Fault signal phase synchronization method and device based on data fusion algorithm

Abstract

The invention relates to the technical field of data analysis, in particular to a fault signal phase synchronization method and device based on a data fusion algorithm. The method comprises the following steps: preprocessing a collected bearing operation data signal; the preprocessed data signals are fused, and reliable phase offset estimation of each operation period and position is calculated; the fused data signals are processed through waveform matching and a phase correction algorithm, phase deviation caused by driving speed and load changes is eliminated, and data signals after phase synchronization are obtained; and based on the data signals after phase synchronization, bearing fault features are extracted, and a phase consistency coefficient is calculated. According to the technical scheme, the identifiability of fault signal characteristics is improved through multi-sensor data fusion and an advanced phase synchronization algorithm, and an important technical means is provided for fault diagnosis of rotating machinery.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and more specifically, to a method and apparatus for synchronizing the phase of fault signals based on a data fusion algorithm. Background Technology

[0002] Due to their periodic and repetitive motion, rotating machinery structures are particularly prone to structural wear and cracking under prolonged high-load conditions. These problems are especially pronounced in bearings, particularly in precision equipment such as machine tools and engines. The operational reliability of bearing structures is crucial to production safety, output efficiency, and the personal safety of operators. Failure to detect and address faults in a timely manner can lead to equipment downtime, production interruptions, and serious safety accidents. Statistics show that over 30% of rotating machinery failures are related to rolling bearing structural problems, primarily caused by manufacturing errors, poor internal lubrication, and defects in coupled operating forces.

[0003] Existing research utilizes various sensors to monitor the operating status of bearings, including thermal imaging, noise, stator current, vibration, and multi-sensor fusion. Vibration signals are generated from all components that produce relative motion, such as balls and cages, and contain operating status information for each bearing component, thus serving as an important information source for structural fault analysis. However, traditional signal feature extraction methods are susceptible to interference from external noise and unstable drive motor speeds, thus requiring advanced data processing methods to correct the raw data and enhance the identifiability of fault features.

[0004] Time Synchronization Analysis (TSA) exhibits comb-like filtering characteristics, enabling truncation and denoising of target signals to extract quasi-periodic features from complex signals. Although this method has been around for over 30 years, it has remained widely used in fault diagnosis of multi-stage gear systems and rotating bearing structures for the past decade. Many researchers have dedicated themselves to improving or applying TSA to solve various problems. Some scholars have used TSA as a data preprocessing method in the Maximum Cyclic Stationary Blind Deconvolution (CYCBD) iterative algorithm. Others have employed TSA to preprocess raw data when using nonlinear autoregressive exogenous (NARX) models for bearing fault time series prediction.

[0005] While the TSA method performs well in fault diagnosis, its synchronization effectiveness relies on time-scale calibration provided by a tachometer due to fluctuations in speed and load during equipment operation. Most researchers install tachometers in their experiments to selectively truncate and average the measured vibration signals during analysis. However, in many engineering scenarios, installing a tachometer is impractical, resulting in the inability to obtain frequency information of interest, particularly the high-frequency information after TSA processing. The phase error caused by direct signal truncation accumulates due to continuous fluctuations in speed and load, ultimately affecting the overall fault analysis results.

[0006] Currently, most researchers use single metrics such as Pearson correlation coefficient, kurtosis, or cross-power spectrum to quantify similarity and perform phase compensation for phase synchronization improvements in TSA algorithms. While these methods are helpful in achieving good alignment results, there is still room for improvement in their application in real-world engineering scenarios. Summary of the Invention

[0007] This invention provides a fault signal phase synchronization method and apparatus based on a data fusion algorithm, which at least solves the phase offset problem of existing rotating structure fault signals.

[0008] According to an embodiment of the present invention, a fault signal phase synchronization method based on a data fusion algorithm is provided, comprising the following steps:

[0009] S101: Preprocess the acquired bearing operation data signals;

[0010] S102: Fuse the preprocessed data signals and calculate a reliable phase offset estimate for each operating cycle and position;

[0011] S103: The fused data signal is processed by waveform matching and phase correction algorithms to eliminate the phase offset caused by changes in drive speed and load, and obtain the phase-synchronized data signal.

[0012] S104: Based on the data signal after phase synchronization, extract bearing fault characteristics and calculate the phase consistency coefficient.

[0013] Furthermore, the method further includes the following steps before step S101:

[0014] Multiple sensors are used to collect vibration signals from the bearing to obtain detailed bearing operation data.

[0015] Further, step S101 includes: performing super-resolution sampling on the acquired original vibration signal of the bearing operation to improve the temporal and spatial resolution of the signal.

[0016] Further, step S102 includes:

[0017] Construct a synchronization information matrix containing all operation cycles and channel data, and record relative phase offset and waveform similarity data;

[0018] Data signals from multiple sensors are fused to calculate a reliable phase offset estimate for each operating cycle and position, forming a synchronization model.

[0019] Furthermore, the window moves along the signal time axis at a set step size, and the waveform similarity at each window position is calculated step by step.

[0020] Furthermore, the synchronization information matrix is ​​a high-dimensional matrix, where each element specifically records the relative phase deviation and the acceleration signal of the corresponding channel under a specific period and channel.

[0021] Furthermore, in step S102, the internal error between channels within each cycle is estimated, and the data from different channels are fused to form a phase synchronization model.

[0022] Furthermore, in step S103, the calculated relative phase error and channel phase error are used to perform the final phase correction on the signal. Two interpolations are used to ensure phase synchronization of data from different channels and phase synchronization of data from different operating cycles.

[0023] According to another embodiment of the present invention, a fault signal phase synchronization device based on a data fusion algorithm is provided, comprising:

[0024] The preprocessing unit is used to preprocess the acquired bearing operation data signals;

[0025] The data fusion unit is used to fuse the preprocessed data signals and calculate a reliable phase offset estimate for each operating cycle and position.

[0026] The phase synchronization unit is used to process the fused data signal through waveform matching and phase correction algorithms to eliminate the phase offset caused by changes in drive speed and load, and obtain the phase-synchronized data signal.

[0027] The phase consistency unit is used to extract bearing fault characteristics from the data signal after phase synchronization and calculate the phase consistency coefficient.

[0028] Furthermore, the device also includes:

[0029] The data acquisition unit is used to collect vibration signals from the bearing using multiple sensors to obtain detailed bearing operation data signals.

[0030] A storage medium storing a program file capable of implementing any of the above-mentioned fault signal phase synchronization methods based on data fusion algorithms.

[0031] A processor for running a program, wherein the program executes a fault signal phase synchronization method based on a data fusion algorithm, as described above.

[0032] The fault signal phase synchronization method and apparatus based on data fusion algorithm in this invention improves the identifiability of fault signal features through multi-sensor data fusion and advanced phase synchronization algorithm, providing an important technical means for fault diagnosis of rotating machinery. Attached Figure Description

[0033] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0034] Figure 1 This is a flowchart of the fault signal phase synchronization method based on data fusion algorithm of the present invention;

[0035] Figure 2 This is a diagram of the algorithm framework in this invention;

[0036] Figure 3 This is a schematic diagram of phase waveform matching in this invention;

[0037] Figure 4 This is a schematic diagram of multi-channel fusion in the present invention;

[0038] Figure 5 The diagram shows the alignment effect of the algorithm under fault condition 1 in this invention; where (a) is the unaligned signal; (b) is the aligned signal; (c) is the detailed waveform of the unaligned signal; (d) is the detailed waveform of the aligned signal; (e) is the visualization of the alignment effect of the unaligned signal; and (f) is the visualization of the alignment effect of the aligned signal.

[0039] Figure 6 This is a signal correction diagram under fault condition 2 in this invention; where (a) is the uncorrected signal; (b) is the corrected signal; (c) is a detailed comparison of the alignment effect of the uncorrected signal; and (d) is a detailed comparison of the alignment effect of the corrected signal.

[0040] Figure 7 This is a preferred module diagram of the fault signal phase synchronization device based on the data fusion algorithm of the present invention. Detailed Implementation

[0041] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0042] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0043] Example 1

[0044] According to an embodiment of the present invention, a fault signal phase synchronization method based on a data fusion algorithm is provided, see [link to relevant documentation]. Figure 1 This includes the following steps:

[0045] S101: Preprocess the acquired bearing operation data signals;

[0046] S102: Fuse the preprocessed data signals and calculate a reliable phase offset estimate for each operating cycle and position;

[0047] S103: The fused data signal is processed by waveform matching and phase correction algorithms to eliminate the phase offset caused by changes in drive speed and load, and obtain the phase-synchronized data signal.

[0048] S104: Based on the data signal after phase synchronization, extract bearing fault characteristics and calculate the phase consistency coefficient.

[0049] The fault signal phase synchronization method based on data fusion algorithm in this embodiment of the invention improves the identifiability of fault signal features through multi-sensor data fusion and advanced phase synchronization algorithm, providing an important technical means for fault diagnosis of rotating machinery.

[0050] This invention aims to address the phase shift problem in fault signals of rotating structures, which severely hinders data analysis and reduces the training effectiveness of machine learning algorithms. To correct phase shifts caused by external noise, this invention introduces a phase synchronization model based on waveform matching and multi-channel data fusion for synchronously collecting vibration data from bearings. This invention employs a data fusion-based phase synchronization method to improve upon traditional TSA methods, ensuring better alignment of time scales after truncating the original signal.

[0051] The basic content of the technical solution of this invention:

[0052] This invention proposes a fault signal phase synchronization method based on a data fusion algorithm. Through waveform matching and multi-channel data fusion, it achieves high-precision phase synchronization and fault feature extraction of bearing vibration data. The specific steps are as follows:

[0053] (1) Data collection: Multiple sensors are used to collect vibration signals of the bearing in order to obtain detailed operating data.

[0054] (2) Preliminary processing: Super-resolution sampling of the original vibration signal is performed to improve the temporal and spatial resolution of the signal, which facilitates subsequent analysis.

[0055] (3) Construction of Synchronization Information Matrix: Construct a synchronization information matrix (SIM) containing all operation cycles and channel data, and record relative phase offset and waveform similarity data.

[0056] (4) Multi-channel data fusion: The data signals from multiple sensors are fused to calculate the reliable phase offset estimate for each operating cycle and position, forming a synchronization model.

[0057] (5) Phase matching and correction: Through waveform matching and phase correction algorithms, phase offset caused by changes in drive speed and load is eliminated, ensuring the consistency of the signal in all cycles.

[0058] (6) Fault feature extraction: Based on the vibration signal after phase synchronization, the bearing fault features are extracted, and the phase consistency coefficient (PCC) is calculated to evaluate the synchronization effect and ensure high-quality fault feature data.

[0059] The technical solution of this invention improves the identifiability of fault signal characteristics through multi-sensor data fusion and advanced phase synchronization algorithms, providing an important technical means for fault diagnosis of rotating machinery.

[0060] The technical solution of the present invention is described in detail below:

[0061] The technical flowchart of this invention is as follows: Figure 2 As shown, the specific implementation of the present invention includes the following steps:

[0062] 1. Establish a phase waveform matching model for the rotating structure.

[0063] The phase difference between bearing test data from different operating cycles after TSA method truncation, caused by unstable drive speed or equipment load changes, is called Relative Phase Error (RPE). This error is measured by comparing vibration signal data from different rotational revolutions under the same fault condition. Local Phase Offset (LPO) refers to the RPE between different cycles under the same fault condition.

[0064] For a set of discrete signals X = {x i |i=1,2,3,...,N} and Y={y i Let |i=1,2,3,...,N}, X and Y be signals measured by the same vibration measuring equipment at different times. Assuming the waveform matching window length is fixed at l, the set of sample points at phase k within the period is denoted as:

[0065]

[0066] Local offset δ (l,k) Similarity to waveform ρ (l,k) It is given by the following formula:

[0067]

[0068] in:

[0069]

[0070]

[0071] Δ is the set window sliding range, with different scales set according to different bearing conditions, and local offset δ. (l,k) ρ is the maximum covariance between the signal segment captured at a position θ within the window's sliding range and the segment of the same length at the original position of the X signal. (l,k) This is the Pearson correlation coefficient between the two signal segments. Then, the window movement step size d is introduced. s ∈N + The position parameter k represents the number of times the search window moves within one period.

[0072] Figure 3 The diagram illustrates the window search process, where the window moves along the signal time axis at a set step size, progressively calculating the waveform similarity at each window position. Using this method, the present invention can systematically analyze signal characteristics throughout the entire cycle, thereby more accurately identifying periodic and non-periodic changes in the signal.

[0073] 2. Establish a phase synchronization information matrix

[0074] This invention defines a matrix called the Correction Information Matrix (CIM) that stores all bearing operating cycles and channel data under a certain fault condition. The CIM records all intermediate calculation results for n ball bearing operating cycles and t channels, including but not limited to local offsets and waveform similarity. The CIM is a high-dimensional matrix where each element specifically records the relative phase deviation and the acceleration signal of the corresponding channel under a specific cycle and channel.

[0075] Specifically, elements in CIM and This represents the relative phase deviation between the c-channel data and the reference channel at bearing position k in the j-th and i-th cycles. Furthermore, X... i,c The acceleration signal X represents the c-channel. i .

[0076]

[0077]

[0078] Under a specific fault condition at a particular location, this invention defines a local offset matrix and a similarity matrix for each sensor. These two matrices record the phase offset and waveform similarity data of each sensor under the specific fault condition:

[0079]

[0080] These matrices are pieced together to form the Corrected Information Matrix (CIM). The CIM not only includes data collected from each sensor but also systematically integrates information from all operating cycles and channels, forming a comprehensive data view:

[0081]

[0082] U={Δ c (X),Ω c (X)|c=1,2,...,t}

[0083] 3. Establish a multi-channel data fusion model

[0084] The purpose of multi-channel fusion is to estimate the internal errors between channels within each cycle and to fuse the data from different channels to form a phase-synchronized model. This step is crucial for handling anomalies in the signal data, making full use of the ample data samples provided by different sensors to improve the accuracy of phase synchronization.

[0085] The constant offset between the i-th and j-th running cycles in channel t is calculated using the following formula:

[0086]

[0087] The optimal channel offset estimate for the i-th running cycle is calculated and denoted as D. i,c .

[0088]

[0089] Subtracting the previously calculated relative phase deviation, we can obtain the final phase offset.

[0090] The waveform similarity is consistent with the waveform similarity estimated within the channel.

[0091]

[0092] Figure 4 This demonstrates the process of extracting fragments from a high-dimensional correction information matrix and performing data comparison calculations at the channel dimension.

[0093] Based on the data calculated above, this invention attempts to estimate the reliable phase offset of the signal in each cycle. This invention denotes the reliable phase offset estimate of the i-th operating cycle at position k as... This quantity must meet the following two criteria:

[0094] 1) Minimize the squared difference of phase offset between the i-th running cycle and all other running cycles;

[0095] 2) All running cycles at position k The sum is zero;

[0096] These two criteria are expressed by the following system of equations:

[0097]

[0098] The above system of equations can be solved using the augmented Lagrange method:

[0099]

[0100] Then, by taking the partial derivative for each variable, we obtain... Best estimate:

[0101]

[0102] 4. Establish a phase interpolation synchronization model

[0103] This section performs final phase correction on the signal using the relative phase error (RPE) and channel phase error (CPO) calculated above. This process is mainly achieved through two interpolations to ensure phase synchronization of data from different channels and data from different operating cycles.

[0104]

[0105] i∈[1,n];c∈[1,t]

[0106] The first linear interpolation maps new phase coordinates using the deviation estimates obtained from the previous series of calculations. This step represents a linear approximation of the reliable phase estimate, aiming to quickly adjust the original data into a new phase frame. The second conformal cubic interpolation resamples the data to restore the original sampling rate. This higher-order interpolation method further reduces the phase estimation error that may be introduced by the first interpolation, while preserving the shape properties of the data, making the interpolated signal closer to the true physical state.

[0107] To evaluate the synchronization effect, this algorithm attempts to define relevant metrics. Because the TSA method truncates and averages the signal, the final waveform may exhibit amplitude compression distortion under phase shift conditions, thus weakening the signal's characteristic strength. Ideally, this invention aims to maximize the TSA result for each cycle. Furthermore, the signal's ||f||² reflects the intensity of internal noise, which the model needs to correct to be as small as possible. Based on this, this invention defines the ratio of these two as the Phase Consistency Coefficient (PCC) and uses maximizing the PCC as the optimization objective of the phase synchronization algorithm.

[0108]

[0109] The key points and areas to be protected in this invention are:

[0110] (1) By fusing vibration data from multiple sensors, the robustness and reliability of the signal are improved, and high-precision phase synchronization is achieved;

[0111] (2) Waveform matching and phase correction algorithms are used to eliminate phase shifts caused by speed and load changes, ensure signal consistency, and extract accurate fault feature signals;

[0112] (3) Construct a phase synchronization model that includes operation cycle and channel data, and evaluate the synchronization effect by combining the phase consistency coefficient (PCC) to ensure accurate extraction of fault signal characteristics.

[0113] Compared with the prior art, the advantages of the present invention are:

[0114] This invention improves the temporal resolution of vibration signals through multi-sensor data fusion and super-resolution sampling techniques, making fault characteristics clearer and more identifiable. The phase matching and correction algorithm effectively eliminates phase shifts caused by speed and load variations, enhancing signal consistency and stability, and significantly reducing the impact of external noise on signal analysis. Employing an advanced phase synchronization model, by constructing a synchronization information matrix and calculating the phase consistency coefficient (PCC), this invention can extract fault characteristic signals more accurately, providing reliable data support for fault diagnosis of rotating machinery. Furthermore, the method of this invention is not only applicable to bearing fault diagnosis but can also be applied to fault analysis of other rotating machinery, demonstrating broad application prospects and practical value. Compared to existing technologies, this invention offers significant improvements in vibration signal processing and fault feature extraction, providing higher accuracy and reliability in fault diagnosis results.

[0115] This invention has been tested on a public dataset from Western Reserve University. Experimental results show that the method of this invention can effectively extract the characteristic waveforms of bearing faults and performs excellently under high load and unstable speed conditions. Figure 5-6 As shown, through multi-sensor data fusion and super-resolution sampling technology, the temporal and spatial resolution of the vibration signal is significantly improved. The phase matching and correction algorithm effectively eliminates phase shift, improving signal consistency and stability. The application of the phase synchronization model and phase consistency coefficient makes fault feature extraction more accurate.

[0116] Specifically, Figure 5 This section demonstrates the effect of the algorithm on signal phase synchronization correction under fault condition 1 in the dataset. (a), (c), and (e) are unaligned signals, while (b), (d), and (f) are aligned signals. (a) and (b) show the signal superposition effect after the ball rotates along the bearing track for one cycle. A circular plot is used to depict the vibration signal of one cycle to show the continuous periodicity of the signal. (c) and (d) show the correction effect of detailed segments extracted from these segments. (e) and (f) use a two-dimensional matrix color map to visually compare the alignment effect of the phase correction algorithm, thus verifying the effectiveness of the correction. The straighter the longitudinal ripples, the better the alignment effect.

[0117] Figure 6 The image shows the effect of this algorithm on signal phase synchronization correction under fault scenario 2 in the dataset. The inner circle of the left image represents the variance of the signal before and after correction. It can be seen that the variance is reduced sharply after correction, with almost no peak values ​​appearing. The right image shows the weakening of the signal's characteristic intensity by phase shift after TSA processing. The dark curve in the image is the signal waveform after TSA processing.

[0118] Experiments show that using the method of this invention significantly reduces the standard deviation of the fault signal, makes the characteristic waveform clearer, and allows for accurate calculation of the speed fluctuation rate. Overall results demonstrate that the method of this invention is feasible for bearing fault diagnosis, providing higher accuracy and reliability in fault detection and analysis, and offering strong support for the operation and maintenance of mechanical equipment.

[0119] This method is not only applicable to rolling bearings, but can also be applied to fault diagnosis of other rotating machinery such as gearboxes, wind turbines, and pumps. By adjusting the phase synchronization and data fusion algorithms, it can be adapted to the characteristics of different mechanical equipment.

[0120] Integrating the method of this invention into a real-time monitoring system enables online monitoring and fault warning of the operating status of mechanical equipment, timely detection of potential problems, reduction of downtime, and improvement of production efficiency.

[0121] Example 2

[0122] According to another embodiment of the present invention, a fault signal phase synchronization device based on a data fusion algorithm is provided, see [link to relevant documentation]. Figure 7 ,include:

[0123] The preprocessing unit 201 is used to preprocess the acquired bearing operation data signals;

[0124] The data fusion unit 202 is used to fuse the preprocessed data signals and calculate a reliable phase offset estimate for each operation cycle and position.

[0125] The phase synchronization unit 203 is used to process the fused data signal through waveform matching and phase correction algorithms to eliminate the phase offset caused by changes in drive speed and load, and obtain the phase-synchronized data signal.

[0126] The phase consistency unit 204 is used to extract bearing fault characteristics from the data signal after phase synchronization and calculate the phase consistency coefficient.

[0127] The fault signal phase synchronization device based on data fusion algorithm in this embodiment of the invention improves the identifiability of fault signal features through multi-sensor data fusion and advanced phase synchronization algorithm, providing an important technical means for fault diagnosis of rotating machinery.

[0128] This invention aims to address the phase shift problem in fault signals of rotating structures, which severely hinders data analysis and reduces the training effectiveness of machine learning algorithms. To correct phase shifts caused by external noise, this invention introduces a phase synchronization model based on waveform matching and multi-channel data fusion for synchronously collecting vibration data from bearings. This invention employs a data fusion-based phase synchronization method to improve upon traditional TSA methods, ensuring better alignment of time scales after truncating the original signal.

[0129] The basic content of the technical solution of this invention:

[0130] This invention proposes a fault signal phase synchronization device based on a data fusion algorithm. Through waveform matching and multi-channel data fusion, it achieves high-precision phase synchronization and fault feature extraction of bearing vibration data. The specific steps are as follows:

[0131] (1) Data collection: Multiple sensors are used to collect vibration signals of the bearing in order to obtain detailed operating data.

[0132] (2) Preliminary processing: Super-resolution sampling of the original vibration signal is performed to improve the temporal and spatial resolution of the signal, which facilitates subsequent analysis.

[0133] (3) Construction of Synchronization Information Matrix: Construct a synchronization information matrix (SIM) containing all operation cycles and channel data, and record relative phase offset and waveform similarity data.

[0134] (4) Multi-channel data fusion: The data signals from multiple sensors are fused to calculate the reliable phase offset estimate for each operating cycle and position, forming a synchronization model.

[0135] (5) Phase matching and correction: Through waveform matching and phase correction algorithms, phase offset caused by changes in drive speed and load is eliminated, ensuring the consistency of the signal in all cycles.

[0136] (6) Fault feature extraction: Based on the vibration signal after phase synchronization, the bearing fault features are extracted, and the phase consistency coefficient (PCC) is calculated to evaluate the synchronization effect and ensure high-quality fault feature data.

[0137] The technical solution of this invention improves the identifiability of fault signal characteristics through multi-sensor data fusion and advanced phase synchronization algorithms, providing an important technical means for fault diagnosis of rotating machinery.

[0138] The technical solution of the present invention is described in detail below:

[0139] The technical flowchart of this invention is as follows: Figure 2 As shown, the specific implementation of the present invention includes the following steps:

[0140] 1. Establish a phase waveform matching model for the rotating structure.

[0141] The phase difference between bearing test data from different operating cycles after TSA method truncation, caused by unstable drive speed or equipment load changes, is called Relative Phase Error (RPE). This error is measured by comparing vibration signal data from different rotational revolutions under the same fault condition. Local Phase Offset (LPO) refers to the RPE between different cycles under the same fault condition.

[0142] For a set of discrete signals X = {x i |i=1,2,3,...,N} and Y={y i Let |i=1,2,3,...,N}, X and Y be signals measured by the same vibration measuring equipment at different times. Assuming the waveform matching window length is fixed at l, the set of sample points at phase k within the period is denoted as:

[0143]

[0144] Local offset δ (l,k) Similarity to waveform ρ (l,k) It is given by the following formula:

[0145]

[0146] in:

[0147]

[0148]

[0149] Δ is the set window sliding range, with different scales set according to different bearing conditions, and local offset δ. (l,k) ρ is the maximum covariance between the signal segment captured at a position θ within the window's sliding range and the segment of the same length at the original position of the X signal. (l,k) This is the Pearson correlation coefficient between the two signal segments. Then, the window movement step size d is introduced. s ∈N + The position parameter k represents the number of times the search window moves within one period.

[0150] Figure 3 The diagram illustrates the window search process, where the window moves along the signal time axis at a set step size, progressively calculating the waveform similarity at each window position. Using this method, the present invention can systematically analyze signal characteristics throughout the entire cycle, thereby more accurately identifying periodic and non-periodic changes in the signal.

[0151] 2. Establish a phase synchronization information matrix

[0152] This invention defines a matrix called the Correction Information Matrix (CIM) that stores all bearing operating cycles and channel data under a certain fault condition. The CIM records all intermediate calculation results for n ball bearing operating cycles and t channels, including but not limited to local offsets and waveform similarity. The CIM is a high-dimensional matrix where each element specifically records the relative phase deviation and the acceleration signal of the corresponding channel under a specific cycle and channel.

[0153] Specifically, elements in CIM and This represents the relative phase deviation between the c-channel data and the reference channel at bearing position k in the j-th and i-th cycles. Furthermore, X... i,c The acceleration signal X represents the c-channel. i .

[0154]

[0155] Under a specific fault condition at a particular location, this invention defines a local offset matrix and a similarity matrix for each sensor. These two matrices record the phase offset and waveform similarity data of each sensor under the specific fault condition:

[0156]

[0157] These matrices are pieced together to form the Corrected Information Matrix (CIM). The CIM not only includes data collected from each sensor but also systematically integrates information from all operating cycles and channels, forming a comprehensive data view:

[0158]

[0159] U={Δ c (X),Ω c (X)|c=1,2,...,t}

[0160] 3. Establish a multi-channel data fusion model

[0161] The purpose of multi-channel fusion is to estimate the internal errors between channels within each cycle and to fuse the data from different channels to form a phase-synchronized model. This step is crucial for handling anomalies in the signal data, making full use of the ample data samples provided by different sensors to improve the accuracy of phase synchronization.

[0162] The constant offset between the i-th and j-th running cycles in channel t is calculated using the following formula:

[0163]

[0164] The optimal channel offset estimate for the i-th running cycle is calculated and denoted as D. i,c .

[0165]

[0166] Subtracting the previously calculated relative phase deviation, we can obtain the final phase offset.

[0167] The waveform similarity is consistent with the waveform similarity estimated within the channel.

[0168]

[0169] Figure 4 This demonstrates the process of extracting fragments from a high-dimensional correction information matrix and performing data comparison calculations at the channel dimension.

[0170] Based on the data calculated above, this invention attempts to estimate the reliable phase offset of the signal in each cycle. This invention denotes the reliable phase offset estimate of the i-th operating cycle at position k as... This quantity must meet the following two criteria:

[0171] 1) Minimize the squared difference of phase offset between the i-th running cycle and all other running cycles;

[0172] 2) All running cycles at position k The sum is zero;

[0173] These two criteria are expressed by the following system of equations:

[0174]

[0175] The above system of equations can be solved using the augmented Lagrange method:

[0176]

[0177] Then, by taking the partial derivative for each variable, we obtain... Best estimate:

[0178]

[0179] 4. Establish a phase interpolation synchronization model

[0180] This section performs final phase correction on the signal using the relative phase error (RPE) and channel phase error (CPO) calculated above. This process is mainly achieved through two interpolations to ensure phase synchronization of data from different channels and data from different operating cycles.

[0181]

[0182] i∈[1,n];c∈[1,t]

[0183] The first linear interpolation maps new phase coordinates using the deviation estimates obtained from the previous series of calculations. This step represents a linear approximation of the reliable phase estimate, aiming to quickly adjust the original data into a new phase frame. The second conformal cubic interpolation resamples the data to restore the original sampling rate. This higher-order interpolation method further reduces the phase estimation error that may be introduced by the first interpolation, while preserving the shape properties of the data, making the interpolated signal closer to the true physical state.

[0184] To evaluate the synchronization effect, this algorithm attempts to define relevant metrics. Because the TSA method truncates and averages the signal, the final waveform may exhibit amplitude compression distortion under phase shift conditions, thus weakening the signal's characteristic strength. Ideally, this invention aims to maximize the TSA result for each cycle. Furthermore, the signal's ||f||² reflects the intensity of internal noise, which the model needs to correct to be as small as possible. Based on this, this invention defines the ratio of these two as the Phase Consistency Coefficient (PCC) and uses maximizing the PCC as the optimization objective of the phase synchronization algorithm.

[0185]

[0186] The key points and areas to be protected in this invention are:

[0187] (1) By fusing vibration data from multiple sensors, the robustness and reliability of the signal are improved, and high-precision phase synchronization is achieved;

[0188] (2) Waveform matching and phase correction algorithms are used to eliminate phase shifts caused by speed and load changes, ensure signal consistency, and extract accurate fault feature signals;

[0189] (3) Construct a phase synchronization model that includes operation cycle and channel data, and evaluate the synchronization effect by combining the phase consistency coefficient (PCC) to ensure accurate extraction of fault signal characteristics.

[0190] Compared with the prior art, the advantages of the present invention are:

[0191] This invention improves the temporal resolution of vibration signals through multi-sensor data fusion and super-resolution sampling techniques, making fault characteristics clearer and more identifiable. The phase matching and correction algorithm effectively eliminates phase shifts caused by speed and load variations, enhancing signal consistency and stability, and significantly reducing the impact of external noise on signal analysis. Employing an advanced phase synchronization model, by constructing a synchronization information matrix and calculating the phase consistency coefficient (PCC), this invention can extract fault characteristic signals more accurately, providing reliable data support for fault diagnosis of rotating machinery. Furthermore, the method of this invention is not only applicable to bearing fault diagnosis but can also be applied to fault analysis of other rotating machinery, demonstrating broad application prospects and practical value. Compared to existing technologies, this invention offers significant improvements in vibration signal processing and fault feature extraction, providing higher accuracy and reliability in fault diagnosis results.

[0192] This invention has been tested on a public dataset from Western Reserve University. Experimental results show that the method of this invention can effectively extract the characteristic waveforms of bearing faults and performs excellently under high load and unstable speed conditions. Figure 5-6 As shown, through multi-sensor data fusion and super-resolution sampling technology, the temporal and spatial resolution of the vibration signal is significantly improved. The phase matching and correction algorithm effectively eliminates phase shift, improving signal consistency and stability. The application of the phase synchronization model and phase consistency coefficient makes fault feature extraction more accurate.

[0193] Specifically, Figure 5 This section demonstrates the effect of the algorithm on signal phase synchronization correction under fault condition 1 in the dataset. (a), (c), and (e) are unaligned signals, while (b), (d), and (f) are aligned signals. (a) and (b) show the signal superposition effect after the ball rotates along the bearing track for one cycle. A circular plot is used to depict the vibration signal of one cycle to show the continuous periodicity of the signal. (c) and (d) show the correction effect of detailed segments extracted from these segments. (e) and (f) use a two-dimensional matrix color map to visually compare the alignment effect of the phase correction algorithm, thus verifying the effectiveness of the correction. The straighter the longitudinal ripples, the better the alignment effect.

[0194] Figure 6 The image shows the effect of this algorithm on signal phase synchronization correction under fault scenario 2 in the dataset. The inner circle of the left image represents the variance of the signal before and after correction. It can be seen that the variance is reduced sharply after correction, with almost no peak values ​​appearing. The right image shows the weakening of the signal's characteristic intensity by phase shift after TSA processing. The dark curve in the image is the signal waveform after TSA processing.

[0195] Experiments show that using the method of this invention significantly reduces the standard deviation of the fault signal, makes the characteristic waveform clearer, and allows for accurate calculation of the speed fluctuation rate. Overall results demonstrate that the method of this invention is feasible for bearing fault diagnosis, providing higher accuracy and reliability in fault detection and analysis, and offering strong support for the operation and maintenance of mechanical equipment.

[0196] This method is not only applicable to rolling bearings, but can also be applied to fault diagnosis of other rotating machinery such as gearboxes, wind turbines, and pumps. By adjusting the phase synchronization and data fusion algorithms, it can be adapted to the characteristics of different mechanical equipment.

[0197] Integrating the method of this invention into a real-time monitoring system enables online monitoring and fault warning of the operating status of mechanical equipment, timely detection of potential problems, reduction of downtime, and improvement of production efficiency.

[0198] Example 3

[0199] A storage medium storing a program file capable of implementing any of the above-mentioned fault signal phase synchronization methods based on data fusion algorithms.

[0200] Example 4

[0201] A processor for running a program, wherein the program executes a fault signal phase synchronization method based on a data fusion algorithm, as described above.

[0202] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0203] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0204] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The system embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of units or modules may be electrical or other forms.

[0205] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0206] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0207] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0208] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A fault signal phase synchronization method based on a data fusion algorithm, characterized in that, Includes the following steps: S101: Preprocess the acquired bearing operation data signals; S102: Fuse the preprocessed data signals and calculate a reliable phase offset estimate for each operating cycle and position; S103: The fused data signal is processed by waveform matching and phase correction algorithms to eliminate the phase offset caused by changes in drive speed and load, and obtain the phase-synchronized data signal. S104: Based on the data signal after phase synchronization, extract bearing fault characteristics and calculate the phase consistency coefficient.

2. The fault signal phase synchronization method based on data fusion algorithm according to claim 1, characterized in that, The method further includes the following step before step S101: Multiple sensors are used to collect vibration signals from the bearing to obtain detailed bearing operation data.

3. The fault signal phase synchronization method based on data fusion algorithm according to claim 1, characterized in that, Step S101 includes: performing super-resolution sampling on the acquired raw vibration signal of the bearing operation to improve the temporal and spatial resolution of the signal.

4. The fault signal phase synchronization method based on data fusion algorithm according to claim 1, characterized in that, Step S102 includes: Construct a synchronization information matrix containing all operation cycles and channel data, and record relative phase offset and waveform similarity data; Data signals from multiple sensors are fused to calculate a reliable phase offset estimate for each operating cycle and position, forming a synchronization model.

5. The fault signal phase synchronization method based on data fusion algorithm according to claim 4, characterized in that, The window moves along the signal time axis in a set step size, and the waveform similarity at each window position is calculated step by step.

6. The fault signal phase synchronization method based on data fusion algorithm according to claim 4, characterized in that, The synchronization information matrix is ​​a high-dimensional matrix, where each element specifically records the relative phase deviation and the acceleration signal of the corresponding channel under a specific period and channel.

7. The fault signal phase synchronization method based on data fusion algorithm according to claim 1, characterized in that, In step S102, the internal error between channels within each cycle is estimated, and the data from different channels are fused to form a phase synchronization model.

8. The fault signal phase synchronization method based on data fusion algorithm according to claim 1, characterized in that, In step S103, the calculated relative phase error and channel phase error are used to perform the final phase correction on the signal. Two interpolations are used to ensure phase synchronization of data from different channels and phase synchronization of data from different operating cycles.

9. A fault signal phase synchronization device based on a data fusion algorithm, characterized in that, include: The preprocessing unit is used to preprocess the acquired bearing operation data signals; The data fusion unit is used to fuse the preprocessed data signals and calculate a reliable phase offset estimate for each operating cycle and position. The phase synchronization unit is used to process the fused data signal through waveform matching and phase correction algorithms to eliminate the phase offset caused by changes in drive speed and load, and obtain the phase-synchronized data signal. The phase consistency unit is used to extract bearing fault characteristics from the data signal after phase synchronization and calculate the phase consistency coefficient.

10. The fault signal phase synchronization device based on data fusion algorithm according to claim 9, characterized in that, The device further includes: The data acquisition unit is used to collect vibration signals from the bearing using multiple sensors to obtain detailed bearing operation data signals.

Images