Vibration diagnosis method and system for turboshaft engine main shaft bearing wear

Abstract

The present application relates to the technical field of bearing wear detection, and discloses a vibration diagnosis method and system for turboshaft engine main shaft bearing wear. The method includes: collecting the vibration signal of the turboshaft engine main shaft bearing; constructing a cumulative amount projection matrix of the vibration signal; reducing the dimension of the cumulative amount projection matrix to obtain a characteristic matrix of the vibration signal; judging whether the bearing is worn based on the characteristic matrix of the vibration signal; and determining the wear location if the bearing is worn, specifically including: calculating the time delay difference of the impact event in the vibration signal based on the characteristic matrix of the vibration signal; constructing a time delay difference feature vector of the impact event; matching the time delay difference feature vector with a preset time delay difference map, and determining the wear location based on the matching result. The present application can improve the accuracy of wear judgment and positioning precision, and has good adaptability under different working conditions.

Description

Technical Field

[0001] The present application relates to the technical field of bearing wear detection, and in particular to a vibration diagnosis method and system for turboshaft engine main shaft bearing wear. Background Art

[0002] In current industry practice, various technologies have been developed to monitor turboshaft engine mainshaft bearing wear. Each technology provides information for bearing condition assessment to a certain extent, but each has significant limitations. Traditional vibration monitoring technology uses vibration sensors installed in key engine locations to collect vibration signals generated by the mainshaft bearing during operation. The bearing's operating condition is determined based on the signal's amplitude, frequency, phase, and other characteristics. When a bearing wears, its vibration characteristics change. However, the actual turboshaft engine operating environment is complex, and vibration signals are easily affected by factors such as vibrations from other components, airflow pulsation, and electromagnetic interference. This results in significant background noise, which can drown out weak characteristic signals associated with bearing wear, making it difficult to accurately extract and identify. Existing vibration monitoring technologies are less sensitive to early, mild wear and often can only detect it after wear has progressed to a certain extent and the vibration characteristics have changed significantly. This hinders early fault warning and maintenance. Furthermore, the turboshaft engine's speed frequently fluctuates during operation, making it difficult for existing technologies to eliminate the impact of speed fluctuations on monitoring results when identifying wear locations. The signal changes caused by the speed change will interfere with the judgment of the wear location characteristics, making the wear identification and wear location judgment results inaccurate and unstable under different speed conditions.

[0003] For example, Chinese patent application CN110646202B discloses a method and apparatus for detecting bearing wear. The method comprises: acquiring a time-domain signal representing the time-varying vibration magnitude of the bearing in its radial direction; converting the time-domain signal into a frequency-domain signal representing the frequency-domain distribution of the bearing's radial vibration; and determining whether the bearing is worn using the frequency-domain signal and multiple stored frequency-domain signal segments for detecting bearing wear, wherein each of the multiple frequency-domain signal segments has characteristics of one of multiple types of wear that may occur on the bearing. This method and apparatus can improve the accuracy of bearing wear detection.

[0004] For example, the Chinese patent application with publication number CN113670615A discloses a bearing unit vibration test method and system, which belongs to the field of bearing detection technology. Vibration detection is performed on the bearing unit to be tested, and its vibration signal is obtained. The vibration signal is converted into a digital signal, and the digital signal is subjected to VMD decomposition to obtain multiple band-limited intrinsic modal functions. The power spectrum energy of each band-limited intrinsic modal function is calculated. When the order of the band-limited intrinsic modal function with the largest power spectrum energy among the band-limited intrinsic modal functions is greater than the set order, the bearing unit to be tested has a wear-related defect; or, if one of the band-limited intrinsic modal functions has a band-limited intrinsic modal function energy greater than the set threshold, the bearing unit to be tested has a wear-related defect. The use of this technical solution can quickly and batch detect bearing units, reducing the detection time.

[0005] The above existing technologies all have the problem raised by this background technology: under different speed conditions, the identification of wear and the judgment results of wear position are inaccurate and unstable.

[0006] The information disclosed in this background technology section is only intended to enhance the understanding of the overall background of the application and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art already known to ordinary technicians in this field. Summary of the Invention

[0007] The technical problem to be solved by this application is to overcome the defects of the existing technology and provide a vibration diagnosis method and system for the wear of the main shaft bearing of a turboshaft engine, so as to improve the accuracy of wear judgment and positioning accuracy and adaptability under different working conditions.

[0008] To solve the above technical problems, this application provides the following technical solutions:

[0009] In one aspect, the present application provides a vibration diagnosis method for turboshaft engine main shaft bearing wear, comprising the following steps:

[0010] Collecting vibration signals of a main shaft bearing of a turboshaft engine; constructing a cumulative projection matrix of the vibration signals;

[0011] Using truncated singular value decomposition to reduce the dimension of the cumulant projection matrix to obtain a characteristic matrix of the vibration signal;

[0012] determining whether the bearing is worn based on a characteristic matrix of the vibration signal;

[0013] If the bearing is worn, determine the location of the wear, including:

[0014] Calculating the time delay difference of the impact event in the vibration signal based on the characteristic matrix of the vibration signal;

[0015] Constructing a time delay difference feature vector of the impact event;

[0016] The time delay difference feature vector is matched with a preset time delay difference map, and the wear position is determined according to the matching result.

[0017] As a preferred solution of the vibration diagnosis method for turboshaft engine main shaft bearing wear described in the present application, wherein: the vibration signal is collected based on a vibration sensor; the cumulant projection matrix is ​​a diagonal slice projection matrix of the third-order cumulant of the vibration signal; the method for constructing the cumulant projection matrix of the vibration signal is as follows:

[0018] Calculate the third-order cumulants of the vibration signal at different time delays;

[0019] extracting diagonal slice data from the third-order cumulant;

[0020] The diagonal slice data are arranged into a cumulant projection matrix of the vibration signal.

[0021] As a preferred solution of the vibration diagnosis method for the turboshaft engine main shaft bearing wear described in the present application, wherein: the truncated singular value decomposition is used to reduce the dimension of the cumulant projection matrix to obtain the characteristic matrix of the vibration signal, specifically including:

[0022] Performing singular value decomposition on the cumulant projection matrix to obtain a singular value diagonal matrix and a left singular vector of the cumulant projection matrix;

[0023] Determine the number of truncated primary singular values, denoted as k; extract the largest k singular values ​​in the singular value diagonal matrix;

[0024] The left singular vectors corresponding to the largest k singular values ​​are arranged into a characteristic matrix of the vibration signal; any column in the characteristic matrix corresponds to a left singular vector.

[0025] As a preferred solution of the vibration diagnosis method for the turboshaft engine main shaft bearing wear described in the present application, wherein: based on the characteristic matrix of the vibration signal, determining whether the bearing is worn specifically includes:

[0026] Extracting characteristic parameters of the vibration signal based on a characteristic matrix of the vibration signal; the characteristic parameters include at least one of amplitude, frequency, kurtosis, and pulse factor;

[0027] Acquiring real-time operating parameters of a turboshaft engine main shaft; the real-time operating parameters include at least one of a rotational speed, a load, and a cumulative operating time;

[0028] After the characteristic parameters of the vibration signal and the real-time operating parameters of the turboshaft engine main shaft are normalized and encoded, they are input into the trained wear prediction model to obtain the judgment result of whether wear exists;

[0029] The wear prediction model is any one of a support vector machine, a random forest model, and a gradient boosting decision tree, and the output is a binary judgment result of whether the bearing is worn or not.

[0030] As a preferred embodiment of the vibration diagnosis method for the turboshaft engine main shaft bearing wear described in the present application, the time delay difference is the time difference between different impact events detected by the sensor in the characteristic matrix of the vibration signal; and the time delay difference of the impact event in the vibration signal is calculated, specifically including:

[0031] Performing Hilbert envelope demodulation on each column of the left singular vectors in the characteristic matrix of the vibration signal to obtain a shock pulse sequence corresponding to each column of the left singular vectors;

[0032] Calculate the cross-correlation coefficient between any two shock pulse sequences at different lag times and form a cross-correlation coefficient sequence corresponding to the two shock pulse sequences;

[0033] The lag time corresponding to the maximum value in the mutual correlation coefficient sequence is extracted as the time delay difference between the impact events corresponding to the two corresponding impact pulse sequences.

[0034] As a preferred solution of the vibration diagnosis method for the turboshaft engine main shaft bearing wear described in the present application, wherein: constructing the time delay difference feature vector of the impact event specifically includes:

[0035] Construct a time delay difference matrix of impact events in the vibration signal; the time delay difference matrix is ​​a strictly upper triangular matrix, and the element in the i-th row and j-th column of the time delay difference matrix is ​​the time delay difference between the i-th impact event and the j-th impact event in the vibration signal; wherein i is greater than j, the value range of i is 2, 3, ..., m, m is the number of columns of the characteristic matrix of the vibration signal; the value range of j is 1, 2, ..., m-1;

[0036] Non-zero elements in the time delay difference matrix are extracted row by row and arranged into a time delay difference feature vector of the impact event.

[0037] As a preferred embodiment of the vibration diagnosis method for turboshaft engine main shaft bearing wear described in the present application, the time delay difference map includes a reference time delay difference feature vector corresponding to each known wear position; the method for constructing the time delay difference map specifically includes: collecting historical vibration signals of bearings at known wear positions, calculating the time delay differences of impact events in the historical vibration signals and constructing a time delay difference feature vector as a reference time delay difference feature vector corresponding to the wear position;

[0038] Matching the time delay difference feature vector with a preset time delay difference map, and determining the wear position based on the matching result specifically includes: calculating the similarity between the time delay difference feature vector and each reference time delay difference feature vector in the time delay difference map; the wear position corresponding to the reference time delay difference feature vector with the highest similarity to the time delay difference feature vector is the current wear position of the turboshaft engine main shaft bearing.

[0039] As a preferred embodiment of the vibration diagnosis method for the turboshaft engine main shaft bearing wear described in the present application, the method for calculating the similarity between the time delay difference feature vector and any reference time delay difference feature vector in the time delay difference map is as follows:

[0040] Normalizing the delay difference feature vector and the reference delay difference feature vector respectively;

[0041] Construct a distance matrix; the element in the pth row and qth column of the distance matrix is ​​recorded as ; is the Euclidean distance between the pth element in the delay difference feature vector and the qth element in the reference delay difference feature vector; the value range of p is 1, 2, ..., , is the length of the delay difference feature vector; the value range of q is 1, 2, ..., , is the length of the reference delay difference feature vector;

[0042] Initialize a cumulative distance matrix; the dimension of the cumulative distance matrix is ​​the same as that of the distance matrix;

[0043] Iteratively calculating cumulative distance values ​​based on the distance matrix, and filling the cumulative distance matrix row by row based on the cumulative distance values;

[0044] The cumulative distance value at the lower right corner of the cumulative distance matrix is ​​extracted as the similarity between the delay difference feature vector and the reference delay difference feature vector.

[0045] As a preferred solution of the vibration diagnosis method for the turboshaft engine main shaft bearing wear described in the present application, wherein: cumulative distance values ​​are iteratively calculated based on a distance matrix, and elements of the cumulative distance matrix are filled row by row based on the cumulative distance values, specifically including:

[0046] Let the element in row p and column q of the cumulative distance matrix be ; Fill the cumulative distance matrix row by row, and fill each row with elements from left to right; for elements with arbitrary values ​​of p and q , if there is a cumulative distance matrix 、 、 If at least one of The value of and 、 、 Otherwise, The value of .

[0047] In a second aspect, the present application provides a vibration diagnosis system for turboshaft engine main shaft bearing wear, comprising a data acquisition module, a data processing module, a wear diagnosis module, and a wear positioning module; wherein:

[0048] The data acquisition module is used to collect the vibration signals of the turboshaft engine main shaft bearing and collect the real-time working parameters of the turboshaft engine main shaft, including speed, load, and accumulated running time;

[0049] The data processing module is used to construct a characteristic matrix of the vibration signal and extract characteristic parameters of the vibration signal based on the characteristic matrix, including amplitude, frequency, kurtosis, and pulse factor;

[0050] The wear diagnosis module determines whether the bearing is worn based on the real-time operating parameters of the turboshaft engine main shaft and the characteristic parameters of the vibration signal, and sends a wear location instruction to the wear location module if the bearing is worn.

[0051] The wear location module is used to respond to the wear location instruction and determine the wear location; the wear location module is configured with a delay difference map; the wear location module constructs a delay difference feature vector of the impact event based on the characteristic matrix of the vibration signal, matches the delay difference feature vector with the delay difference map, and determines the wear location according to the matching result.

[0052] Compared with the prior art, the beneficial effects achieved by this application are as follows:

[0053] This application realizes structured encoding of the high-order statistical characteristics of the original vibration signal by constructing a diagonal slice projection matrix of the third-order cumulative amount of the vibration signal, effectively suppresses Gaussian noise, amplifies the impact components related to wear in the vibration signal, and makes the vibration characteristics of wear more obvious.

[0054] Based on the characteristic matrix of the vibration signal, multiple characteristic parameters such as amplitude, frequency, kurtosis, and pulse factor are extracted. Combined with real-time working parameters such as the speed, load, and cumulative running time of the turboshaft engine main shaft, a joint analysis is performed through a machine learning model to comprehensively reflect the operating status of the bearing, thereby improving the accuracy of judging whether the bearing is worn.

[0055] By calculating the cross-correlation coefficients between different shock pulse sequences, the time delay difference of the shock event is accurately extracted. A time delay difference feature vector is constructed and matched with a preset time delay difference map, enabling precise determination of the bearing wear location. This method leverages the direct correlation between the shock event propagation path and time domain characteristics and the wear location, effectively overcoming the influence of speed fluctuations on wear location determination. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be derived from these drawings without inventive effort. Among them:

[0057] Figure 1 A flow chart of the vibration diagnosis method for turboshaft engine main shaft bearing wear provided in this application;

[0058] Figure 2 Schematic diagram of the vibration diagnosis system for turboshaft engine main shaft bearing wear provided in this application. DETAILED DESCRIPTION

[0059] The technical solution of the present application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations on the technical solution of the present application. Unless there is a conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0060] Example 1

[0061] This embodiment introduces a vibration diagnosis method for the wear of the main shaft bearing of a turboshaft engine. Figure 1 , the method comprises the following steps:

[0062] Collecting vibration signals of a main shaft bearing of a turboshaft engine; constructing a cumulative projection matrix of the vibration signals;

[0063] The vibration signal is collected based on a vibration sensor. In this embodiment, a vibration sensor is preferably installed on the bearing seat of the turboshaft engine to accurately obtain vibration signals related to the main shaft bearing.

[0064] The cumulant projection matrix is ​​a diagonal slice projection matrix of the third-order cumulant of the vibration signal. The method for constructing the cumulant projection matrix of the vibration signal is as follows:

[0065] The third-order cumulants of the vibration signal at different time delays are calculated; diagonal slice data in the third-order cumulants are extracted; and the diagonal slice data are arranged into a cumulant projection matrix of the vibration signal.

[0066] This application prefers the Hankel matrix as the specific form of the cumulant projection matrix. Third-order cumulants are high-order statistics. Arranging the diagonal slice data in the third-order cumulants into a matrix according to the Hankel structure achieves structured encoding of the high-order statistical characteristics of the original vibration signal. The nonlinear impact characteristics of the original vibration signal are projected into the space composed of high-order statistics, achieving Gaussian noise suppression (the high-order cumulants of Gaussian noise are zero) and amplifying the impact component of wear-related non-Gaussian noise in the vibration signal.

[0067] The truncated singular value decomposition is used to reduce the dimension of the cumulant projection matrix to obtain the characteristic matrix of the vibration signal; specifically, the method includes:

[0068] Performing singular value decomposition on the cumulant projection matrix to obtain a singular value diagonal matrix and a left singular vector of the cumulant projection matrix; performing singular value decomposition on the cumulant projection matrix using a mathematical algorithm software library to obtain a singular value diagonal matrix and a left singular matrix; wherein the singular value diagonal matrix contains singular values, and any column in the left singular matrix is ​​a left singular vector;

[0069] Determine the number of truncated principal singular values, denoted as k; extract the largest k singular values ​​in the singular value diagonal matrix; in an embodiment of the present application, the method for determining the number of truncated principal singular values ​​is as follows: calculate the total energy of all elements in the cumulant projection matrix; if the sum of the energies of the elements in the k column vectors in the cumulant projection matrix corresponding to the largest k singular values ​​accounts for more than 95% of the total energy of all elements in the cumulant projection matrix, then k is the number of truncated principal singular values.

[0070] The left singular vectors corresponding to the k largest singular values ​​are arranged into a vibration signal feature matrix; each column in the feature matrix corresponds to a left singular vector. Through singular value truncation, only the left singular vectors corresponding to the k main singular values ​​are retained, achieving feature dimensionality reduction. This removes minor information such as noise corresponding to smaller singular values ​​while retaining wear-related impact events corresponding to larger singular values, thereby highlighting the signal characteristics.

[0071] The impact signal generated by bearing wear propagates directly or is reflected to the sensor along various physical paths, such as along the inner race, outer race, and rolling elements. This embodiment calculates a diagonal slice matrix of the vibration signal's third-order cumulants, with each column corresponding to a third-order cumulant slice with a different time delay. This matrix implicitly captures the multipath propagation delay information of the impact event in the vibration signal. The dominant impact event is extracted through truncated singular value decomposition. The column space of left singular vectors reflects the physical propagation path of the impact signal. Each column of left singular vectors represents a component of the impact event.

[0072] Determining whether the bearing is worn based on the characteristic matrix of the vibration signal; specifically, comprising:

[0073] Extracting characteristic parameters of the vibration signal based on a characteristic matrix of the vibration signal; the characteristic parameters include at least one of amplitude, frequency, kurtosis, and pulse factor;

[0074] Acquiring real-time operating parameters of a turboshaft engine main shaft; the real-time operating parameters include at least one of a rotational speed, a load, and a cumulative operating time;

[0075] After the characteristic parameters of the vibration signal and the real-time operating parameters of the turboshaft engine main shaft are normalized and encoded, they are input into the trained wear prediction model to obtain the judgment result of whether wear exists;

[0076] The wear prediction model is any one of a support vector machine, a random forest model, and a gradient boosting decision tree, and the output is a binary judgment result of whether the bearing is worn or not.

[0077] By constructing a cumulative projection matrix of the vibration signal and further performing dimensionality reduction on it, the resulting feature matrix highlights impact events caused by wear in the vibration signal. Analysis of the feature matrix can determine the presence and extent of bearing wear. By analyzing historical vibration data from a large number of bearings with known wear conditions and calibrating experimental data, the patterns and characteristics of characteristic parameters corresponding to typical wear patterns can be determined. For example, if the amplitude extracted from the feature matrix is ​​significantly higher than the amplitude under normal operating conditions and exceeds a preset threshold, it indicates bearing wear. Bearing wear can also lead to the appearance or enhancement of specific frequency components. For example, inner race wear peaks at frequencies that are multiples of the rotational frequency, while outer race wear exhibits significant frequency components at characteristic frequencies associated with the rolling elements. Spectral analysis indicates that the bearing is worn if the energy of the frequency components extracted from the feature matrix increases significantly at these frequencies. Furthermore, if the kurtosis value extracted from the feature matrix exceeds the normal range or the pulse factor increases significantly, and the trend of the corresponding indicators in known wear cases is consistent, bearing wear is present.

[0078] This embodiment preferably uses a machine learning model (such as a support vector machine) to automatically learn the laws and characteristics of the above-mentioned characteristic parameters, while considering real-time operating parameters such as the speed, load, and cumulative operating time of the turboshaft engine main shaft to perform a joint analysis, thereby improving the accuracy of wear judgment.

[0079] If the bearing is not worn, continue to collect the vibration signal of the turboshaft engine main shaft bearing and continue to judge whether the bearing is worn; if the bearing is worn, determine the wear position, and continue to collect the vibration signal of the turboshaft engine main shaft bearing and perform subsequent processing and judgment.

[0080] The method for determining the location of wear is as follows:

[0081] Calculating the time delay difference of the impact event in the vibration signal based on the characteristic matrix of the vibration signal;

[0082] The time delay difference is the time difference between different impact events detected by the sensor in the characteristic matrix of the vibration signal; and calculating the time delay difference of the impact event in the vibration signal specifically includes:

[0083] Hilbert envelope demodulation is performed on each column of left singular vectors in the characteristic matrix of the vibration signal to obtain a shock pulse sequence corresponding to each column of left singular vectors; Hilbert transform is performed on each column of left singular vectors, and its envelope signal is calculated as the shock pulse sequence to eliminate high-frequency carrier interference and highlight the time domain characteristics of the shock event in the vibration signal.

[0084] Calculate the cross-correlation coefficient between any two shock pulse sequences at different lag times and form a cross-correlation coefficient sequence corresponding to the two shock pulse sequences;

[0085] The lag time corresponding to the maximum value in the mutual correlation coefficient sequence is extracted as the time delay difference between the impact events corresponding to the two corresponding impact pulse sequences.

[0086] When bearings wear, the impact events caused by wear travel along different propagation paths to the vibration sensor. While the waveforms of the impact pulse sequences corresponding to these different propagation paths are similar, they exhibit time delay differences due to the different times they arrive at the sensor. By calculating the cross-correlation coefficients of two impact pulse sequences at different lag times, we can determine the time difference between the impact events corresponding to these two sequences arriving at the sensor, known as the time delay. This time delay reflects the time difference in the propagation paths of the wear-induced impact events within the bearing. When the lag time equals the time delay difference, the cross-correlation coefficient is maximized, indicating that the waveforms of the impact pulse sequences at that lag time are most similar.

[0087] Constructing a time delay difference characteristic vector of the impact event; specifically comprising: constructing a time delay difference matrix of the impact event in the vibration signal; the time delay difference matrix is ​​a strictly upper triangular matrix, and the element in the i-th row and j-th column of the time delay difference matrix is ​​the time delay difference between the i-th impact event and the j-th impact event in the vibration signal; wherein i is greater than j, the value range of i is 2, 3, ..., m, m is the number of columns of the characteristic matrix of the vibration signal, that is, the number of impact events in the characteristic matrix; the value range of j is 1, 2, ..., m-1;

[0088] Non-zero elements in the time delay difference matrix are extracted row by row and arranged into a time delay difference feature vector of the impact event.

[0089] Bearing wear triggers periodic impacts as the bearing rotates. The propagation path and time-domain characteristics of the impact events are directly related to the wear location, and the time delay difference records the wear location information. As the speed changes, the time difference between the periodic impacts also changes, meaning the magnitude of the impact event's time delay difference also changes. This embodiment constructs a time delay difference matrix for the impact events in the vibration signal. This matrix is ​​composed of the time delay differences between all two impact modes, and then constructs a time delay difference eigenvector for the impact event. Even if any of the time delay differences fluctuates with the speed, the relative magnitudes of the multiple time delay differences in the time delay difference eigenvector remain constant (all delay differences are scaled proportionally). Therefore, the time delay difference eigenvector retains the wear location information while eliminating the effects of speed fluctuations on the wear location information, making the wear location determination more accurate.

[0090] The time delay difference feature vector is matched with a preset time delay difference map, and the wear position is determined according to the matching result.

[0091] The time delay difference map includes a reference time delay difference feature vector corresponding to each known wear position; the method of constructing the time delay difference map specifically includes: collecting historical vibration signals of bearings at known wear positions, calculating the time delay difference of impact events in the historical vibration signals and constructing a time delay difference feature vector as a reference time delay difference feature vector for the corresponding wear position; for wear positions where historical vibration signals are missing, the historical vibration signals are supplemented through simulation experiments, and the corresponding reference time delay difference feature vector is constructed.

[0092] Matching the time delay difference feature vector with a preset time delay difference map, and determining the wear position based on the matching result specifically includes: calculating the similarity between the time delay difference feature vector and each reference time delay difference feature vector in the time delay difference map; the wear position corresponding to the reference time delay difference feature vector with the highest similarity to the time delay difference feature vector is the current wear position of the turboshaft engine main shaft bearing.

[0093] The method for calculating the similarity between the delay difference feature vector and any reference delay difference feature vector in the delay difference map is as follows:

[0094] The delay difference eigenvector and the reference delay difference eigenvector are normalized respectively; the scaling of the delay difference caused by the rotation speed change is eliminated by the normalization process, and the relative proportion of the delay difference is retained.

[0095] Construct a distance matrix; the element in the pth row and qth column of the distance matrix is ​​recorded as ; is the Euclidean distance between the pth element in the delay difference feature vector and the qth element in the reference delay difference feature vector; the value range of p is 1, 2, ..., , is the length of the delay difference feature vector; the value range of q is 1, 2, ..., , is the length of the reference delay difference feature vector;

[0096] Initialize a cumulative distance matrix; the dimension of the cumulative distance matrix is ​​the same as that of the distance matrix;

[0097] Iteratively calculating cumulative distance values ​​based on the distance matrix, and filling the cumulative distance matrix row by row based on the cumulative distance values;

[0098] The cumulative distance value at the lower right corner of the cumulative distance matrix is ​​extracted as the similarity between the delay difference feature vector and the reference delay difference feature vector.

[0099] Iteratively calculating cumulative distance values ​​based on the distance matrix, and filling the cumulative distance matrix row by row based on the cumulative distance values, specifically including:

[0100] Let the element in row p and column q of the cumulative distance matrix be ; Fill the cumulative distance matrix row by row, and fill each row with elements from left to right; for elements with arbitrary values ​​of p and q , if there is a cumulative distance matrix 、 、 If at least one of The value of and 、 、 Otherwise, The value of .

[0101] By iteratively calculating cumulative distance values, each calculation accumulates the minimum Euclidean distance between the delay difference feature vector and the reference delay difference feature vector. This dynamically aligns the time axes of the cumulative delay difference feature vector and the reference delay difference feature vector, ignoring the scaling of the delay difference due to speed changes and focusing on comparing the waveform similarity of the delay difference feature vectors. This achieves the optimal match between the delay difference feature vector calculated in real-time detection and the delay difference map through a nonlinear alignment path, avoiding misjudgment of wear location due to changes in delay difference size.

[0102] Speed ​​fluctuations will cause the absolute value of the time delay difference to scale overall, but the periodic characteristics of the time delay difference corresponding to different wear positions remain unchanged. For example, when the inner ring is worn, the impact event is transmitted through the rotating rolling element, modulated by the load zone, and the periodic change of the time delay difference is affected by the speed and bearing geometric parameters. When the outer ring is worn, the impact event is mainly transmitted along the stationary outer ring, the path is fixed, and the periodicity of the time delay difference is relatively stable. When the rolling element is worn, the impact event occurs periodically with the revolution of the rolling element, and the time delay difference shows a quasi-periodic oscillation. Based on the method provided in the present application, in the preset time delay difference map, multiple time delay differences in the reference time delay difference feature vector corresponding to outer ring wear are close to the theoretical time delay difference of outer ring wear, for example, 0.01 seconds; during actual detection, multiple time delay differences in the time delay difference feature vector constructed by the vibration signal are close to integer multiples of 0.01 seconds, and the highest similarity with the reference time delay difference feature vector of outer ring wear is calculated, and the wear position is determined to be the outer ring. This method successfully converts the spatial characteristics of the wear position into the relative proportion of the time delay difference, and achieves high-precision wear positioning under variable speed conditions through the elastic alignment capability of cumulative distance calculation similarity.

[0103] Example 2

[0104] This embodiment is the second embodiment of the present application; it is based on the same inventive concept as embodiment 1, and Figure 2 This embodiment introduces a vibration diagnosis system for turboshaft engine main shaft bearing wear, including a data acquisition module, a data processing module, a wear diagnosis module, and a wear positioning module; wherein:

[0105] The data acquisition module is used to collect vibration signals from the turboshaft engine main shaft bearings and collect real-time operating parameters of the turboshaft engine main shaft, including speed, load, and accumulated running time; the module includes a vibration sensor, which continuously samples at a set sampling frequency to obtain vibration signals related to the main shaft bearings, providing a data basis for subsequent analysis.

[0106] The data processing module is used to construct a characteristic matrix for the vibration signal and, based on this matrix, extract characteristic parameters of the vibration signal, including amplitude, frequency, kurtosis, and impulse factor. This module first constructs a cumulant projection matrix, calculates the third-order cumulants of the vibration signal at different time delays, extracts diagonal slice data, and arranges them into a matrix using a Hankel structure. This achieves structured encoding of the high-order statistical characteristics of the original vibration signal, suppresses Gaussian noise, and amplifies wear-related impact components. Truncated singular value decomposition is then used to reduce the dimensionality of the cumulant projection matrix. The left singular vectors corresponding to the largest k singular values ​​are extracted to form the vibration signal's characteristic matrix, highlighting wear-related impact events and providing effective data for wear assessment and location. The module also extracts characteristic parameters based on the vibration signal's characteristic matrix for wear identification.

[0107] The wear diagnosis module determines bearing wear based on the turboshaft engine's real-time operating parameters and the characteristic parameters of the vibration signal. If the bearing is worn, it sends a wear location instruction to the wear location module. After normalizing and encoding these parameters, they are input into a trained wear prediction model, such as a support vector machine, random forest model, or gradient boosting decision tree, to output a determination of bearing wear. This module comprehensively analyzes the characteristic parameters of the vibration signal and the engine's real-time operating parameters, and uses the trained machine learning model to accurately determine bearing wear.

[0108] The wear location module is used to respond to the wear location instruction and determine the wear location; the wear location module is configured with a delay difference map; the wear location module constructs a delay difference feature vector of the impact event based on the characteristic matrix of the vibration signal, matches the delay difference feature vector with the delay difference map, and determines the wear location based on the matching result. After determining that the bearing is worn, the module performs Hilbert envelope demodulation on each column of the left singular vector of the vibration signal characteristic matrix to obtain an impact pulse sequence, and then calculates the mutual correlation coefficient of any two impact pulse sequences at different lag times, finds the lag time corresponding to the maximum value as the delay difference, and constructs a delay difference feature vector. The delay difference feature vector is matched with the delay difference map, and the wear location is determined by calculating the similarity.

[0109] The specific functions of the above modules are realized by referring to the relevant contents of the vibration diagnosis method for turboshaft engine main shaft bearing wear described in Example 1, and are not described in detail here.

[0110] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0111] The above describes the embodiments of the present application in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also make many forms without departing from the purpose and scope of protection of this application, all of which are protected by this application.

Claims

1. A vibration diagnosis method for turboshaft engine main shaft bearing wear, characterized by: The following steps are involved: Collecting vibration signals of a main shaft bearing of a turboshaft engine; constructing a cumulative projection matrix of the vibration signals; Using truncated singular value decomposition to reduce the dimension of the cumulant projection matrix to obtain a characteristic matrix of the vibration signal; determining whether the bearing is worn based on a characteristic matrix of the vibration signal; If the bearing is worn, determine the location of the wear, including: Based on the characteristic matrix of the vibration signal, calculating the time delay difference of the impact event in the vibration signal; the time delay difference is the time difference between different impact events detected by the sensor in the characteristic matrix of the vibration signal; calculating the time delay difference of the impact event in the vibration signal specifically includes: Performing Hilbert envelope demodulation on each column of the left singular vectors in the characteristic matrix of the vibration signal to obtain a shock pulse sequence corresponding to each column of the left singular vectors; Calculate the cross-correlation coefficient between any two shock pulse sequences at different lag times and form a cross-correlation coefficient sequence corresponding to the two shock pulse sequences; extracting the lag time corresponding to the maximum value in the mutual correlation coefficient sequence as the time delay difference between the impact events corresponding to the two corresponding impact pulse sequences; Constructing a time delay difference feature vector of the impact event; The time delay difference feature vector is matched with a preset time delay difference map, and the wear position is determined according to the matching result; specifically, the similarity between the time delay difference feature vector and each reference time delay difference feature vector in the time delay difference map is calculated respectively; the wear position corresponding to the reference time delay difference feature vector with the highest similarity to the time delay difference feature vector is the current wear position of the turboshaft engine main shaft bearing.

2. The vibration diagnosis method for turboshaft engine main shaft bearing wear according to claim 1, characterized in that: The vibration signal is collected based on a vibration sensor; the cumulant projection matrix is ​​a diagonal slice projection matrix of the third-order cumulant of the vibration signal; the method for constructing the cumulant projection matrix of the vibration signal is as follows: Calculate the third-order cumulants of the vibration signal at different time delays; extracting diagonal slice data from the third-order cumulant; The diagonal slice data are arranged into a cumulant projection matrix of the vibration signal.

3. The vibration diagnosis method for turboshaft engine main shaft bearing wear according to claim 2, characterized in that: The truncated singular value decomposition is used to reduce the dimension of the cumulant projection matrix to obtain the characteristic matrix of the vibration signal, specifically including: Performing singular value decomposition on the cumulant projection matrix to obtain a singular value diagonal matrix and a left singular vector of the cumulant projection matrix; Determine the number of truncated primary singular values, denoted as k; extract the largest k singular values ​​in the singular value diagonal matrix; The left singular vectors corresponding to the largest k singular values ​​are arranged into a characteristic matrix of the vibration signal; any column in the characteristic matrix corresponds to a left singular vector.

4. The vibration diagnosis method for turboshaft engine main shaft bearing wear according to claim 3, characterized in that: Based on the characteristic matrix of the vibration signal, determining whether the bearing is worn specifically includes: Extracting characteristic parameters of the vibration signal based on a characteristic matrix of the vibration signal; the characteristic parameters include at least one of amplitude, frequency, kurtosis, and pulse factor; Acquiring real-time operating parameters of a turboshaft engine main shaft; the real-time operating parameters include at least one of a rotational speed, a load, and a cumulative operating time; After the characteristic parameters of the vibration signal and the real-time operating parameters of the turboshaft engine main shaft are normalized and encoded, they are input into the trained wear prediction model to obtain the judgment result of whether wear exists; The wear prediction model is any one of a support vector machine, a random forest model, and a gradient boosting decision tree, and the output is a binary judgment result of whether the bearing is worn or not.

5. The vibration diagnosis method for turboshaft engine main shaft bearing wear according to claim 4, characterized in that: Constructing the time delay difference feature vector of the impact event specifically includes: Construct a time delay difference matrix of impact events in the vibration signal; the time delay difference matrix is ​​a strictly upper triangular matrix, and the element in the i-th row and j-th column of the time delay difference matrix is ​​the time delay difference between the i-th impact event and the j-th impact event in the vibration signal; wherein i is greater than j, the value range of i is 2, 3, ..., m, m is the number of columns of the characteristic matrix of the vibration signal; the value range of j is 1, 2, ..., m-1; Non-zero elements in the time delay difference matrix are extracted row by row and arranged into a time delay difference feature vector of the impact event.

6. The vibration diagnosis method for turboshaft engine main shaft bearing wear according to claim 5, characterized in that: The time delay difference map includes a reference time delay difference feature vector corresponding to each known wear position; The method of constructing the time delay difference map specifically includes: collecting historical vibration signals of bearings with known wear positions, calculating the time delay differences of impact events in the historical vibration signals and constructing a time delay difference feature vector as a reference time delay difference feature vector corresponding to the wear position.

7. The vibration diagnosis method for turboshaft engine main shaft bearing wear according to claim 6, characterized in that: The method for calculating the similarity between the delay difference feature vector and any reference delay difference feature vector in the delay difference map is as follows: Normalizing the delay difference feature vector and the reference delay difference feature vector respectively; Construct a distance matrix; the element in the pth row and qth column of the distance matrix is ​​recorded as ; is the Euclidean distance between the pth element in the delay difference feature vector and the qth element in the reference delay difference feature vector; the value range of p is 1, 2, ..., , is the length of the delay difference feature vector; the value range of q is 1, 2, ..., , is the length of the reference delay difference feature vector; Initialize a cumulative distance matrix; the dimension of the cumulative distance matrix is ​​the same as that of the distance matrix; Iteratively calculating cumulative distance values ​​based on the distance matrix, and filling the cumulative distance matrix row by row based on the cumulative distance values; The cumulative distance value at the lower right corner of the cumulative distance matrix is ​​extracted as the similarity between the delay difference feature vector and the reference delay difference feature vector.

8. The vibration diagnosis method for turboshaft engine main shaft bearing wear according to claim 7, characterized in that: Iteratively calculating cumulative distance values ​​based on the distance matrix, and filling the cumulative distance matrix row by row based on the cumulative distance values, specifically including: Let the element in row p and column q of the cumulative distance matrix be ; Fill the elements of the cumulative distance matrix row by row, and fill each row with elements from left to right; for elements with arbitrary values ​​of p and q , if there is a cumulative distance matrix 、 、 If at least one of The value of and 、 、 Otherwise, The value of .

9. A vibration diagnosis system for turboshaft engine main shaft bearing wear, which is used to implement the vibration diagnosis method for turboshaft engine main shaft bearing wear according to any one of claims 1 to 8, characterized in that: It includes data acquisition module, data processing module, wear diagnosis module and wear positioning module; among which: The data acquisition module is used to collect the vibration signals of the turboshaft engine main shaft bearing and collect the real-time working parameters of the turboshaft engine main shaft, including speed, load, and accumulated running time; The data processing module is used to construct a characteristic matrix of the vibration signal and extract characteristic parameters of the vibration signal based on the characteristic matrix, including amplitude, frequency, kurtosis, and pulse factor; The wear diagnosis module determines whether the bearing is worn based on the real-time operating parameters of the turboshaft engine main shaft and the characteristic parameters of the vibration signal, and sends a wear location instruction to the wear location module if the bearing is worn. The wear location module is used to respond to the wear location instruction and determine the wear location; the wear location module is configured with a delay difference map; the wear location module constructs a delay difference feature vector of the impact event based on the characteristic matrix of the vibration signal, matches the delay difference feature vector with the delay difference map, and determines the wear location according to the matching result.

Images