A method for bearing wear resistance detection

Abstract

The present application belongs to the technical field of bearing wear resistance detection, and particularly relates to a method for bearing wear resistance detection. The method comprises the steps of data acquisition, primary detection, secondary detection and comprehensive evaluation. The data acquisition acquires real-time temperature and sound data, including environmental temperature and background noise interference. The primary detection performs difference calculation, Kalman filtering and compensation correction on the temperature data, and performs time-frequency spectrum analysis and frequency band noise reduction on the sound data. The primary wear index is calculated by fusing the two, and if the primary wear index is less than a first threshold value, it is determined to be qualified, otherwise, it enters the secondary detection. The secondary detection collects the rolling body, inner ring and outer ring slide wear images, and calculates the secondary wear index through a double-branch convolution network after preprocessing and comparing with the non-wear image. The comprehensive wear degree is obtained by weighted fusion of the two-stage indexes, and if the comprehensive wear degree is greater than a second threshold value, it is determined to be unqualified and an alarm is given, otherwise, it is qualified. The method dynamically eliminates environmental interference, combines multiple parameters and image features, and improves the detection accuracy and reliability.

Description

Technical Field

[0001] This invention belongs to the field of bearing wear resistance testing technology, and particularly relates to a method for testing bearing wear resistance. Background Technology

[0002] Bearings are essential components in modern machinery. Their primary function is to support rotating mechanical parts, reduce the coefficient of friction during movement, and ensure rotational accuracy. The wear resistance of bearings directly affects the operational stability, service life, and safety of equipment; therefore, accurate wear resistance testing is crucial for ensuring reliable equipment operation. Current technologies often neglect the dynamic characteristics of environmental interference and fail to incorporate image features into wear resistance testing, resulting in often unsatisfactory test results. Summary of the Invention

[0003] In view of the technical problems existing in the background art, the present invention proposes a method for testing the wear resistance of bearings.

[0004] To achieve the above objectives, the technical solution adopted by the present invention includes the following steps:

[0005] Data acquisition: Real-time data during bearing operation is acquired, including temperature data and sound data. The data contains interference parameters, including ambient temperature and background noise.

[0006] Primary inspection: The collected temperature and sound data are preprocessed, and the temperature and sound change data during the current inspection period are processed to calculate the primary wear index. If the primary wear index is less than the first threshold, the bearing is judged to be qualified in terms of wear resistance. If it is greater than or equal to the first threshold, secondary inspection is performed.

[0007] Secondary detection: Acquire bearing surface wear images and preprocess the images; the bearing surface wear images include rolling element images, inner ring slide images and outer ring slide images, and the secondary wear index is obtained by differentiating the bearing surface wear images with preset no-wear images;

[0008] Comprehensive assessment: The comprehensive wear degree is determined by weighted calculation based on the primary wear index and the secondary wear index. If the comprehensive wear degree is greater than the second threshold, the bearing wear resistance is deemed unqualified and an alarm is triggered; otherwise, it is deemed qualified and the test results are recorded.

[0009] Preferably, the specific implementation of preprocessing the collected temperature and sound data in the primary detection includes:

[0010] Temperature data preprocessing is used to eliminate environmental temperature interference in the real-time temperature data collected during the runtime period;

[0011] Audio data preprocessing is used to eliminate background noise interference in the real-time audio data acquired during runtime.

[0012] Preferably, the temperature data preprocessing is implemented as follows:

[0013] Real-time temperature of bearing outer ring collected and real-time temperature of the detection environment Calculate the original difference between the bearing temperature and the ambient temperature, eliminate common-mode interference from the ambient temperature, and the difference sequence is as follows: ;

[0014] The smoothed value of the difference sequence is used as the state variable. Based on the Kalman filter recursion, the filtered smooth difference sequence is obtained: ,in This is the optimal estimate at time t after Kalman filtering. This is the optimal estimate at time t-1. This represents the Kalman filter gain.

[0015] When the ambient temperature deviates from the bearing's rated operating temperature beyond a set threshold, correction is required, and a correction factor must be calculated. First, the ambient temperature deviation is defined. ,in To determine the rated operating temperature of the bearing, a quadratic polynomial fitting compensation coefficient is used. , ,in These are the coefficients of the constant term, the linear term, and the quadratic term, respectively.

[0016] Final effective temperature The result is the superposition of the compensated filtered difference and the ambient temperature, calculated as follows: .

[0017] Preferably, the specific implementation of the sound data preprocessing is as follows:

[0018] The collected sound signals Perform a short-time Fourier transform to obtain the time spectrum. , representing the first The time spectrum values ​​of each time frame at frequency f;

[0019] Based on the bearing rotational speed, the core characteristic frequency is calculated first. Next, frequency bands are divided into effective frequency bands. noise frequency band ,in The maximum frequency of the signal;

[0020] Initialize noise power spectrum ,in The initial noise power spectrum, The noise power spectrum is updated based on the total number of frames in the initial stationary phase. ,in For the first Frame in frequency Real-time noise power spectrum at the location;

[0021] Frequency-band noise suppression, time spectrum The time-spectrum values ​​are obtained by differentiating based on frequency bands. Finally, the denoised time spectrum was analyzed. The inverse short-time Fourier transform yields the denoised bearing target sound signal.

[0022] Preferably, the specific implementation of processing the temperature change data and sound change data during the current detection period to calculate the primary wear index is as follows:

[0023] First, preprocessed temperature data from the bearing's historical normal operating period was selected, and a normal temperature trend curve was obtained using nonlinear least squares fitting. And calculate the standard deviation of temperature fluctuations during that period. As a normal fluctuation threshold;

[0024] Calculate the instantaneous deviation of the temperature data from the normal temperature trend curve for the current detection period. ,in The data is preprocessed based on the temperature data for the current detection period, and the slopes of the current temperature trend curve and the normal temperature trend curve are obtained through linear approximation. ;

[0025] Calculate the initial value of temperature wear ,in, The instantaneous deviation value during the current detection period The average value;

[0026] Next, the correlation coefficient R between temperature trend and sound trend is calculated for the preprocessed sound data to obtain the sound conditioning coefficient. ,in This is the baseline correlation coefficient between temperature and sound trends during normal operation.

[0027] The initial wear index is obtained by normalizing the product of the initial temperature wear value and the sound adjustment coefficient.

[0028] Preferably, the image preprocessing in the secondary detection involves histogram equalization, which changes the histogram distribution of the image and increases image contrast.

[0029] Preferably, the secondary wear index is obtained by differentiating the bearing surface wear image from the preset no-wear image in the secondary detection as follows:

[0030] Feature differentiation is performed on the pre-processed bearing surface wear image and the preset no-wear image to obtain independent sub-images of rolling elements, inner ring slide, and outer ring slide;

[0031] A dual-branch contrastive convolutional network is constructed. The worn sub-image and the corresponding unworn sub-image of the region are respectively input into the two branches of the network to extract the wear feature vector and the baseline feature vector.

[0032] The wear feature vector of each region is compared with the baseline feature vector to obtain the regional difference features;

[0033] The regional difference features are weighted and fused to obtain the global difference features. The global difference features are then input into the fully connected layer for mapping and normalization, and then used as the secondary wear index.

[0034] Preferably, the specific implementation of constructing a dual-branch contrastive convolutional network, in which the worn sub-image and the corresponding unworn sub-image of the region are input into the two branches of the network respectively, and the worn feature vector and the baseline feature vector are extracted, is as follows:

[0035] First, the scale of each sub-image region is normalized and uniformly adjusted to the preset size;

[0036] A dual-branch contrastive convolutional network is constructed, with each branch containing 5 convolutional layers and 2 residual connection modules. The first branch takes a worn sub-image as input and extracts local wear features through 3×3 convolutional kernels; the second branch takes a non-worn sub-image as input and extracts baseline features through the same network structure.

[0037] Global average pooling is performed on the output features of the two branches respectively to obtain wear feature vectors and baseline feature vectors with consistent dimensions.

[0038] Preferably, the specific implementation of performing a differential operation on the wear feature vector and the reference feature vector of each region to obtain the regional difference features is as follows:

[0039] For elements at the same position in two vectors, calculate the absolute difference value (the absolute value of the difference between the two element values) and the weighted difference value (|wear element - reference element| × reference element); sum the absolute difference value and the weighted difference value in an equal weighted manner to obtain the composite difference value of the element.

[0040] A 3×1 sliding window is used to slide along the vector dimension. The sum and maximum value of the composite difference values ​​within the window are calculated, and the average of the two is taken as the enhanced difference value at the center of the window. The elements at the edge of the window keep the original composite difference values ​​unchanged, resulting in an enhanced one-dimensional difference vector.

[0041] Feature purification is performed on the enhanced one-dimensional difference vector. The median of all elements in the vector is calculated, and elements greater than the median are retained as effective difference features, while elements less than or equal to the median are removed. The retained effective difference features are arranged in their original position order to form the regional difference features of the region.

[0042] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0043] 1. Temperature data is used to eliminate common-mode interference through difference analysis, combined with Kalman filtering and smoothing, and environmental deviations are corrected using quadratic polynomial compensation coefficients; sound data is subjected to short-time Fourier transform for frequency band noise reduction, and the noise power spectrum is dynamically updated to solve the problem of existing technologies ignoring the dynamic characteristics of environmental interference.

[0044] 2. By integrating temperature and sound data, fitting a normal temperature trend curve, calculating temperature deviation and trend slope, and adjusting the correlation coefficient between the two, a primary wear index is formed, overcoming the limitations of single-parameter detection and improving the accuracy of preliminary judgment.

[0045] 3. Acquire images of the bearing surface, perform histogram equalization, extract wear and no-wear features using a bi-branch convolutional network, and obtain secondary indices through composite difference calculation.

[0046] 4. Weighted fusion of primary and secondary indices yields a comprehensive wear degree, with dual thresholds for judgment. Primary tests directly determine compliance, improving efficiency, while secondary tests enhance the accuracy of non-compliance determination, comprehensively improving detection reliability. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a schematic diagram of the structural process of a method for testing the wear resistance of bearings. Detailed Implementation

[0049] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described below in conjunction with the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0050] Numerous specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways than those described herein, and therefore the invention is not limited to the specific embodiments disclosed in the following specification.

[0051] In this embodiment, bearings, as key components bearing rotational motion in various mechanical equipment, directly affect the operational stability, service life, and safety of the equipment due to their wear resistance. Existing methods for testing bearing wear resistance mostly rely on single-sensor technology, ignoring dynamic interference from ambient temperature and background noise, and typically do not incorporate surface image features, resulting in low detection accuracy and high false alarm and false negative rates. Therefore, a method for testing bearing wear resistance is proposed, with the specific implementation steps as follows: Figure 1 As shown.

[0052] First, data acquisition is performed to obtain real-time data during the bearing's operation, including temperature data and sound data. The data contains interference parameters, including ambient temperature and background noise.

[0053] Then, a primary test is performed. The collected temperature and sound data are preprocessed, and the temperature and sound changes during the current testing period are processed to calculate the primary wear index. If the primary wear index is less than the first threshold, the bearing is deemed to have qualified wear resistance. If it is greater than or equal to the first threshold, a secondary test is performed.

[0054] The specific implementation of preprocessing the collected temperature and sound data in the primary detection includes: temperature data preprocessing to eliminate environmental temperature interference in the collected real-time temperature data during the runtime; and sound data preprocessing to eliminate background noise interference in the collected real-time sound data during the runtime.

[0055] Specifically, the temperature data preprocessing is implemented by: acquiring the real-time temperature of the bearing outer ring. and real-time temperature of the detection environment Calculate the original difference between the bearing temperature and the ambient temperature, eliminate common-mode interference from the ambient temperature, and the difference sequence is as follows: The smoothed value of the difference sequence is used as the state variable. Based on the Kalman filter recursion, the filtered smooth difference sequence is obtained: ,in This is the optimal estimate at time t after Kalman filtering. This is the optimal estimate at time t-1. The Kalman filter gain is used; when the deviation between the ambient temperature and the bearing's rated operating temperature exceeds a set threshold, correction is required, and the correction factor is calculated; first, the ambient temperature deviation is defined. ,in To determine the rated operating temperature of the bearing, a quadratic polynomial fitting compensation coefficient is used. , ,in These are the coefficients of the constant term, the linear term, and the quadratic term, respectively; the final effective temperature. The result is the superposition of the compensated filtered difference and the ambient temperature, calculated as follows: .

[0056] The specific implementation of the sound data preprocessing is as follows: processing the acquired sound signals... Perform a short-time Fourier transform to obtain the time spectrum. , representing the first The time-spectral values ​​of each time frame at frequency f; the core characteristic frequency is calculated based on the bearing rotation speed, first calculating the fundamental frequency. Next, frequency bands are divided into effective frequency bands. noise frequency band ,in The maximum frequency of the signal; initialize the noise power spectrum. ,in The initial noise power spectrum, The noise power spectrum is updated based on the total number of frames in the initial stationary phase. ,in For the first Frame in frequency Real-time noise power spectrum at the location;

[0057] Frequency-band noise suppression, time spectrum The time-spectrum values ​​are obtained by differentiating based on frequency bands. Finally, the denoised time spectrum was analyzed. The inverse short-time Fourier transform yields the denoised bearing target sound signal.

[0058] After preprocessing the temperature and sound data, it is necessary to process the temperature and sound change data for the current detection period to calculate the primary wear index. Specifically, firstly, preprocessed temperature data from the bearing's historical normal operating period is selected, and a normal temperature trend curve is obtained using nonlinear least squares fitting. And calculate the standard deviation of temperature fluctuations during that period. As a normal fluctuation threshold; for the temperature data of the current detection period, calculate its instantaneous deviation from the normal temperature trend curve. ,in The data is preprocessed based on the temperature data for the current detection period, and the slopes of the current temperature trend curve and the normal temperature trend curve are obtained through linear approximation. ;

[0059] Calculate the initial value of temperature wear ,in, The instantaneous deviation value during the current detection period The average value is then calculated; next, for the preprocessed sound data, the trend slope of the curve for the current detection period is calculated. Calculate the correlation coefficient R between temperature trend and sound trend to obtain the sound modulation coefficient. ,in The baseline correlation coefficient between temperature and sound trends during normal operation is used; the product of the initial temperature wear value and the sound adjustment coefficient is normalized to obtain the primary wear index.

[0060] If the primary wear index is less than the first threshold, the bearing is deemed to have qualified wear resistance; if it is greater than or equal to the first threshold, a secondary test is performed.

[0061] Secondary detection: Acquire bearing surface wear images and preprocess them. The bearing surface wear images include rolling element images, inner raceway images, and outer raceway images. The secondary wear index is calculated by differentiating the bearing surface wear images from preset no-wear images. The image preprocessing involves histogram equalization to change the histogram distribution and increase image contrast.

[0062] The process of calculating the secondary wear index by differentiating the bearing surface wear image with a preset no-wear image involves distinguishing features between the preprocessed bearing surface wear image and the preset no-wear image to obtain independent sub-images of the rolling element, inner raceway, and outer raceway. A dual-branch contrastive convolutional network is constructed, and the wear sub-image and the corresponding no-wear sub-image of the region are input into the two branches of the network to extract wear feature vectors and reference feature vectors. Differentiation operations are performed on the wear feature vector and the reference feature vector of each region to obtain regional difference features. The regional difference features are weighted and fused to obtain global difference features. The global difference features are input into a fully connected layer for mapping and normalization to serve as the secondary wear index.

[0063] Specifically, after feature differentiation is performed on the preprocessed bearing surface wear image and the preset no-wear image to obtain independent sub-images of the rolling element, inner raceway, and outer raceway, a dual-branch contrastive convolutional neural network is constructed to efficiently compare and analyze the wear area and the no-wear area. The two branches of the network take the wear area sub-image and the corresponding no-wear sub-image as inputs, respectively. Each branch uses the same convolutional and pooling layer structure to normalize the scale of each region's sub-image, adjusting it to a preset size. The dual-branch contrastive convolutional network consists of 5 convolutional layers and 2 residual connection modules in each branch. The first branch takes the wear sub-image as input and extracts local wear features using 3×3 convolutional kernels; the second branch takes the no-wear sub-image as input and extracts baseline features using the same network structure. Global average pooling is then performed on the output features of both branches to obtain wear feature vectors and baseline feature vectors with consistent dimensions. These features extract local texture and shape information through convolutional layers and undergo spatial dimensionality reduction through pooling layers to reduce computational complexity while preserving important high-level features. The output of each branch is the feature vector for that region. For elements at the same position in the two vectors, the absolute difference value (the absolute value of the difference between the two element values) and the weighted difference value (|wear element - reference element| × reference element) are calculated separately. The absolute difference value and the weighted difference value are then summed with equal weights to obtain the composite difference value for that element. A 3×1 sliding window is used to slide along the vector dimension, calculating the sum and maximum value of the composite difference values ​​within the window. The average of these two values ​​is taken as the enhanced difference value at the center of the window. Elements at the window edges retain their original composite difference values, resulting in an enhanced one-dimensional difference vector. Feature purification is performed on this enhanced one-dimensional difference vector by calculating the median of all elements. Elements greater than the median are retained as effective difference features, while elements less than or equal to the median are removed. The retained effective difference features are arranged in their original position order to form the regional difference features. Finally, after calculating the difference features for each region, a weighted fusion method is used to merge the difference features of each region to comprehensively evaluate the overall wear of the bearing. The weighted fusion method takes into account that different regions may have different importance in the wear process. Therefore, it is necessary to weight the difference features of each region according to the preset weight coefficient, sum the weighted difference features of all regions to obtain the global difference features, input the global difference features into the fully connected layer for mapping, and after normalization, use them as the secondary wear index.

[0064] Finally, a comprehensive evaluation is conducted. The overall wear degree is determined by weighted calculation based on the primary wear index and the secondary wear index. If the overall wear degree is greater than the second threshold, the bearing wear resistance is deemed unqualified and an alarm is triggered; otherwise, it is deemed qualified and the test results are recorded.

[0065] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments for application in other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for testing the wear resistance of bearings, characterized in that, Includes the following steps: Data acquisition: Real-time data during bearing operation is acquired, including temperature data and sound data. The data contains interference parameters, including ambient temperature and background noise. Primary inspection: The collected temperature and sound data are preprocessed, and the temperature and sound change data during the current inspection period are processed to calculate the primary wear index. If the primary wear index is less than the first threshold, the bearing is judged to be qualified in terms of wear resistance. If it is greater than or equal to the first threshold, secondary inspection is performed. Secondary detection: Acquire bearing surface wear images and preprocess the images; the bearing surface wear images include rolling element images, inner ring slide images and outer ring slide images, and the secondary wear index is obtained by differentiating the bearing surface wear images with preset no-wear images; Comprehensive assessment: The comprehensive wear degree is determined by weighted calculation based on the primary wear index and the secondary wear index. If the comprehensive wear degree is greater than the second threshold, the bearing wear resistance is deemed unqualified and an alarm is triggered; otherwise, it is deemed qualified and the test results are recorded. The specific implementation of processing temperature and sound change data during the current detection period to calculate the primary wear index is as follows: First, preprocessed temperature data from the bearing's historical normal operating period was selected, and a normal temperature trend curve was obtained using nonlinear least squares fitting. And calculate the standard deviation of temperature fluctuations during that period. As a normal fluctuation threshold; Calculate the instantaneous deviation of the temperature data from the normal temperature trend curve for the current detection period. ,in The data is preprocessed based on the temperature data for the current detection period, and the slopes of the current temperature trend curve and the normal temperature trend curve are obtained through linear approximation. ; Calculate the initial value of temperature wear ,in, The instantaneous deviation value during the current detection period The average value; Next, the trend slope of the curve for the current detection period is calculated for the preprocessed audio data. Calculate the correlation coefficient R between temperature trend and sound trend to obtain the sound modulation coefficient. ,in This is the baseline correlation coefficient between temperature and sound trends during normal operation. The initial wear index is obtained by normalizing the product of the initial temperature wear value and the sound adjustment coefficient.

2. The method for testing the wear resistance of bearings according to claim 1, characterized in that, The specific implementation of preprocessing the collected temperature and sound data in the primary detection includes: Temperature data preprocessing is used to eliminate environmental temperature interference in the real-time temperature data collected during the runtime period; Audio data preprocessing is used to eliminate background noise interference in the real-time audio data acquired during runtime.

3. The method for testing the wear resistance of bearings according to claim 2, characterized in that, The specific implementation of the temperature data preprocessing is as follows: Real-time temperature of bearing outer ring collected and real-time temperature of the detection environment Calculate the original difference between the bearing temperature and the ambient temperature, eliminate common-mode interference from the ambient temperature, and the difference sequence is as follows: ; The smoothed value of the difference sequence is used as the state variable. Based on the Kalman filter recursion, the filtered smooth difference sequence is obtained: ,in This is the optimal estimate at time t after Kalman filtering. This is the optimal estimate at time t-1. This represents the Kalman filter gain. When the ambient temperature deviates from the bearing's rated operating temperature beyond a set threshold, correction is required, and a correction factor must be calculated. First, the ambient temperature deviation is defined. ,in To determine the rated operating temperature of the bearing, a quadratic polynomial fitting compensation coefficient is used. , ,in These are the coefficients of the constant term, the linear term, and the quadratic term, respectively. Final effective temperature The result is the superposition of the compensated filtered difference and the ambient temperature, calculated as follows: .

4. The method for testing the wear resistance of bearings according to claim 2, characterized in that, The specific implementation of the audio data preprocessing is as follows: The collected sound signals Perform a short-time Fourier transform to obtain the time spectrum. , representing the first The time spectrum values ​​of each time frame at frequency f; Based on the bearing rotational speed, the core characteristic frequency is calculated first. Next, frequency bands are divided into effective frequency bands. noise frequency band ,in The maximum frequency of the signal; Initialize noise power spectrum ,in The initial noise power spectrum, The noise power spectrum is updated based on the total number of frames in the initial stationary phase. ,in For the first Frame in frequency Real-time noise power spectrum at the location; Frequency-band noise suppression, time spectrum The time-spectrum values ​​are obtained by differentiating based on frequency bands. Finally, the denoised time spectrum was analyzed. The inverse short-time Fourier transform yields the denoised bearing target sound signal.

5. The method for testing the wear resistance of bearings according to claim 1, characterized in that, The preprocessing of the image in the secondary detection involves histogram equalization, which changes the histogram distribution of the image and increases image contrast.

6. The method for testing the wear resistance of bearings according to claim 1, characterized in that, The secondary wear index is obtained by differentiating the bearing surface wear image from the preset no-wear image in the secondary detection process as follows: Feature differentiation is performed on the pre-processed bearing surface wear image and the preset no-wear image to obtain independent sub-images of rolling elements, inner ring slide, and outer ring slide; A dual-branch contrastive convolutional network is constructed. The worn sub-image and the corresponding unworn sub-image of the region are respectively input into the two branches of the network to extract the wear feature vector and the baseline feature vector. The wear feature vector of each region is compared with the baseline feature vector to obtain the regional difference features; The regional difference features are weighted and fused to obtain the global difference features. The global difference features are then input into the fully connected layer for mapping and normalization, and then used as the secondary wear index.

7. The method for testing the wear resistance of bearings according to claim 6, characterized in that, The specific implementation of constructing a dual-branch contrastive convolutional network, in which the worn sub-image and the corresponding unworn sub-image of the region are input into the two branches of the network respectively, and the worn feature vector and the baseline feature vector are extracted, is as follows: First, the scale of each sub-image region is normalized and uniformly adjusted to the preset size; A dual-branch contrastive convolutional network is constructed, with each branch containing 5 convolutional layers and 2 residual connection modules. The first branch takes a worn sub-image as input and extracts local wear features through 3×3 convolutional kernels; the second branch takes a non-worn sub-image as input and extracts baseline features through the same network structure. Global average pooling is performed on the output features of the two branches respectively to obtain wear feature vectors and baseline feature vectors with consistent dimensions.

8. A method for testing the wear resistance of bearings according to claim 6, characterized in that, The specific implementation of performing a differential operation between the wear feature vector and the baseline feature vector of each region to obtain the regional difference features is as follows: For elements at the same position in two vectors, calculate the absolute difference value (the absolute value of the difference between the two element values) and the weighted difference value (|wear element - reference element| × reference element); sum the absolute difference value and the weighted difference value in an equal weighted manner to obtain the composite difference value of the element. A 3×1 sliding window is used to slide along the vector dimension. The sum and maximum value of the composite difference values ​​within the window are calculated, and the average of the two is taken as the enhanced difference value at the center of the window. The elements at the edge of the window keep the original composite difference values ​​unchanged, resulting in an enhanced one-dimensional difference vector. Feature purification is performed on the enhanced one-dimensional difference vector. The median of all elements in the vector is calculated, and elements greater than the median are retained as effective difference features, while elements less than or equal to the median are removed. The retained effective difference features are arranged in their original position order to form the regional difference features of the region.

Images