Aero turbine engine sensor signal processing method and system based on CSI-EMD

Abstract

This invention relates to a method and system for processing aero-turbine engine sensor signals based on CSI-EMD, belonging to the field of aero-turbine engine technology. The invention addresses the aero-turbine engine sensor signals requiring decomposition by using cubic spline interpolation to expand data points and increase the length of short-time-series signals. Then, in the step of generating the upper and lower envelopes of the signals, the left and right endpoints of the upper and lower envelopes are determined using the fixed extremum method and the extremum translation method. Subsequent steps are the same as traditional EMD, ultimately yielding the decomposed intrinsic mode functions. This invention effectively decomposes aero-turbine engine sensor signals into several intrinsic mode functions at different time scales, improving the endpoint effect problem and the upper and lower envelope intersection problem during the decomposition process. Compared to traditional empirical mode decomposition methods, this invention more effectively decomposes signals into several intrinsic mode functions.

Description

Technical Field

[0001] This invention belongs to the field of aero-turbine engine technology and relates to an aero-turbine engine sensor signal processing method and system based on CSI-EMD. Background Technology

[0002] As the power source for modern aircraft, the reliable operation of turbofan engines is crucial to aircraft safety. Therefore, monitoring the health status of turbofan engines and predicting their remaining service life is essential. Numerous research findings exist on methods for predicting the service life of turbofan engines. The mainstream method utilizes Empirical Mode Decomposition (EMD) to decompose the original time series of the engine into several subsequences. Each subsequence is then modeled separately, and finally, the prediction results of all subsequences are summed to obtain the prediction result for the original time series.

[0003] However, during signal decomposition, if the number of data points in the signal to be decomposed is small, the envelope obtained by cubic spline interpolation (CSI) will be poor, and the upper and lower envelopes are prone to crossing. Furthermore, since the envelope averaging is calculated by performing spline interpolation on the upper and lower extreme points of the original data separately and then averaging, the extreme points at the breakpoints cannot be determined during spline interpolation unless they are also the extreme points of the data. This introduces fitting errors into the spline interpolation. During the "sieving" process of EMD decomposition, due to the uncertainty of the extreme values ​​at the endpoints, each spline interpolation will have fitting errors. This leads to the continuous accumulation of errors, resulting in a large error at the endpoints of the first intrinsic mode function (IMF). The decomposition of the second IMF is based on the original data minus the residual term of the first IMF, which leads to an even larger error in the second IMF. In the continuous decomposition process, the error gradually propagates inward from the endpoints, and in severe cases, it can even contaminate the entire data. Therefore, how to suppress the endpoint effect of empirical mode decomposition is an urgent problem to be solved. Summary of the Invention

[0004] The technical problem to be solved by this invention is how to decompose the sensor signal of an aero-turbine engine into several intrinsic mode functions with different time scales while improving the endpoint effect problem and the crossover problem of the upper and lower envelopes during the decomposition process.

[0005] The present invention solves the above-mentioned technical problems through the following technical solutions:

[0006] On the one hand, the CSI-EMD-based signal processing method for aero-turbine engine sensors includes the following steps: S1. Process the aircraft turbine engine sensor signal to be processed. The data points are augmented by interpolating three data points between every two data points using cubic spline interpolation to obtain the augmented signal. ; S2. Obtain the signal using quadratic polynomial interpolation. All local maxima and local minima; S3. Interpolate all the obtained local maxima using a cubic spline interpolation function to form the upper envelope of the signal, and interpolate all the local minima to form the lower envelope of the signal, and calculate the envelope average of the upper and lower envelopes. ; S4, using signals Subtract the envelope average line Get the difference ; S5. Determine the difference. Does it meet the set conditions for the intrinsic modulus function? If not, then the difference will be... Repeat steps S2 to S4 as a new signal until the condition is met, and record the difference. As the first eigenmode function, denoted as ; S6, using signals Subtract the first eigenmode function The residual components are obtained. ; S7. Transfer the residual components Repeat steps S2 to S6 as a new signal until the residual component is decomposed into a monotonic function or has one and only one extremum. At this point, the signal... It is decomposed into multiple intrinsic moduli and a residual component, i.e. ,in, Let i be the i-th intrinsic modulus function. This is the final residual component.

[0007] This invention addresses the decomposition of aero-turbine engine sensor signals by using cubic spline interpolation to expand the data points and increase the length of short-time-series signals. Then, in the step of generating the upper and lower envelopes of the signal, the left and right endpoints of the upper and lower envelopes are determined using the fixed extremum method and the extremum translation method. Subsequent steps are the same as traditional EMD, ultimately yielding the decomposed intrinsic mode functions. This invention effectively decomposes aero-turbine engine sensor signals into several intrinsic mode functions at different time scales, improving the endpoint effect and upper and lower envelope intersection problems during the decomposition process. Compared to traditional empirical mode decomposition methods, this invention more effectively decomposes signals into several intrinsic mode functions.

[0008] Furthermore, the signal acquisition method using quadratic polynomial interpolation described in step S2... The specific methods for finding all local maxima and local minima are as follows: Using the function values ​​of the target signal at three different points, a function related to the signal is constructed. Approximate quadratic polynomials In quadratic polynomial The extreme points are used as the target signal The approximate extreme points are then determined using the fixed extreme point method to determine the left and right endpoints of the upper and lower envelopes, and the extreme point at the endpoints is determined using the extreme value translation method.

[0009] Furthermore, the method for determining the left and right endpoints of the upper and lower envelopes using the fixed extreme value method is as follows: Select the two endpoints of the expanded time series signal. If the first extreme value is a maximum value, then the left endpoint of the signal is taken as the minimum value point; otherwise, the left endpoint of the signal is taken as the maximum value point. If the last extreme value is a maximum value, then the right endpoint of the signal is taken as the minimum value point; otherwise, the right endpoint of the signal is taken as the maximum value point. The method for determining the extreme points at the endpoints using the extreme value translation method is as follows: If the maximum point at the left endpoint is determined by the fixed extreme value method, then the first minimum point is translated to the left end as the minimum point at the left endpoint; conversely, if the minimum point at the left endpoint is determined by the fixed extreme value method, then the first maximum point is translated to the left end as the maximum point at the left endpoint. If the maximum point at the right endpoint is determined by the fixed extreme value method, then the last minimum point is translated to the right end as the minimum point at the right endpoint; conversely, if the minimum point at the right endpoint is determined by the fixed extreme value method, then the last maximum point is translated to the right end as the maximum point at the right endpoint.

[0010] Furthermore, the envelope average line of the upper and lower envelope lines mentioned in step S3 The calculation formula is as follows:

[0011] in, The upper envelope, The lower envelope, It is the envelope average line.

[0012] Furthermore, the setting conditions described in step S5 are as follows: the number of local extrema and zero-crossing points must be equal or differ by at most one throughout the entire time range; and the mean of the upper envelope and lower envelope is zero at any time.

[0013] On the other hand, the CSI-EMD-based aero-turbine engine sensor signal processing system includes: a signal augmentation module, an extreme point determination module, an envelope averaging module, an intrinsic mode function acquisition module, and a decomposition termination module. The aforementioned signal enhancement module will process the aircraft turbine engine sensor signal. The data points are augmented by interpolating three data points between every two data points using cubic spline interpolation to obtain the augmented signal. ; The extreme point determination module uses quadratic polynomial interpolation to obtain the signal. All local maxima and local minima; The envelope averaging module uses a cubic spline interpolation function to interpolate all local maxima to form the upper envelope of the signal, and interpolates all local minima to form the lower envelope of the signal, and then calculates the envelope average of the upper and lower envelopes. ; The intrinsic modulus function acquisition module uses signals Subtract the envelope average line Get the difference And determine the difference. Does it meet the set conditions for the intrinsic modulus function? If not, then the difference will be... Repeat steps S2 to S4 as a new signal until the condition is met, and record the difference. As the first eigenmode function, denoted as ; The decomposition termination module uses signals Subtract the first eigenmode function The residual components are obtained. ;residual components Repeat steps S2 to S6 as a new signal until the residual component is decomposed into a monotonic function or has one and only one extremum. At this point, the signal... It is decomposed into multiple intrinsic moduli and a residual component, i.e. ,in, Let i be the i-th intrinsic modulus function. This is the final residual component.

[0014] Furthermore, the extreme point determination module describes the use of quadratic polynomial interpolation to obtain the signal. The specific methods for finding all local maxima and local minima are as follows: Using the function values ​​of the target signal at three different points, a function related to the signal is constructed. Approximate quadratic polynomials In quadratic polynomial The extreme points are used as the target signal The approximate extreme points are then determined using the fixed extreme point method to determine the left and right endpoints of the upper and lower envelopes, and the extreme point at the endpoints is determined using the extreme value translation method.

[0015] Furthermore, the method for determining the left and right endpoints of the upper and lower envelopes using the fixed extreme value method is as follows: Select the two endpoints of the expanded time series signal. If the first extreme value is a maximum value, then the left endpoint of the signal is taken as the minimum value point; otherwise, the left endpoint of the signal is taken as the maximum value point. If the last extreme value is a maximum value, then the right endpoint of the signal is taken as the minimum value point; otherwise, the right endpoint of the signal is taken as the maximum value point. The method for determining the extreme points at the endpoints using the extreme value translation method is as follows: If the maximum point at the left endpoint is determined by the fixed extreme value method, then the first minimum point is translated to the left end as the minimum point at the left endpoint; conversely, if the minimum point at the left endpoint is determined by the fixed extreme value method, then the first maximum point is translated to the left end as the maximum point at the left endpoint. If the maximum point at the right endpoint is determined by the fixed extreme value method, then the last minimum point is translated to the right end as the minimum point at the right endpoint; conversely, if the minimum point at the right endpoint is determined by the fixed extreme value method, then the last maximum point is translated to the right end as the maximum point at the right endpoint.

[0016] Furthermore, the envelope average line of the upper and lower envelope lines described in the envelope average line module... The calculation formula is as follows:

[0017] in, The upper envelope, The lower envelope, It is the envelope average line.

[0018] Furthermore, the setting conditions described in the intrinsic modulus function acquisition module are: the number of local extrema and zero-crossing points must be equal or differ by at most one throughout the entire time range; and the mean of the upper envelope and lower envelope is zero at any time.

[0019] The advantages of this invention are: This invention addresses the decomposition of aero-turbine engine sensor signals by using cubic spline interpolation to expand the data points and increase the length of short-time-series signals. Then, in the step of generating the upper and lower envelopes of the signal, the left and right endpoints of the upper and lower envelopes are determined using the fixed extremum method and the extremum translation method. Subsequent steps are the same as traditional EMD, ultimately yielding the decomposed intrinsic mode functions. This invention effectively decomposes aero-turbine engine sensor signals into several intrinsic mode functions at different time scales, improving the endpoint effect and upper and lower envelope intersection problems during the decomposition process. Compared to traditional empirical mode decomposition methods, this invention more effectively decomposes signals into several intrinsic mode functions. Attached Figure Description

[0020] Figure 1 This is a flowchart of a CSI-EMD-based sensor signal processing method for aero-turbine engines according to an embodiment of the present invention; Figure 2 The sensor signal of the aircraft turbine engine to be processed in this embodiment of the invention; Figure 3 The signal obtained by expanding the data points of the aircraft turbine engine sensor signal to be processed according to the embodiment of the present invention is the signal after cubic spline interpolation. Figure 4 The upper and lower envelopes and the average envelope of the aero-turbine engine sensor signal in this embodiment of the invention are obtained using the traditional empirical mode decomposition method. Figure 5 The upper and lower envelopes and the average envelope line are obtained by the CSI-EMD-based aero-turbine engine sensor signal processing method in this embodiment of the invention. Figure 6 The signal decomposition result of the aero-turbine engine sensor signal in this embodiment of the invention is obtained by the traditional empirical mode decomposition method; Figure 7 The signal decomposition result is obtained by the CSI-EMD-based aero-turbine engine sensor signal processing method according to an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments: Example 1 like Figure 1 As shown, the CSI-EMD-based signal processing method for aero-turbine engine sensors of the present invention includes the following steps: Step 1: Process the short time series signal to be processed. The signal data points are augmented using cubic spline interpolation to obtain the augmented signal. ; The method of expanding the data points of the signal using cubic spline interpolation specifically includes: interpolating three data points between every two data points using cubic spline interpolation, thereby ensuring that only the number of data points is increased during the interpolation process without changing the numerical values ​​of the original signal data points. Assuming the original signal has a number of data points... Then the number of data points after interpolation augmentation is . indivual; Step 2: Use quadratic polynomial interpolation (parabolic method) to find all local maxima and local minima of the signal, and use the fixed extremum method and extremum translation method to determine the maxima and minima at the left and right endpoints of the signal; The method of finding all local maxima and local minima of a signal using quadratic polynomial interpolation specifically includes: constructing a function of the target signal at three different points, which is analogous to the signal... Approximate quadratic polynomials In quadratic polynomial The extreme points are used as the target signal Approximate extreme points; The method of determining the left and right endpoints of the upper and lower envelopes using the fixed extreme value method specifically includes: selecting the two endpoints of the expanded time series signal; if the first extreme value is a maximum value, then the left endpoint of the signal is taken as the minimum value point, otherwise the left endpoint of the signal is taken as the maximum value point; if the last extreme value is a maximum value, then the right endpoint of the signal is taken as the minimum value point, otherwise the right endpoint of the signal is taken as the maximum value point. The method of using extreme value translation to determine the extreme points at the endpoints specifically includes: if the maximum point at the left endpoint is determined by the fixed extreme value method, then the first minimum point is translated to the left endpoint as the minimum point at the left endpoint; conversely, if the minimum point at the left endpoint is determined by the fixed extreme value method, then the first maximum point is translated to the left endpoint as the maximum point at the left endpoint. If the maximum point at the right endpoint is determined by the fixed extreme value method, then the last minimum point is translated to the right endpoint as the minimum point at the right endpoint; conversely, if the minimum point at the right endpoint is determined by the fixed extreme value method, then the last maximum point is translated to the right endpoint as the maximum point at the right endpoint.

[0023] Step 3: Interpolate all the obtained maxima points using a cubic spline interpolation function to form the upper envelope of the signal, and interpolate all the minima points to form the lower envelope of the signal. Calculate the envelope average of the upper and lower envelopes. The envelope average line of the upper and lower envelope lines. The calculation formula is as follows:

[0024] in, The upper envelope, The lower envelope, It is the envelope average line.

[0025] Step 4: Process the obtained envelope average line using the traditional EMD method, using signal... Subtract the average Get the difference The formula is specifically defined as follows: .

[0026] Step 5, Judgment Whether an intrinsic modulus function meets the two conditions is: (1) the number of local extrema and zero-crossings must be equal or differ by at most one over the entire time range; (2) the mean of the upper and lower envelopes is zero at any given time. If it does not meet the conditions, then... Repeat steps 2-4 as the new signal to be decomposed until the condition is met. Then, at this point... As the first intrinsic modulus function, it is denoted as: ,in This is the first intrinsic modulus function; Step 6, use signals Subtract the first eigenmode function The residual components are obtained. The formula for calculating the residual component is as follows: .

[0027] Step 7: Remove the residual components Repeat steps 2-6 with the new decomposed signal until the residual component is decomposed into a monotonic function or has one and only one extremum, at which point the decomposition stops. Decomposed into One intrinsic mode function and one residual component are used to obtain the signal. The decomposition is as follows:

[0028] in, For the first Intrinsic modulus function, This is the final residual component.

[0029] pass Figure 4 and Figure 5 The comparison clearly shows that the method of the present invention can effectively alleviate the envelope crossing problem and endpoint effect problem in the traditional EMD method during the process of fitting the envelope.

[0030] pass Figure 6 and Figure 7 The comparison clearly shows that the method of this invention is more effective in decomposing the original signal into several eigenmode functions at different time scales.

[0031] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for processing sensor signals of aero-turbine engines based on CSI-EMD, characterized in that, Includes the following steps: S1. Process the aircraft turbine engine sensor signal to be processed. The data points are augmented by interpolating three data points between every two data points using cubic spline interpolation to obtain the augmented signal. ; S2. Obtain the signal using quadratic polynomial interpolation. Find all local maxima and local minima using the following method: Construct a function value of the target signal at three different points, which is related to the signal... Approximate quadratic polynomials In quadratic polynomial The extreme points are used as target signals The approximate extreme points are then determined using the fixed extreme point method to determine the left and right endpoints of the upper and lower envelopes, and the extreme point at the endpoints is determined using the extreme value translation method. The method for determining the left and right endpoints of the upper and lower envelopes using the fixed extreme value method is as follows: Select the two endpoints of the expanded time series signal. If the first extreme value is a maximum value, then the left endpoint of the signal is taken as the minimum value point; otherwise, the left endpoint of the signal is taken as the maximum value point. If the last extreme value is a maximum value, then the right endpoint of the signal is taken as the minimum value point; otherwise, the right endpoint of the signal is taken as the maximum value point. The method for determining the extreme points at the endpoints using the extreme value translation method is as follows: If the maximum point at the left endpoint is determined by the fixed extreme value method, then the first minimum point is translated to the left end as the minimum point at the left endpoint; conversely, if the minimum point at the left endpoint is determined by the fixed extreme value method, then the first maximum point is translated to the left end as the maximum point at the left endpoint. If the maximum point at the right endpoint is determined by the fixed extreme value method, then the last minimum point is translated to the right end as the minimum point at the right endpoint; conversely, if the minimum point at the right endpoint is determined by the fixed extreme value method, then the last maximum point is translated to the right end as the maximum point at the right endpoint. S3. Interpolate all the obtained local maxima using a cubic spline interpolation function to form the upper envelope of the signal, and interpolate all the local minima to form the lower envelope of the signal, and calculate the envelope average of the upper and lower envelopes. ; S4, using signals Subtract the envelope average line Get the difference ; S5. Determine the difference. Does it meet the set conditions for the intrinsic modulus function? If not, then the difference will be... Repeat steps S2 to S4 as a new signal until the condition is met, and record the difference. As the first eigenmode function, denoted as ; S6, using signals Subtract the first eigenmode function The residual components are obtained. ; S7. Transfer the residual components Repeat steps S2 to S6 as a new signal until the residual component is decomposed into a monotonic function or has one and only one extremum. At this point, the signal... It is decomposed into multiple intrinsic moduli and a residual component, i.e. ,in, Let i be the i-th intrinsic modulus function. This is the final residual component.

2. The method for processing aircraft turbine engine sensor signals based on CSI-EMD according to claim 1, characterized in that, The envelope average line of the upper and lower envelopes mentioned in step S3 The calculation formula is as follows: in, The upper envelope, The lower envelope, It is the envelope average line.

3. The method for processing aero-turbine engine sensor signals based on CSI-EMD according to claim 2, characterized in that, The conditions set in step S5 are: the number of local extrema and zero-crossing points must be equal or differ by at most one throughout the entire time range; and the mean of the upper envelope and lower envelope is zero at any time.

4. A CSI-EMD-based sensor signal processing system for aero-turbine engines, characterized in that, include: Signal augmentation module, extreme point determination module, envelope averaging module, intrinsic modulus function acquisition module, and decomposition termination module; The aforementioned signal enhancement module will process the aircraft turbine engine sensor signal. The data points are augmented by interpolating three data points between every two data points using cubic spline interpolation to obtain the augmented signal. ; The extreme point determination module uses quadratic polynomial interpolation to obtain the signal. Find all local maxima and local minima using the following method: Construct a function value of the target signal at three different points, which is related to the signal... Approximate quadratic polynomials In quadratic polynomial The extreme points are used as target signals The approximate extreme points are then determined using the fixed extreme point method to determine the left and right endpoints of the upper and lower envelopes, and the extreme point at the endpoints is determined using the extreme value translation method. The method for determining the left and right endpoints of the upper and lower envelopes using the fixed extreme value method is as follows: Select the two endpoints of the expanded time series signal. If the first extreme value is a maximum value, then the left endpoint of the signal is taken as the minimum value point; otherwise, the left endpoint of the signal is taken as the maximum value point. If the last extreme value is a maximum value, then the right endpoint of the signal is taken as the minimum value point; otherwise, the right endpoint of the signal is taken as the maximum value point. The method for determining the extreme points at the endpoints using the extreme value translation method is as follows: If the maximum point at the left endpoint is determined by the fixed extreme value method, then the first minimum point is translated to the left end as the minimum point at the left endpoint; conversely, if the minimum point at the left endpoint is determined by the fixed extreme value method, then the first maximum point is translated to the left end as the maximum point at the left endpoint. If the maximum point at the right endpoint is determined by the fixed extreme value method, then the last minimum point is translated to the right end as the minimum point at the right endpoint; conversely, if the minimum point at the right endpoint is determined by the fixed extreme value method, then the last maximum point is translated to the right end as the maximum point at the right endpoint. The envelope averaging module uses a cubic spline interpolation function to interpolate all local maxima to form the upper envelope of the signal, and interpolates all local minima to form the lower envelope of the signal, and then calculates the envelope average of the upper and lower envelopes. ; The intrinsic modulus function acquisition module uses signals Subtract the envelope average line Get the difference And determine the difference. Does it meet the set conditions for the intrinsic modulus function? If not, then the difference will be... Repeat steps S2 to S4 as a new signal until the condition is met, and record the difference. As the first eigenmode function, denoted as ; The decomposition termination module uses signals Subtract the first eigenmode function The residual components are obtained. ; the residual components Repeat steps S2 to S6 as a new signal until the residual component is decomposed into a monotonic function or has one and only one extremum. At this point, the signal... It is decomposed into multiple intrinsic moduli and a residual component, i.e. ,in, Let i be the i-th intrinsic modulus function. This is the final residual component.

5. The CSI-EMD-based aero-turbine engine sensor signal processing system according to claim 4, characterized in that, The envelope average line module describes the upper and lower envelope lines as envelope average lines. The calculation formula is as follows: in, The upper envelope, The lower envelope, It is the envelope average line.

6. The CSI-EMD-based aero-turbine engine sensor signal processing system according to claim 5, characterized in that, The conditions set in the intrinsic modulus function acquisition module are: the number of local extrema and zero-crossings must be equal or differ by at most one throughout the entire time range; and the mean of the upper envelope and lower envelope is zero at any time.

Images