Physical data driven aircraft lubricating oil temperature fault diagnosis and prediction method and system

Abstract

The invention discloses a physical data driven aircraft lubricating oil temperature fault diagnosis and prediction method and system, and belongs to the technical field of electrical data processing. The method comprises the following steps: S01, calculating a lubricating oil temperature characteristic value through a characteristic extraction model according to engine rotating speed and lubricating oil temperature data; s02, judging whether the lubricating oil temperature characteristic value is within a first upper limit threshold value and a first lower limit threshold value or not, if yes, determining that the characteristic value is normal, then executing the step S03, and updating the first upper limit threshold value and the first lower limit threshold value according to the characteristic value; if not, ending; s03, inputting the lubricating oil temperature characteristic value into the lubricating oil temperature fault prediction model to obtain a prediction value, and if the prediction value is within a second upper limit threshold value and a second lower limit threshold value, considering that the probability that the lubricating oil temperature fault does not occur in the future time is low; if not, it is considered that the probability that the lubricating oil temperature fault occurs in the future time is high, and fault early warning needs to be sent out. The method has better interpretability and predictability; the uncertainty and subjectivity of traditional fixed threshold diagnosis can be overcome.

Description

Technical Field

[0001] The present invention relates to a physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method and system, belonging to the technical field of fault diagnosis and prediction. Background Art

[0002] The Chinese patent application with the publication number CN117932550A discloses an online lubricating oil early warning method based on an adaptive threshold. The method includes the following steps: loading operation data, constructing an autoregressive integrated moving average model based on a random search strategy, obtaining prediction data based on the autoregressive integrated moving average model and the operation data, and splicing the operation data and the prediction data into splicing data as verification data; loading initial training data, constructing a self-associative kernel regression model with a max-pooling strategy; inputting the initial training data and the verification data into the self-associative kernel regression model with the max-pooling strategy to construct an adaptive threshold; and realizing the state early warning of the lubricating oil by judging whether the expected value of the verification data exceeds the adaptive threshold.

[0003] This invention combines two models, namely, an autoregressive integrated moving average model and a self-associative kernel regression model with a max-pooling strategy, and can realize a faster and more accurate online lubricating oil state recognition and early warning ability. However, this invention directly uses the original lubricating oil data for prediction, and its interpretability and predictability are poor. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, the invention aims to provide a physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method and system, which have better interpretability and predictability.

[0005] To achieve the above object, the present invention provides a physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method, which includes the following steps: S01: Calculate the lubricating oil temperature characteristic value according to the engine speed and lubricating oil temperature data through a feature extraction model; S02: Judge whether the lubricating oil temperature characteristic value is within a first upper limit threshold and a first lower limit threshold. If it is within the first upper limit threshold and the first lower limit threshold, judge that the lubricating oil temperature characteristic value is normal, then execute step S03, and update the first upper limit threshold and the first lower limit threshold according to the lubricating oil temperature characteristic value; if it is not within the first upper limit threshold and the first lower limit threshold, judge that the lubricating oil temperature characteristic value is abnormal, do not update the first upper limit threshold and the first lower limit threshold, and then end; S03: Input the lubricating oil temperature characteristic value into the lubricating oil temperature fault prediction model to obtain a predicted value , compare the predicted value with the second upper threshold and the second lower threshold. If the predicted value is within the second upper threshold and the second lower threshold, it is considered that the probability of no lubricating oil temperature fault occurring in the future is low; otherwise, it is considered that the probability of a lubricating oil temperature fault occurring in the future is high, and a fault warning needs to be issued.

[0006] To achieve the above-mentioned invention purpose, the present invention also provides a physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction system, which includes a storage medium and one or more processors. The storage medium stores a computer program, and the computer program is called by one or more processors to implement the above-mentioned physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method.

[0007] Compared with the prior art, the present invention has the following beneficial effects: 1. The process of extracting the lubricating oil temperature characteristic value proposed by the present invention incorporates a feature extraction model. Compared with directly extracting the lubricating oil temperature value, the lubricating oil temperature error value adopted by the present invention contains physical model information and has better interpretability and predictability; 2. In the process of constructing the lubricating oil temperature threshold, the adaptive threshold constructed through historical data can overcome the uncertainty and subjectivity of traditional fixed-threshold diagnosis. The continuous threshold update ensures the long-term effectiveness of the adaptive threshold and can effectively reduce the false alarm rate; 3. The prediction method based on the autoregressive integrated moving average model can consider seasonal factors to a certain extent. Compared with traditional prediction methods, it has better accuracy and stability and can accurately give a warning of lubricating oil temperature faults. Brief Description of the Drawings

[0008] Figure 1 is a flowchart of the physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method provided by the present invention.

[0009] Figure 2 is a schematic diagram of the tailing phenomenon.

[0010] Figure 3 is a schematic diagram of the truncation phenomenon. Detailed Embodiments

[0011] To make the technical means, creative features, achieved purposes and functions of the present invention easy to understand, the present invention will be further described below in conjunction with specific embodiments.

[0012] In the present invention, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0013] Figure 1 is a flowchart of the physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method provided by the present invention; as Figure 1As shown in the figure, the physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method provided by the present invention includes the following steps: S01: Calculate the lubricating oil temperature characteristic value according to the engine speed and lubricating oil temperature data through the feature extraction model; S02: Determine whether the lubricating oil temperature characteristic value is within the first upper limit threshold and the first lower limit threshold. If it is within the first upper limit threshold and the first lower limit threshold, it is determined that the lubricating oil temperature characteristic value is normal, and then step S03 is executed, and the first upper limit threshold and the first lower limit threshold are updated according to the lubricating oil temperature characteristic value; if it is not within the first upper limit threshold and the first lower limit threshold, it is determined that the lubricating oil temperature characteristic value is abnormal, the first upper limit threshold and the first lower limit threshold are not updated, and then the process ends; S03: Input the lubricating oil temperature characteristic value into the lubricating oil temperature fault prediction model to obtain a predicted value , compare the predicted value with the second upper limit threshold and the second lower limit threshold. If the predicted value is within the second upper limit threshold and the second lower limit threshold, it is considered that the probability of no lubricating oil temperature fault in the future is low; if not, it is considered that the probability of lubricating oil temperature fault in the future is high, and a fault warning needs to be issued. In the present invention, .

[0014] The process of extracting the lubricating oil temperature characteristic value proposed by the present invention incorporates a feature extraction model. Compared with directly extracting the lubricating oil temperature value, the lubricating oil temperature error value adopted by the present invention contains physical model information and has better interpretability and predictability.

[0015] In the present invention, the lubricating oil temperature fault prediction model is trained by an ARIMA model. The data form of the ARIMA model is: , In the formula, c is a constant term, p is the number of autoregressive terms; q is the number of moving average terms; are the coefficients of the AR model respectively; are the coefficients of the MR model respectively; is the characteristic value time series value; is the error sequence value; and the characteristic value time series value is differenced by d orders;

[0016] The lubricating oil temperature fault prediction model is trained by an ARIMA model. Training the ARIMA model into a lubricating oil temperature fault prediction model includes the following steps: S1-1: Obtain the historical lubricating oil temperature time series characteristic value dataset from the database ; S1-2: Slide and segment the historical lubricating oil temperature time series characteristic value data according to the input vector dimension tf into input samples , corresponding to the input sample The test label is: , input the input sample into the ARIMA model, and the estimated label output by it is: , where i is the sample serial number, tf is the input vector dimension, and tp is the output vector dimension; S1-3: Calculate the prediction using the following formula and the deviation MSE from the test label and record it: , Then, adjust the parameters of the ARIMA model. The parameters include the number of autoregressive terms p, the order of differencing d, and the number of moving average terms q, and repeat the deviation calculation and recording process. Take the parameter values of the ARIMA model when the minimum deviation value MSE is obtained as the final value.

[0017] In the present invention, training the ARIMA model into a lubricating oil temperature fault prediction model further includes the following steps: S1-4: Perform autocorrelation analysis on the historical lubricating oil temperature time series eigenvalue dataset, and calculate the k-order autocorrelation coefficient , k = 1, 2,..., N, and its calculation process is as follows: , In the formula, N is the eigenvalue data length, is the average value of N eigenvalue data; is an intermediate variable.

[0018] In the present invention, training the ARIMA model into a lubricating oil temperature fault prediction model further includes the following steps: S1-5: Perform partial autocorrelation analysis on the historical lubricating oil temperature time series eigenvalue dataset, and calculate the k-order partial autocorrelation coefficient , k = 1, 2,..., n; the k-order partial autocorrelation coefficient is transformed from the autocorrelation coefficient , and its calculation process is: .

[0019] When performing autocorrelation and partial autocorrelation analysis on the historical lubricating oil temperature time series eigenvalue data, observe whether the autocorrelation plot and partial autocorrelation plot of the time series show, for example Figure 2 a trailing phenomenon or, for example Figure 3The truncation phenomenon is that the autocorrelation coefficient or partial autocorrelation coefficient still maintains a high value after a certain order or shows obvious periodic fluctuations after a certain order; the truncation phenomenon is that the autocorrelation coefficient or partial autocorrelation coefficient is close to zero after a certain order or drops sharply after a certain order.

[0020] If the time series autocorrelogram is tailing and the time series partial autocorrelogram is truncated, the difference order d and the number of moving average items q are both 0, and the number of autoregressive items p is the number of truncated items in the partial autocorrelogram; if the time series autocorrelogram is truncated and the partial autocorrelogram is tailing, the number of autoregressive items p and the difference order d are both 0, and the number of moving average items q is the number of truncated items in the time series autocorrelogram; if the above two conditions are not met, the values of the autoregressive items p, the difference order d, and the moving average items q are determined through S1-2 and step S1-3.

[0021] The prediction method based on the ARIMA model has better accuracy and stability than traditional prediction methods, and can accurately warn of lubricating oil temperature failures.

[0022] In the present invention, the first upper threshold and the first lower threshold are determined by a lubricating oil temperature threshold model, and the determination process includes: S2-1: Select the latest M-1 historical feature values and oil temperature characteristic value ; S2-2: Update the first upper threshold value by the following formula and the first lower threshold : ; , Where, is the adjustment factor; , Where, is the probability distribution; Indicates that the probability distribution is selected within the value range Take the maximum value The value of; , Where, are the probability distributions of the first upper threshold and the first lower threshold before update The mean and variance of .

[0023] The adaptive threshold constructed by the present invention through historical data can overcome the uncertainty and subjectivity of traditional fixed threshold diagnosis. The long-term effectiveness of the adaptive threshold guaranteed by continuous threshold update can effectively reduce the false alarm rate.

[0024] In the present invention, the feature extraction model calculates the lubricating oil temperature eigenvalue from the engine speed and lubricating oil temperature data, specifically including: S4-1: Calculate the theoretical value of the lubricating oil temperature according to the following formula : , where, is the engine speed, is a parameter; S4-2: Calculate the lubricating oil temperature eigenvalue according to the following formula: , where, is the measured value of the lubricating oil temperature, and A is the number of data when calculating the lubricating oil temperature eigenvalue.

[0025] In the present invention, a stable segment extraction algorithm is used to extract the time period when the engine speed is stable.

[0026] In the present invention, the stable segment extraction algorithm includes the following steps: S3-1: Starting from the head of the engine speed data in the cruise stage, select the engine speed data with the window length, and record the start and end positions of the window; S3-2: Judge whether the standard deviation of the selected window data is lower than the preset value; S3-3: If it is lower than the preset value, go to step S3-4; if it is not lower than the preset value, move the window start position one bit backward, and then return to step S3-2; S5-4: Judge whether the standard deviation of the new window data is lower than the known preset value S3-5: If it is lower than the preset value, go to step S3-6; if it is not lower than the preset value, move the window tail position one bit backward, and then return to step S3-2; S3-6: Save the start and end positions of the window position, and end the program.

[0027] The present invention also provides a physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction system, which includes a storage medium and one or more processors. The storage medium stores a computer program, and the computer program is called by one or more processors to implement the above-mentioned physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method.

[0028] The present invention also provides a computer program product, which compiles the above-mentioned physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method into a computer program called and executed by one or more processors using a computer language.

[0029] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and all these changes and improvements fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A physical data-driven method for aircraft lubricating oil temperature fault diagnosis and prediction, characterized in that The steps include: S01: Calculating a lubricating oil temperature characteristic value based on engine speed and lubricating oil temperature data using a feature extraction model; S02: Determining whether the lubricating oil temperature characteristic value is within a first upper threshold and a first lower threshold. If so, determining that the lubricating oil temperature characteristic value is normal, and then executing step S03 to update the first upper threshold and the first lower threshold based on the lubricating oil temperature characteristic value; If it is not within the first upper threshold and the first lower threshold, the lubricating oil temperature characteristic value is determined to be abnormal, the first upper threshold and the first lower threshold are not updated, and the process ends; S03: Input the lubricating oil temperature characteristic value into the lubricating oil temperature fault prediction model to obtain a predicted value , compare the predicted value with the second upper threshold and the second lower threshold. If the predicted value is within the second upper threshold and the second lower threshold, it is considered that the probability of no lubricating oil temperature fault in the future is low; otherwise, it is considered that the probability of a lubricating oil temperature fault in the future is high, and a fault warning needs to be issued.

2. The physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method according to claim 1, wherein The lubricating oil temperature fault prediction model is trained by the ARIMA model. Training the ARIMA model into the lubricating oil temperature fault prediction model includes the following steps: S1-1: Obtain the historical lubricating oil temperature time series eigenvalue dataset from the database ; S1-2: Slide and segment the historical lubricating oil temperature time series eigenvalue data into input samples according to the input vector dimension tf The corresponding test label for the input sample is: Input the input sample into the ARIMA model, and the estimated label output by it is: where i is the sample serial number, tf is the input vector dimension, and tp is the output vector dimension; ​ S1-3: Calculate the prediction using the following formula and the test label to calculate and record the deviation MSE ， Then, adjust the parameters of the ARIMA model and repeat the deviation calculation and recording process to obtain the parameter value of the ARIMA model when the minimum deviation value MSE is obtained as the final value.

3. The physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method according to claim 2, wherein Training the ARIMA model into an oil temperature fault prediction model includes the following steps: S1-4: Perform autocorrelation analysis on the historical lubricating oil temperature time series eigenvalue dataset and calculate the k-th order autocorrelation coefficient , where k = 1, 2, …, N, and the calculation process is as follows: ， Where N is the eigenvalue data Length is the average value of N eigenvalue data; is an intermediate variable.

4. The physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method according to claim 3, wherein Training the ARIMA model into an oil temperature fault prediction model includes the following steps: S1-5: Conduct partial autocorrelation analysis on the historical lubricating oil temperature time series eigenvalue dataset, and calculate the kth-order partial autocorrelation coefficient , where k = 1, 2, …, n; the kth-order partial autocorrelation coefficient is transformed from the autocorrelation coefficient , and its calculation process is as follows: 。 5. The physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method according to claim 1, wherein The first upper threshold and the first lower threshold are determined by a lubricating oil temperature threshold model, and the determination process includes: S2-1: Select the latest M-1 historical eigenvalue and the lubricating oil temperature eigenvalue ; S2-2: Update the first upper limit threshold and the first lower limit threshold respectively according to the following formula and the first lower limit threshold : ； ， In the formula, is the adjustment coefficient; ， is a probability distribution; represents the value that makes the probability distribution take the maximum value within the value range; ， Where, are the probability distributions of the first upper threshold and the first lower threshold before update The mean and variance of .

6. The physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method according to claim 1, wherein, The stable segment extraction algorithm is used to extract the period when the engine speed is stable.

7. The physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method according to claim 6, wherein The plateau extraction algorithm includes the following steps: S3-1: Starting from the head of the engine speed data in the cruise phase, select the engine speed data of the window length and record the start and end positions of the window; S3-2: Determine whether the standard deviation of the selected window data is lower than a preset value; S3-3: If it is lower than the preset value, proceed to step S3-4; if it is not lower than the preset value, the window start position is moved back one position, and then return to step S3-2; S3-4: Determine whether the standard deviation of the new window data is lower than the known preset value S3-5: If it is lower than the preset value, proceed to step S3-6; if it is not lower than the preset value, the window tail position is moved back one position, and then return to step S3-2; S3-6: Save the start and end positions of the window and end the program.

8. The physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method according to claim 6, wherein The feature extraction model calculates the lubricating oil temperature feature values based on the engine speed and lubricating oil temperature data, specifically including: S4-1: Calculate the theoretical value of the lubricating oil temperature according to the following formula : , where is the engine speed, is a parameter; S4-2: Calculate the lubricating oil temperature characteristic value according to the following formula: , Wherein, is the measured value of the lubricating oil temperature, and A is the number of data when calculating the characteristic value of the lubricating oil temperature.

9. A physical data driven aircraft oil temperature fault diagnosis and prediction system, characterized in that: The method comprises a storage medium and one or more processors, wherein the storage medium stores a computer program, and the computer program is called by the one or more processors to implement the physical data-driven aircraft lubricating oil temperature fault diagnosis and prediction method according to any one of claims 1 to 8.

Images