Bearing fault identification method based on signal evaluation index and multi-sensor fusion

Abstract

The invention provides a bearing fault identification method based on signal evaluation indexes and multi-sensor fusion. The method comprises the following steps: collecting vibration signals of N sensors at the same moment as input; the collected signals are optimized through the optimal equalization factor; performing weighted fusion on the optimized signal by using a weight calculation method to obtain a fused signal; decomposing the fusion signal to obtain a corresponding component signal; selecting a signal containing the most abundant fault information from the obtained component signals, and recording the signal as y (t); and calculating an autocorrelation function yc (t) of the y (t), carrying out spectral analysis on the yc (t), extracting the fault characteristic frequency of the bearing, and carrying out bearing fault identification. According to the method, the weight coefficient is adaptively determined by calculating the evaluation index values of the input data measured by different sensors; and combining the fused signal with a signal decomposition algorithm for subsequent processing. Effective fusion of bearing fault feature information and enhancement of fault features are realized, and the fault type of the bearing is accurately judged.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and in particular to a bearing fault identification method based on signal evaluation index and multi-sensor fusion. Background Art

[0002] As one of the most widely used and numerous parts, rolling bearings are widely used in mechanical equipment in many fields such as electricity, metallurgy, and rail transportation. Its working status is directly related to key factors such as the equipment's processing accuracy, production efficiency, and operational safety. Data show that more than 44% of rotating mechanical equipment failures are related to rolling bearing problems. Due to the harsh working environment and heavy loads, rolling bearings are prone to compound failures, which are far more harmful than single failures. Therefore, the effective extraction of rolling bearing compound failure characteristics has become a major challenge in the field of rotating machinery fault diagnosis. In-depth research on this issue is of great significance for improving the reliability of equipment operation and ensuring personnel safety. When a rolling bearing fails, the collision between the damaged area and other components will generate an impact signal, and different fault types correspond to different characteristic frequencies. By identifying these characteristic frequencies, the fault type of the bearing can be accurately determined.

[0003] In the field of bearing fault diagnosis, the vibration signals collected by sensors are an important basis for evaluating the health status of bearings. These vibration signals contain rich information about the bearing status. How to accurately extract features from these signals is a key link in identifying bearing fault status. In the field of bearing fault feature extraction, signal processing technology has gradually shifted from traditional Fourier transform and short-time Fourier transform to advanced time-frequency analysis technologies, such as wavelet transform and variational mode decomposition (VMD). These technologies can more effectively capture the non-stationary characteristics and local features of signals, thereby improving the recognition accuracy of fault features. In addition, adaptive filtering and autocorrelation noise reduction techniques are also applied to signal preprocessing to eliminate noise and interference and enhance the clarity of fault features. The continuous progress and optimization of these signal processing technologies have laid a solid technical foundation for the accurate extraction of bearing fault features.

[0004] At present, most fault diagnosis methods based on vibration signal analysis focus on the research of single sensor signals. However, single sensor data analysis has inherent limitations. Its diagnostic results are often limited by factors such as sensor type, installation location, and environmental noise, resulting in incomplete information, which will affect the effect of fault identification. In order to overcome these limitations, multi-sensor data fusion technology has emerged. Multi-sensor fusion can make up for the shortcomings of a single sensor in information acquisition. By integrating the data of multiple sensors, it not only expands the dimension and scope of monitoring, but also improves the accuracy and reliability of fault diagnosis through the complementarity and verification between data. It plays an increasingly important role in bearing fault diagnosis. The weighted method in the multi-sensor data fusion algorithm is suitable for data-level fusion. Since there is a certain similarity between the data collected by different sensors at the same time on the same section, and the noise is random. Therefore, the weighted fusion algorithm is used to assign different weights to each signal, which will further highlight the features and reduce the noise. The most important issue in the weighted fusion algorithm is how to determine the weight value of each signal. The reasonable allocation of weight coefficients can ensure that the fusion result is closer to the actual situation. In a multi-sensor system, different sensors may have different measurement accuracy, sensitivity, and reliability. By giving higher weights to sensors with high precision and high reliability, the fusion results can be made more dependent on these high-quality data, thereby improving the overall accuracy. Summary of the invention

[0005] According to the technical problems raised above, a bearing fault identification method based on signal evaluation index and multi-sensor fusion is provided. In view of the above problems, the present invention adaptively determines the weight coefficient by calculating the evaluation index value of the input data measured by different sensors; and combines the fused signal with the signal decomposition algorithm for subsequent processing; realizes the effective fusion of bearing fault feature information and the enhancement of fault features, accurately determines the composite fault type of the bearing, and has engineering application value.

[0006] The technical means adopted by the present invention are as follows:

[0007] A bearing fault identification method based on signal evaluation index and multi-sensor fusion, characterized by comprising:

[0008] S1, collect the vibration signals of N sensors at the same time as input;

[0009] S2, optimizing the signal collected in S1 through the optimal equalization factor ε;

[0010] S3, using a weight calculation method to perform weighted fusion on the optimized signal in S2 to obtain a fused signal;

[0011] S4, decomposing the fused signal in S3 using a signal decomposition algorithm to obtain k decomposed component signals;

[0012] S5. Among the k component signals obtained in S4, select the signal with the richest fault information and record it as y(t); calculate the autocorrelation function y of y(t) c (t), and for y c (t) Perform spectrum analysis to extract the fault characteristic frequency of the bearing and identify the bearing fault.

[0013] Further, the vibration signal is represented as X n (i), where n represents the nth sensor, n = 1, 2, 3 ... N, N is the number of sensors, and i = 1, 2, 3 ... L, L is the discrete signal length.

[0014] Furthermore, the step S2 specifically includes:

[0015] All vibration signals X collected in S1 n (i) is used as input and the optimal equalization factor ε is used for data optimization. The optimization process is as follows:

[0016]

[0017] in, is the optimized signal.

[0018] Furthermore, the step S3 specifically includes:

[0019] S31, normalize the signal optimized in S2:

[0020]

[0021] Where x is A value in the data set, min(x) is the minimum value in the data set, max(x) is the maximum value in the data set, and x′ is the normalized data;

[0022] According to the normalized data, construct the signal evaluation index vector set A, A = {A1, A2, A3, A4, ..., A K}, where A K represents the Kth signal evaluation index. For the signal evaluation index vector set A, the values ​​of each signal evaluation index corresponding to the data of each sensor are calculated respectively. The signal evaluation index value corresponding to the data of the nth sensor is expressed as m. nA1 、m nA2 、m nA3 ,......,m nAK Indicates that the values ​​of each signal evaluation index of all sensor corresponding data are represented by matrix B:

[0023]

[0024] The index value is processed as non-negative, and the jth evaluation index corresponding to the nth sensor is The non-negative processing process is:

[0025]

[0026] Where, j = 1, 2, 3, 4, ... K;

[0027] Each index value is processed as non-negative to obtain a new index set Sq:

[0028]

[0029] S32, calculate the weight of each signal evaluation index corresponding to the data of each sensor in the index set Sq after non-negative processing, and the jth evaluation index corresponding to the nth sensor The weight of the evaluation index of the j-th signal for:

[0030]

[0031] The average value of the K weights corresponding to the same sensor data is obtained. That is the final weight corresponding to the nth sensor data:

[0032]

[0033] The optimized signals corresponding to N sensors are weighted fused according to the obtained weight coefficients, and the fused signal is obtained as

[0034]

[0035] in, The fusion signal.

[0036] Furthermore, the step S4 specifically includes:

[0037] S41, use the robust mean decomposition algorithm to fusion signal in S3 Perform signal decomposition;

[0038] Calculate the fusion signal All local extreme points of i , find two adjacent points e i and e i+1 The average value m i :

[0039]

[0040] Calculate the envelope estimate a based on the extreme points i :

[0041]

[0042] The moving average method is used to process the local mean function m 11 (t) and envelope estimation function a 11 (t), calculate the demodulated signal:

[0043]

[0044] Among them, s 11 (t) is the demodulated signal;

[0045] S42, if s 11 The envelope estimation function of (t) satisfies a 12 (t)=1, then s 11 (t) is taken as the original signal, and S41 is repeated to iterate until s 1n (t), until the final signal satisfies -1≤s 1n (t)≤1, envelope estimation function a 1(n+1) (t) satisfies a 1(n+1) (t) = 1;

[0046] Multiply all the envelope estimation functions generated during the iteration process to obtain the envelope signal:

[0047]

[0048] The envelope signal a1(t) and the frequency modulated signal s 1n (t) and the first PF component can be obtained by multiplying:

[0049] PF1(t)=a1(y)s 1n (t)

[0050] After decomposition, we can get k component signals y1(t), y2(t),…y k (t).

[0051] Furthermore, the step S5 specifically includes:

[0052] Calculate the autocorrelation function y of the signal y(t) selected in S4 that contains the most fault information c (t), and for y c (t) Spectral analysis of the signal is performed to extract the fault characteristic frequency of the bearing and realize bearing fault identification;

[0053] Bearing outer ring fault characteristic frequency fo for:

[0054]

[0055] Bearing inner ring fault characteristic frequency f i for:

[0056]

[0057] Bearing rolling element fault characteristic frequency f b for:

[0058]

[0059] Bearing cage failure frequency f c for:

[0060]

[0061] Among them, Z is the number of rolling elements, D is the bearing pitch diameter, d is the ball diameter, and f r is the rotation frequency.

[0062] Compared with the prior art, the present invention has the following advantages:

[0063] The bearing fault identification method based on signal evaluation index and multi-sensor fusion provided by the present invention introduces an optimal equalization factor to optimize the collected data to reduce the observation error of each sensor; the weight coefficient during multi-sensor fusion is adaptively determined by a signal evaluation index vector set; the fused data is combined with a signal decomposition algorithm, and the bearing fault characteristics are extracted by spectral analysis of the component signals; the effective monitoring of the bearing operating status and the accurate identification of the bearing fault type are achieved, and the method has engineering application value.

[0064] Based on the above reasons, the present invention can be widely promoted in the field of fault diagnosis technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0066] Figure 1 This is a block diagram of the bearing fault identification method based on signal evaluation index and multi-sensor fusion of the present invention.

[0067] Figure 2 The figure is a block diagram of the multi-sensor fusion method in an embodiment of the present invention.

[0068] Figure 3 This is a graph showing the composite fault identification results of a rolling bearing in an embodiment of the present invention. DETAILED DESCRIPTION

[0069] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0070] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0071] like Figure 1 As shown, the present invention provides a bearing fault identification method based on signal evaluation index and multi-sensor fusion, comprising:

[0072] S1, collect the vibration signals of N sensors at the same time as input;

[0073] In specific implementation, as a preferred embodiment of the present invention, the vibration signal is represented as X n (i), where n represents the nth sensor, n = 1, 2, 3 ... N, N is the number of sensors, and i = 1, 2, 3 ... L, L is the discrete signal length.

[0074] S2, optimizing the signal collected in S1 through the optimal equalization factor ε;

[0075] In specific implementation, as a preferred embodiment of the present invention, step S2 specifically includes:

[0076] All vibration signals X collected in S1 n (i) is used as input and the optimal equalization factor ε is used for data optimization. The optimization process is as follows:

[0077]

[0078] in, is the optimized signal.

[0079] S3, using the weight calculation method to perform weighted fusion on the optimized signal in S2 to obtain a fused signal; Figure 2 shown.

[0080] In specific implementation, as a preferred embodiment of the present invention, step S3 specifically includes:

[0081] S31, normalize the signal optimized in S2:

[0082]

[0083] Where x is A value in a data set, min(x) is the minimum value in the data set, max(x) is the maximum value in the data set, and x′ is the normalized value.

[0084] According to the normalized data, construct the signal evaluation index vector set A, A = {A1, A2, A3, A4, ..., A K}, where A K Represents the Kth signal evaluation index. For the signal evaluation index vector set A, the values ​​of each signal evaluation index corresponding to the data of each sensor are calculated respectively. The signal evaluation index value corresponding to the data of the nth sensor is expressed as m. nA1 、m nA2 、m nA3 ,......,m nAK Indicates that the values ​​of each signal evaluation index of all sensor corresponding data are represented by matrix B:

[0085]

[0086] The index value is processed as non-negative, and the jth evaluation index corresponding to the nth sensor is The non-negative processing process is:

[0087]

[0088] Where, j = 1, 2, 3, 4, ... K;

[0089] Each index value is processed as non-negative to obtain a new index set Sq:

[0090]

[0091] S32, calculate the weight of each signal evaluation index corresponding to the data of each sensor in the index set Sq after non-negative processing, which accounts for the weight of the signal evaluation index of this type. The weight of the evaluation index of the j-th signal Take this as an example to explain the calculation process:

[0092]

[0093] The average value of the K weights corresponding to the same sensor data is obtained. That is the final weight corresponding to the nth sensor data:

[0094]

[0095] The optimized signals corresponding to the N sensors (obtained in S2) are weighted fused according to the weight coefficients obtained in the previous step, and the fused signal is obtained as

[0096]

[0097] in, The fusion signal.

[0098] S4, decompose the fusion signal in S3 using the signal decomposition algorithm to obtain k decomposed component signals y1(t), y2(t), ...y k (t);

[0099] In specific implementation, as a preferred embodiment of the present invention, step S4 specifically includes:

[0100] S41, fusion signal in S3 Perform robust local mean decomposition and calculate the fusion signal All local extreme points of i , find two adjacent points e i and e i+1 The average value m i :

[0101]

[0102] Calculate the envelope estimate a based on the extreme points i :

[0103]

[0104] The moving average method is used to process the local mean function m 11 (t) and envelope estimation function a 11 (t), calculate the demodulated signal:

[0105]

[0106] Among them, s 11 (t) is the demodulated signal;

[0107] S42, if s 11 The envelope estimation function of (t) satisfies a 12 (t)=1, then s 11 (t) is taken as the original signal, and S41 is repeated to iterate until s 1n (t), until the final signal satisfies -1≤s 1n (t)≤1, envelope estimation function a 1(n+1) (t) satisfies a 1(n+1) (t) = 1;

[0108] Multiply all the envelope estimation functions generated during the iteration process to obtain the envelope signal:

[0109]

[0110] The envelope signal a1(t) and the frequency modulated signal S 1n (t) and the first PF component can be obtained by multiplying:

[0111] PF1(t)=a1(t)S 1n (t)

[0112] In implementation, the component is separated from the original signal to obtain u1(t)=x(t)-PF1(t), which is used as the original signal to repeat the above process for the kth time until the final u k It is a monotonic function, and finally multiple PF components can be obtained.

[0113] Robust local mean decomposition optimizes the local mean decomposition method from three aspects, namely, the optimization of boundary conditions, the optimization of boundary conditions, and the optimization of boundary conditions. After the decomposition algorithm is optimized, it is decomposed to obtain k decomposed component signals y1(t), y2(t),…y k (t);

[0114] S5. Among the k component signals obtained in S4, select the signal with the richest fault information and record it as y(t); calculate the autocorrelation function y of y(t) c (t), and for y c (t) Perform spectrum analysis to extract the fault characteristic frequency of the bearing and identify the bearing fault.

[0115] In specific implementation, as a preferred embodiment of the present invention, step S5 specifically includes:

[0116] Calculate the autocorrelation function y of the signal y(t) selected in S4 that contains the most fault information c (t), and for y c (t) Spectral analysis of the signal is performed to extract the fault characteristic frequency of the bearing and realize bearing fault identification;

[0117] Bearing outer ring fault characteristic frequency f o for:

[0118]

[0119] Bearing inner ring fault characteristic frequency f i for:

[0120]

[0121] Bearing rolling element fault characteristic frequency f b for:

[0122]

[0123] Bearing cage failure frequency f c for:

[0124]

[0125] Among them, Z is the number of rolling elements, D is the bearing pitch diameter, d is the ball diameter, and f r is the rotation frequency.

[0126] Example

[0127] like Figure 1 As shown, the present invention provides a bearing fault identification method based on signal evaluation index and multi-sensor fusion. In this embodiment, when a rolling bearing has a composite fault of the outer ring and the inner ring, typical vibration data is randomly selected for analysis and verification (taking the number of sensors as 3, the number of evaluation indexes as 4, ε=0.55, the number of rolling elements Z=7, the bearing pitch diameter D=36, and the ball diameter d=9.6 as an example), and the kurtosis index for measuring the impact characteristics is applied when calculating the weight; the average value index for measuring the overall level of the signal; the multi-scale diversity entropy index with high consistency and high calculation efficiency; and the square envelope Gini index index that is very sensitive to the nonlinear characteristics and fault characteristics in the signal when the bearing fails. Where A is the kurtosis, B is the square envelope Gini index, C is the multi-scale diversity entropy, and E is the average value.

[0128] Table 1 Parameters represented by each symbol and calculation formula of each characteristic frequency of intermediate bearing

[0129]

[0130]

[0131] According to Table 1 and the bearing geometry, the characteristic frequencies calculated are: f r =25.44Hz, f o =65.3Hz, f i =112.8Hz, f c =9.33Hz. The signal of rolling bearing compound fault is analyzed according to the method of the present invention. The results are as follows Figure 3 shown. Figure 3 (a1)-(a2) are the time domain and spectrum diagrams of the composite fault signal respectively. The weights of the three signals calculated by the fusion method proposed in this paper are 0.2381, 0.3954, and 0.3666. The time domain and spectrum diagrams of the signals obtained by fusing the vibration signals of the three sensors are shown in Figure 3 (a3)-(a4) are shown. The time domain image obtained by performing robust local mean decomposition on the fused signal is shown in Figure 3 As shown in (a5), among the component signals obtained by robust local mean decomposition, the first layer of component signals can usually provide the basic characteristics of the signal and have the richest information, so the first layer of component signals is selected for analysis. Figure 3 (a6) is the spectrum of the first layer component signal. Figure 3 (a7)-(a8) is Figure 3 (a6) Autocorrelation function and its normalized spectrum. Figure 3 The corresponding relationship between the typical frequency components and the characteristic frequencies in (a8) is shown in Table 2.

[0132] Table 2 Figure 3 (a8) Correspondence table between typical frequencies and characteristic frequencies

[0133] Typical frequency components Characteristic frequency Fault type 294.2 (294.2+fr+2fc) / 3=112.74 Inner race fault 698.2 (698.2-fr) / 6＝112.12 Inner race fault 1311 1311 / 20＝65.55 Outer ring failure

[0134] analyze Figure 3 (a8) and Table 2, it can be found that the spectrum of the signal obtained by the proposed method has the following characteristics:

[0135] (1) The noise component is greatly reduced, achieving effective noise suppression.

[0136] (2) There is a typical and prominent frequency component of 1311 Hz, which corresponds to the characteristic frequency of the bearing outer ring fault. The calculation process is shown in Table 2, marked as ★ in the figure.

[0137] (3) There are typical and prominent frequency components of 294.2 Hz and 698.2 Hz. These two frequencies correspond to the characteristic frequencies of the inner ring fault of the bearing. The calculation process is shown in Table 2, and the figure is marked as .

[0138] These prominent frequency components are completely consistent with the composite fault type of the bearing, which proves that the method proposed in the present invention can effectively fuse the information of multi-sensor data and accurately judge and identify the composite fault type of the bearing.

[0139] It can be seen that: according to the bearing composite fault identification method based on signal evaluation index and multi-sensor fusion proposed by the present invention, in the spectrum analysis of the obtained signal, the noise component is effectively suppressed, and the frequency component corresponding to the fault characteristic frequency is prominent, which realizes the accurate identification of the bearing composite fault type. It has excellent engineering application value.

[0140] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0141] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0142] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0143] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0144] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0145] If the integrated unit is implemented in the form of 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, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0146] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A bearing fault identification method based on signal evaluation index and multi-sensor fusion, characterized in that: include: S1, collect the vibration signals of N sensors at the same time as input; S2, optimizing the signal collected in S1 through the optimal equalization factor ε; S3, using a weight calculation method to perform weighted fusion on the optimized signal in S2 to obtain a fused signal; S4, decomposing the fused signal in S3 using a signal decomposition algorithm to obtain k decomposed component signals; S5. Among the k component signals obtained in S4, select the signal with the richest fault information and record it as y(t); calculate the autocorrelation function y of y(t) c (t), and for y c (t) Perform spectrum analysis to extract the fault characteristic frequency of the bearing and identify the bearing fault.

2. The bearing fault identification method based on signal evaluation index and multi-sensor fusion according to claim 1 is characterized in that: The vibration signal is represented as X n (i), where n represents the nth sensor, n = 1, 2, 3 ... N, N is the number of sensors, and i = 1, 2, 3 ... L, L is the discrete signal length.

3. The bearing fault identification method based on signal evaluation index and multi-sensor fusion according to claim 1 is characterized in that: The step S2 specifically includes: All vibration signals X collected in S1 n (i) is used as input and the optimal equalization factor ε is used for data optimization. The optimization process is as follows: in, is the optimized signal.

4. The bearing fault identification method based on signal evaluation index and multi-sensor fusion according to claim 1 is characterized in that: The step S3 specifically includes: S31, normalize the signal optimized in S2: Where x is A value in the data set, min(x) is the minimum value in the data set, max(x) is the maximum value in the data set, and x′ is the normalized data; According to the normalized data, construct the signal evaluation index vector set A, A = {A1, A2, A3, A4, ..., A K }, where A K represents the Kth signal evaluation index. For the signal evaluation index vector set A, the values ​​of each signal evaluation index corresponding to the data of each sensor are calculated respectively. The signal evaluation index value corresponding to the data of the nth sensor is expressed as m. nA1 、m nA2 、m nA3 ,......,m nAK Indicates that the values ​​of each signal evaluation index of all sensor corresponding data are represented by matrix B: The index value is processed as non-negative, and the jth evaluation index corresponding to the nth sensor is The non-negative processing process is: Where, j = 1, 2, 3, 4, ... K; Each index value is processed as non-negative to obtain a new index set Sq: S32, calculate the weight of each signal evaluation index corresponding to the data of each sensor in the index set Sq after non-negative processing, and the jth evaluation index corresponding to the nth sensor The weight of the evaluation index of the j-th signal for: The average value of the K weights corresponding to the same sensor data is obtained. That is the final weight corresponding to the nth sensor data: The optimized signals corresponding to N sensors are weighted fused according to the obtained weight coefficients, and the fused signal is obtained as in, The fusion signal.

5. The bearing fault identification method based on signal evaluation index and multi-sensor fusion according to claim 1 is characterized in that: The step S4 specifically includes: S41, use the robust mean decomposition algorithm to fusion signal in S3 Perform signal decomposition; Calculate the fusion signal All local extreme points of i , find two adjacent points e i and e i+1 The average value m i : Calculate the envelope estimate a based on the extreme points i : The moving average method is used to process the local mean function m 11 (t) and envelope estimation function a 11 (t), calculate the demodulated signal: Among them, s 11 (t) is the demodulated signal; S42, if s 11 The envelope estimation function of (t) satisfies a 12 (t)=1, then s 11 (t) is taken as the original signal, and S41 is repeated to iterate until s 1n (t), until the final signal satisfies -1≤s 1n (t)≤1, envelope estimation function a 1(n+1) (t) satisfies a 1(n+1) (t) = 1; Multiply all the envelope estimation functions generated during the iteration process to obtain the envelope signal: The envelope signal a1(t) and the frequency modulated signal s 1n (t) and the first PF component can be obtained by multiplying: PF1(t)=a1(t)s 1n (t) After decomposition, we can get k component signals y1(t), y2(t),…y k (t).

6. The bearing fault identification method based on signal evaluation index and multi-sensor fusion according to claim 1 is characterized in that: The step S5 specifically includes: Calculate the autocorrelation function y of the signal y(t) selected in S4 that contains the most fault information c (t), and for y c (t) Spectral analysis of the signal is performed to extract the fault characteristic frequency of the bearing and realize bearing fault identification; Bearing outer ring fault characteristic frequency f o for: Bearing inner ring fault characteristic frequency f i for: Bearing rolling element fault characteristic frequency f b for: Bearing cage failure frequency f c for: Among them, Z is the number of rolling elements, D is the bearing pitch diameter, d is the ball diameter, and f r is the rotation frequency.

Images