function to calculate the weight of each remaining life prediction result according to the distance between the environmental parameter data and the environmental parameter clustering center of each selected life prediction model, which avoids the risk of losing valid information in a single prediction model, reduces the fluctuation error of the prediction results, and also realizes the nonlinear fusion of the prediction results.

[0029] It is proposed to modify the abnormal values ​​in several final life prediction results into corresponding values ​​in the life function line, and optimize the corresponding life prediction model according to the modified final life prediction results and the corresponding environmental parameter data and state parameter data, thereby improving the evolvability and adaptability of the life prediction model. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 A schematic flow chart of a life prediction method for a sealing structure of an aviation hydraulic pipeline system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0033] like Figure 1 As shown, the embodiment of the present application provides a life prediction method for a sealing structure of an aviation hydraulic pipeline system, comprising the following steps: Step 1: construct a life test data set under different environmental parameter test data, perform cluster analysis according to the environmental parameters, and obtain several cluster life test cluster data sets.

[0034] Step 2: Pre-train a machine learning model based on each cluster of life test clustering data sets to obtain several life prediction models.

[0035] Step three: Collect environmental parameter data and status parameter data of the sealing structure of the aviation hydraulic pipeline system.

[0036] Step 4: Calculate the distance between the environmental parameter data and the environmental parameter cluster center corresponding to each life prediction model, and select at least two life prediction models according to the distance from near to far.

[0037] Step five, respectively using each selected life prediction model to perform life prediction according to the environmental parameter data and the state parameter data to obtain at least two remaining life prediction results.

[0038] Step six, calculating the weight of each remaining life prediction result according to the distance between the environmental parameter data and the environmental parameter cluster center of each selected life prediction model, and performing weighted summation to obtain the final life prediction result.

[0039] Exemplarily, step one includes: The same type of aviation hydraulic pipeline system sealing structure is subjected to life tests under different environmental parameter test data in advance, and a life test data set under different environmental parameter test data is constructed. Each set of data in the life test data set includes environmental parameter test data, state parameter test data at a certain moment and the corresponding test remaining life.

[0040] The life test data sets under different environmental parameter test data are clustered according to the environmental parameters to obtain several cluster life test data sets.

[0041] Exemplarily, in step 1, the cluster analysis adopts one of a K-means clustering algorithm, a hierarchical clustering algorithm, and a spectral clustering algorithm.

[0042] Specifically, in this embodiment, a total of 32 life test data sets under different environmental parameter test data are constructed. The K-means clustering algorithm is used to perform cluster analysis on the 32 life test data sets according to the environmental parameters. The specific process is as follows: the environmental parameter test data in the 32 life test data sets are selected as clustering features, and all environmental parameter test data are standardized; the number of clusters (i.e., K value) is selected, and the elbow method is used in this embodiment to obtain a K value of 4; the K-means clustering algorithm is used to cluster the life test data sets under different environmental parameter test data according to the K value, and 4 clusters of life test cluster data sets are obtained.

[0043] Exemplarily, in step 2, each life prediction model adopts the same model architecture.

[0044] The machine learning model adopts one of a feedforward neural network model, a convolutional neural network model, and a support vector machine model.

[0045] Specifically, in this embodiment, in order to match the number of clusters of the life test clustering data set, four life prediction models are set, and each life prediction model is obtained by pre-training a feedforward neural network model with a cluster of life test clustering data sets. In this embodiment, the model architecture of the feedforward neural network includes an input layer, a hidden layer and an output layer, wherein the input layer is used to input environmental parameter data and state parameter data (environmental parameter test data and state parameter test data are input during the training process, and the corresponding test remaining life is used as the target prediction value), the hidden layer uses the ReLU activation function, and the output layer is used to output the remaining life prediction result.

[0046] In this embodiment, the process of pre-training each feedforward neural network model includes: standardizing the data in the corresponding life test clustering data set to ensure stability and efficiency during the model training process; selecting a loss function (mean square error is selected in this embodiment) to measure the difference between the remaining life prediction result output by the output layer and the corresponding target prediction value; selecting an optimizer (Adam optimizer is selected in this embodiment) to update the model weights; dividing the life test clustering data set into a training set and a validation set, using the training set to train the feedforward neural network model until the loss function converges or reaches a preset maximum number of iterations, and using the validation set to evaluate the model performance to prevent overfitting.

[0047] Exemplarily, in step 2, when the machine learning model adopts a feedforward neural network model or a convolutional neural network model, before pre-training each machine learning model, the initial weight of each type of data in the corresponding life test clustering data set is set.

[0048] Exemplarily, the initial weight of each type of data in the life test clustering data set corresponding to the setting adopts one of the entropy weight method and the coefficient of variation method.

[0049] Specifically, in this embodiment, the entropy weight method is used to set the initial weight of each type of data in the four clusters of life test clustering data sets. The entropy weight method is an objective weighting method that uses information entropy to calculate the weights of various types of data. According to the explanation of the basic principles of information theory, information is a measure of the degree of order of a system, and entropy is a measure of the degree of disorder of a system. For a certain type of data, the smaller the information entropy value, the greater the degree of discreteness of the data, and the greater the impact of the data on the comprehensive evaluation (i.e., weight), and vice versa. The specific weight setting process includes: calculating the information entropy of each type of data, which reflects the amount of information and uncertainty of the data; calculating the redundancy of each type of data, that is, 1 minus the information entropy, which reflects the degree to which the data provides effective information; dividing the redundancy of each type of data by the sum of the redundancies of all data to obtain the initial weight of each type of data.

[0050] By setting the initial weight of each type of data in the corresponding life test clustering data set in this way, the corresponding life prediction model can learn the key features in the corresponding life test clustering data set more quickly during training, so that the model converges quickly.

[0051] Exemplarily, in step three, the environmental parameter data includes: pressure data, temperature data, vibration data, flow data, and the state parameter data includes: pressure difference data and contact stress data.

[0052] Specifically, in this embodiment, the pressure data uses the average working pressure of the pipeline where the sealing structure is located, the temperature data uses the average working temperature of the pipeline where the sealing structure is located, the vibration data uses the average vibration intensity of the pipeline where the sealing structure is located, and the flow data uses the fluid flow of the pipeline where the sealing structure is located. The pressure difference data uses the pressure difference on both sides of the sealing structure, and the contact stress data includes the maximum contact stress, the minimum contact stress and the contact stress standard deviation of the contact surface of the sealing structure.

[0053] Exemplarily, in step 4, the distance between the environmental parameter data and the environmental parameter cluster center corresponding to each life prediction model is calculated using one of the methods of Euclidean distance, Mahalanobis distance, cosine similarity, and Manhattan distance.

[0054] Specifically, in this embodiment, the Mahalanobis distance is used to calculate the distance between the environmental parameter data and the environmental parameter cluster center corresponding to each life prediction model. Since the environmental parameter cluster center corresponding to each cluster of life test cluster data set has been obtained when cluster analysis is performed according to the environmental parameters in step 1, and thus, the environmental parameter cluster center corresponding to each life prediction model is known, it is only necessary to directly substitute the environmental parameter data and the environmental parameter cluster center corresponding to each life prediction model into the existing Mahalanobis distance calculation formula. In other possible embodiments, Euclidean distance, cosine similarity, Manhattan distance and the like may also be used.

[0055] In this embodiment, a total of two life prediction models, one with the closest distance and the other with the second closest distance, are selected to perform life prediction based on the environmental parameter data and the state parameter data to obtain two remaining life prediction results.

[0056] Exemplarily, in step six, a Gaussian kernel function is used to calculate the weight of each remaining life prediction result according to the distance between the environmental parameter data and the environmental parameter cluster center of each selected life prediction model.

[0057] Specifically, in this embodiment, before calculating the weight based on the distance, the distance between the environmental parameter data and each selected environmental parameter cluster center is first standardized; then a Gaussian kernel function is used to convert each standardized distance into a corresponding weight; the weight is then normalized so that the sum of the weights of all remaining life prediction results is equal to 1; and the normalized weight is used to perform a weighted summation on each remaining life prediction result to obtain the final life prediction result.

[0058] Exemplarily, step six also includes: step seven, obtaining several final life prediction results within a preset time period and performing linear fitting to generate a life function line, modifying the abnormal values ​​in the several final life prediction results to corresponding values ​​in the life function line, and optimizing and training the corresponding life prediction model based on the modified final life prediction results and the corresponding environmental parameter data and state parameter data.

[0059] The linear fitting adopts the least square method.

[0060] Specifically, the least squares method finds the best fitting line by minimizing the sum of the squares of the vertical distances from all data points to the fitting line. Suppose that several final life prediction results within a preset time period are expressed as (x 1 ,y 1 ),(x 2 ,y 2 ),...,(x n ,y n ), where x represents the time, x n represents the time value of the nth final life prediction result within the preset time period, y represents the predicted remaining life, and y n represents the predicted remaining life value of the nth final life prediction result within the preset time period, n represents the total number of final life prediction results within the preset time period, and the equation of the life function line is set to y=mx+b, where m represents the slope of the line and b represents the intercept of the line. The goal of the least squares method is to find the error function Minimum m and b values: .

[0061] Among them, xi Indicates the time value of the final life prediction result of the i-th time period within the preset time period, y i Represents the predicted remaining life value of the i-th final life prediction result within a preset time period.

[0062] In this embodiment, taking the preset time period of 24 hours as an example, 25 final life prediction results are obtained at intervals of 1 hour and linear fitting is performed to generate a life function line. If the error between a certain final life prediction result and the corresponding value in the life function line is greater than a preset threshold (the moment values ​​are equal, the error is the difference in the predicted remaining life, which is set to 1 hour in this embodiment), the final life prediction result is modified to the corresponding value in the life function line, and the corresponding life prediction model is optimized and trained based on the modified final life prediction result (the modified final life prediction result can be used to reversely modify the remaining life prediction result before the weighted summation) and the corresponding environmental parameter data and state parameter data.

[0063] The embodiment of the present application constructs a life test data set under different environmental parameter test data, combines cluster analysis, machine learning models, and calculates weights based on distance, takes into account the nonlinear effects of different environmental combinations on the aging of the sealing structure, improves the correlation between the environmental parameters and the distribution of the prediction model training data, and improves the accuracy and reliability of the sealing structure life prediction.

[0064] By pre-training a machine learning model based on each cluster of life test clustering data sets, several life prediction models are obtained, which improves the targetedness of the training and thus improves the reliability of the life prediction model.

[0065] The initial weight of each type of data in the life test clustering data set corresponding to the proposed setting adopts one of the entropy weight method and the coefficient of variation method. The initial weight is set for each cluster of life test clustering data sets through the weighting method, which improves the rationality of the initial weight, enables the training process of the life prediction model to converge quickly, and improves the training efficiency.

[0066] The proposed method uses Gaussian kernel function to calculate the weight of each remaining life prediction result according to the distance between the environmental parameter data and the environmental parameter clustering center of each selected life prediction model, which avoids the risk of losing valid information in a single prediction model, reduces the fluctuation error of the prediction results, and also realizes the nonlinear fusion of the prediction results.

[0067] It is proposed to modify the abnormal values ​​in several final life prediction results into corresponding values ​​in the life function line, and optimize the corresponding life prediction model according to the modified final life prediction results and the corresponding environmental parameter data and state parameter data, thereby improving the evolvability and adaptability of the life prediction model.

[0068] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0069] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A life prediction method for a sealing structure of an aviation hydraulic pipeline system, characterized in that: The following steps are involved: Step 1: construct life test data sets under different environmental parameter test data, perform cluster analysis according to environmental parameters, and obtain several cluster life test cluster data sets; Step 2: pre-train a machine learning model based on each cluster of life test clustering data sets to obtain several life prediction models; Step 3, collecting environmental parameter data and state parameter data of the sealing structure of the aviation hydraulic pipeline system; Step 4, calculating the distance between the environmental parameter data and the environmental parameter cluster center corresponding to each life prediction model, and selecting at least two life prediction models according to the distance from near to far; Step 5, respectively using each selected life prediction model to perform life prediction according to the environmental parameter data and the state parameter data to obtain at least two remaining life prediction results; Step six, calculating the weight of each remaining life prediction result according to the distance between the environmental parameter data and the environmental parameter cluster center of each selected life prediction model, and performing weighted summation to obtain the final life prediction result.

2. The life prediction method of the sealing structure of an aviation hydraulic pipeline system according to claim 1 is characterized in that: Step one includes: Perform life tests on the same type of aviation hydraulic pipeline system sealing structure under different environmental parameter test data in advance, and construct life test data sets under different environmental parameter test data, wherein each set of data in the life test data set includes environmental parameter test data, state parameter test data at a certain moment, and corresponding test remaining life; The life test data sets under different environmental parameter test data are clustered according to the environmental parameters to obtain several cluster life test data sets.

3. The life prediction method of the sealing structure of an aviation hydraulic pipeline system according to claim 1 is characterized in that: In step 1, the cluster analysis adopts one of the K-means clustering algorithm, the hierarchical clustering algorithm, and the spectral clustering algorithm.

4. The life prediction method of the sealing structure of an aviation hydraulic pipeline system according to claim 1 is characterized in that: In step 2, each life prediction model uses the same model architecture; The machine learning model adopts one of a feedforward neural network model, a convolutional neural network model, and a support vector machine model.

5. The life prediction method of the sealing structure of an aviation hydraulic pipeline system according to claim 4 is characterized in that: In step 2, when the machine learning model adopts a feedforward neural network model or a convolutional neural network model, before pre-training each machine learning model, the initial weight of each type of data in the corresponding life test clustering data set is set.

6. The life prediction method of the sealing structure of an aviation hydraulic pipeline system according to claim 5 is characterized in that: The initial weight of each type of data in the life test clustering data set corresponding to the setting adopts one of the entropy weight method and the coefficient of variation method.

7. The life prediction method of the sealing structure of an aviation hydraulic pipeline system according to claim 1 is characterized in that: In step three, the environmental parameter data include: pressure data, temperature data, vibration data, flow data, and the state parameter data include: pressure difference data and contact stress data.

8. The life prediction method of the sealing structure of an aviation hydraulic pipeline system according to claim 1 is characterized in that: In step 4, the distance between the environmental parameter data and the environmental parameter cluster center corresponding to each life prediction model is calculated by using one of the methods of Euclidean distance, Mahalanobis distance, cosine similarity, and Manhattan distance.

9. The life prediction method of the sealing structure of an aviation hydraulic pipeline system according to claim 1, characterized in that: In step six, a Gaussian kernel function is used to calculate the weight of each remaining life prediction result according to the distance between the environmental parameter data and the environmental parameter cluster center of each selected life prediction model.

10. The life prediction method of the sealing structure of an aviation hydraulic pipeline system according to claim 1, characterized in that: The method further includes the following steps after step 6: step 7, obtaining several final life prediction results within a preset time period and performing linear fitting to generate a life function straight line, modifying abnormal values ​​in the several final life prediction results to corresponding values ​​in the life function straight line, and optimizing and training the corresponding life prediction model according to the modified final life prediction results and the corresponding environmental parameter data and state parameter data; The linear fitting adopts the least square method.

Images