Rolling bearing state evaluation method, system, terminal device and medium

Abstract

The application provides a rolling bearing state evaluation method, system, terminal equipment and medium, belongs to the field of rolling bearing fault prediction and health management, including obtaining the full life vibration signal data of multiple bearings; based on the surface area defect law, the degradation feature set of the bearing is extracted; according to the degradation feature set, a bearing degradation model for predicting the degradation amount of the bearing is constructed, and the bearing degradation model is trained to obtain the trained bearing degradation model; the vibration signal of the rolling bearing to be evaluated is obtained, the vibration signal is input into the trained bearing degradation model, the degradation amount prediction value of the rolling bearing to be evaluated is obtained, and the residual life prediction value of the rolling bearing to be evaluated is calculated based on the first arrival time and the degradation amount prediction value; according to the life prediction value, the service state of the rolling bearing to be evaluated is evaluated to obtain the state evaluation result of the rolling bearing to be evaluated. The application can improve the accuracy of the service state evaluation of the rolling bearing.

Description

Technical Field

[0001] This invention belongs to the field of rolling bearing failure prediction and health management, specifically involving a rolling bearing condition assessment method, system, terminal equipment, and medium. Background Technology

[0002] During operation, mechanical equipment inevitably degrades due to various factors. This degradation process first occurs in components that generate relative motion, especially rolling bearings. Therefore, to ensure that mechanical equipment operates in a safe condition, it is essential to assess the operational status of rolling bearings. Remaining life prediction methods are widely recognized as a fundamental and effective method for assessing the condition of rolling bearings. If the remaining service life of a rolling bearing can be predicted, its current service condition can be assessed. Currently, in the field of rolling bearing remaining life prediction, scholars have proposed a series of methods, which can generally be categorized into expert knowledge base-based, data-driven, physical model-based, and hybrid methods. Expert knowledge base-based methods require specialized knowledge of fault data, resulting in high acquisition costs; data-driven methods show decreased accuracy in remaining life prediction as the prediction time span increases; the accuracy of physical model-based methods largely depends on the accuracy of the physical model used; hybrid prediction methods combine the advantages of physical models and data-driven approaches, but these methods are algorithmically complex and difficult to model. Due to the limitations of different methods, the lack of clarity in the failure mechanism of rolling bearings, the scarcity of degradation data, and especially the neglect of historical operating data under normal service conditions, these methods cannot accurately assess the service condition of rolling bearings.

[0003] Therefore, there is an urgent need for a method that can accurately assess the service condition of rolling bearings. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method, system, terminal equipment and medium for assessing the condition of rolling bearings, so as to improve the accuracy of assessing the service condition of rolling bearings.

[0005] This invention provides a method for evaluating the condition of rolling bearings, comprising:

[0006] The vibration signal of the rolling bearing to be evaluated is input into a pre-trained bearing degradation model to obtain the predicted value of the degradation amount of the rolling bearing to be evaluated; wherein, the bearing degradation model is trained based on a degradation feature set, the degradation feature set is extracted from the whole life vibration signal data based on the surface area defect law, and the degradation feature set includes at least one degradation feature, the degradation feature includes local fluctuation feature and degradation trend feature;

[0007] Based on the first arrival time and the predicted degradation value, the predicted remaining life of the rolling bearing to be evaluated is calculated; wherein, the first arrival time is the moment when the state value first exceeds the threshold.

[0008] Based on the predicted remaining life value, the service status of the rolling bearing to be evaluated is assessed to obtain the status assessment result of the rolling bearing to be evaluated.

[0009] Furthermore, the bearing degradation model is trained in the following manner:

[0010] Acquire full-life vibration signal data from multiple bearings of the same model as the rolling bearing to be evaluated;

[0011] Based on the surface area defect law, a degradation feature set of the bearing is extracted from the full-life vibration signal data; the degradation feature set is used to characterize the correlation between the bearing's service life and its service time.

[0012] Based on the degradation feature set, a bearing degradation model is constructed to predict the amount of bearing degradation. The bearing degradation model is then trained using the full-life vibration signal data until the loss value of the bearing degradation model is less than a preset loss threshold, thus obtaining the trained bearing degradation model.

[0013] Furthermore, the expression for the degradation feature is:

[0014]

[0015] in, Indicates the first The degradation characteristics of individual bearings Indicates the first The service life of each bearing. Indicates the first The first bearing Each service period, Indicates characteristics of degradation trend. A scaling factor to control the intensity of noise. It is a fixed parameter that is related to the properties of the material. This indicates noise and is used to describe local fluctuations during bearing degradation.

[0016] Furthermore, the expression for the bearing degradation model is:

[0017]

[0018] in, Indicates the bearing's service life The amount of degradation at that time This indicates the initial state of the bearing. The drift coefficient represents the difference between bearings and follows a normal distribution, i.e. , It is a fixed parameter that is related to the properties of the material. The diffusion coefficient is used to describe the degree of fluctuation during bearing degradation. This represents standard Brownian motion, used to represent the inherent variability of random degradation processes over time.

[0019] Furthermore, the step of calculating the predicted remaining life of the rolling bearing to be evaluated based on the first arrival time and the predicted degradation amount specifically involves: determining a given threshold for failure prediction according to industry standards or a relative method. , This indicates that the bearing, calculated based on the first arrival time, has exceeded a given threshold from its current state. The required time is determined by estimating all unknown parameters in the probability density function using the predicted degradation value, and then using the remaining service life density function of the rolling bearing. Calculate the probability that the remaining life of the rolling bearing is in the range [0, +∞). and Integrating the product over [0, +∞) yields... Predicted remaining life of rolling bearings at all times .

[0020] Furthermore, the aforementioned and The calculation is performed as follows:

[0021] Through calculation formula

[0022]

[0023] get Density function of remaining service life of rolling bearings ;in, This indicates that the bearing, calculated based on the first arrival time, has exceeded a given threshold from its current state. The time required , It is a value greater than 0, in order to ensure that the rolling bearing is in The failure point has not yet been reached. This indicates a pre-defined threshold. express The state value at time t, where c represents the diffusion coefficient. express variance express The mean, , , , It is a fixed parameter that is related to the properties of the material;

[0024] Through calculation formula

[0025]

[0026] get Predicted remaining life of rolling bearings at all times .

[0027] Further, the step of evaluating the service status of the rolling bearing to be evaluated based on the predicted service life value specifically involves adding the predicted remaining service life value to the service time of the rolling bearing to obtain the total service life of the rolling bearing.

[0028] The condition assessment index is obtained by dividing the predicted remaining life of the rolling bearing by the total life of the rolling bearing.

[0029] The service condition of the rolling bearing to be evaluated is assessed based on the aforementioned condition assessment indicators.

[0030] Furthermore, the step of evaluating the service condition of the rolling bearing to be evaluated based on the condition evaluation index includes:

[0031] When the status evaluation index is 0, the service status of the rolling bearing to be evaluated is determined to be unusable.

[0032] When the condition evaluation index is 100%, the service condition of the rolling bearing to be evaluated is determined to be excellent.

[0033] Furthermore, the state assessment index The following calculation formula is used to obtain...

[0034]

[0035] in, Indicates the status assessment index, , This indicates the time the bearing has been in service.

[0036] Furthermore, the full-lifetime vibration signal data was obtained through accelerated degradation experiments.

[0037] The present invention also provides a rolling bearing condition assessment system, comprising:

[0038] The degradation prediction module is used to input the vibration signal of the rolling bearing to be evaluated into a pre-trained bearing degradation model to obtain the predicted value of the degradation of the rolling bearing to be evaluated; wherein, the bearing degradation model is trained based on a degradation feature set, the degradation feature set is extracted from the whole life vibration signal data based on the surface area defect law, and the degradation feature set includes at least one degradation feature, including local fluctuation feature and degradation trend feature;

[0039] The life prediction module is used to calculate the remaining life prediction value of the rolling bearing to be evaluated based on the first arrival time and the predicted degradation value; wherein, the first arrival time is the moment when the state value first exceeds a threshold.

[0040] The condition assessment module is used to assess the service condition of the rolling bearing to be assessed based on the remaining life prediction value, and obtain the condition assessment result of the rolling bearing to be assessed.

[0041] Furthermore, the system also includes a model training module, which is used to train the bearing degradation model in the following manner:

[0042] The data acquisition module acquires full-life vibration signal data from multiple bearings of the same model as the rolling bearing to be evaluated.

[0043] Based on the surface area defect law, the feature extraction module extracts the degradation feature set of the bearing from the full-life vibration signal data; the degradation feature set is used to characterize the correlation between the bearing's service life and its service time.

[0044] The model training module constructs a bearing degradation model for predicting the amount of bearing degradation based on the degradation feature set, and trains the bearing degradation model using the full-life vibration signal data until the loss value of the bearing degradation model is less than a preset loss threshold, thus obtaining the trained bearing degradation model.

[0045] Furthermore, the expression for the degradation feature is:

[0046]

[0047] in, Indicates the first The degradation characteristics of individual bearings Indicates the first The service life of each bearing. Indicates the first The first bearing Each service period, Indicates characteristics of degradation trend. A scaling factor to control the intensity of noise. It is a fixed parameter that is related to the properties of the material. This indicates noise and is used to describe local fluctuations during bearing degradation.

[0048] Furthermore, the expression for the bearing degradation model is:

[0049]

[0050] in, Indicates the bearing's service life The amount of degradation at that time This indicates the initial state of the bearing. The drift coefficient represents the difference between bearings and follows a normal distribution, i.e. , It is a fixed parameter that is related to the properties of the material. The diffusion coefficient is used to describe the degree of fluctuation during bearing degradation. This represents standard Brownian motion, used to represent the inherent variability of random degradation processes over time.

[0051] Furthermore, the step of calculating the predicted remaining life of the rolling bearing to be evaluated based on the first arrival time and the predicted degradation amount specifically involves: determining a given threshold for failure prediction according to industry standards or a relative method. , This indicates that the bearing, calculated based on the first arrival time, has exceeded a given threshold from its current state. The required time is determined by estimating all unknown parameters in the probability density function using the predicted degradation value, and then using the remaining service life density function of the rolling bearing. Calculate the probability that the remaining life of the rolling bearing is in the range [0, +∞). and Integrating the product over [0, +∞) yields... Predicted remaining life of rolling bearings at all times .

[0052] Furthermore, the aforementioned and The calculation is performed as follows:

[0053] Through calculation formula

[0054]

[0055] get Density function of remaining service life of rolling bearings ;in, This indicates that the bearing, calculated based on the first arrival time, has exceeded a given threshold from its current state. The time required , It is a value greater than 0, in order to ensure that the rolling bearing is in The failure point has not yet been reached. This indicates a pre-defined threshold. express The state value at time t, where c represents the diffusion coefficient. express variance express The mean, , , , It is a fixed parameter that is related to the properties of the material;

[0056] Through calculation formula

[0057]

[0058] get Predicted remaining life of rolling bearings at all times .

[0059] Further, the step of evaluating the service status of the rolling bearing to be evaluated based on the predicted service life value specifically involves adding the predicted remaining service life value to the service time of the rolling bearing to obtain the total service life of the rolling bearing.

[0060] The condition assessment index is obtained by dividing the predicted remaining life of the rolling bearing by the total life of the rolling bearing.

[0061] The service condition of the rolling bearing to be evaluated is assessed based on the aforementioned condition assessment indicators.

[0062] Furthermore, the step of evaluating the service condition of the rolling bearing to be evaluated based on the condition evaluation index includes:

[0063] When the status evaluation index is 0, the service status of the rolling bearing to be evaluated is determined to be unusable.

[0064] When the condition evaluation index is 100%, the service condition of the rolling bearing to be evaluated is determined to be excellent.

[0065] Furthermore, the state assessment index The following calculation formula is used to obtain...

[0066]

[0067] in, Indicates the status assessment index, , This indicates the time the bearing has been in service.

[0068] Furthermore, the full-lifetime vibration signal data was obtained through accelerated degradation experiments.

[0069] The present invention also provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method.

[0070] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0071] The beneficial effects of this invention are:

[0072] This invention discloses a method, system, terminal device, and medium for assessing the condition of rolling bearings. Based on the surface area defect law, it extracts a set of bearing degradation features from full-life vibration signal data. The obtained degradation features include local fluctuation characteristics and degradation trend characteristics. On one hand, relying on the surface area defect law allows for a more accurate description of the bearing degradation mechanism. On the other hand, extracting bearing degradation features from full-life vibration signal data overcomes the shortcomings of traditional methods that ignore historical operating data under normal service conditions. Furthermore, the integration of local fluctuation characteristics and degradation trend characteristics improves the interpretability of the degradation features, better reflects the physical characteristics of rolling bearings in actual working conditions, and improves the accuracy of the degradation features, thereby enhancing the accuracy of rolling bearing condition assessment. Based on the first arrival time and the predicted degradation amount, a predicted remaining life value is calculated. Finally, based on the predicted life value, the service condition of the rolling bearing to be assessed is evaluated. The method allows for flexible adjustment of the judgment threshold for the remaining life prediction to address differences in the remaining life of the bearing under different working conditions, thereby improving the accuracy of rolling bearing condition assessment. Attached Figure Description

[0073] Figure 1 This is a flowchart of a rolling bearing condition assessment method according to one embodiment of this application;

[0074] Figure 2 This is a schematic diagram of a degradation feature set in one embodiment of this application;

[0075] Figure 3 This is a schematic diagram of the probability density function of the remaining useful life in one embodiment of this application;

[0076] Figure 4 This is a comparison chart of the remaining service life prediction results at different monitoring points in one embodiment of this application;

[0077] Figure 5 This is a schematic diagram of the service status of the target rolling bearing at different monitoring points in one embodiment of this application;

[0078] Figure 6 This is a schematic diagram of the structure of a rolling bearing condition assessment system according to one embodiment of this application;

[0079] Figure 7 This is a schematic diagram of the structure of a terminal device in one embodiment of this application. Detailed Implementation

[0080] Existing methods for assessing the condition of rolling bearings include expert knowledge base-based, data-driven, physical model-based, and hybrid methods. Expert knowledge base-based methods achieve prediction by comparing observed data with a pre-defined fault database using expert systems or fuzzy systems. This method typically requires specialized knowledge of the fault data, which is costly to acquire in practice. Data-driven methods utilize historical equipment condition data to extract features related to changes in the monitored object's condition. Through statistical analysis, pattern recognition, and machine learning, they attempt to simulate a fuzzy functional relationship between sensor data and equipment condition, thereby achieving condition assessment and remaining life prediction. However, this method requires establishing a condition representation function for the rolling bearing, and the model's representational ability decreases with increasing prediction time span, leading to a decline in remaining life prediction accuracy. Physical model-based methods predict bearing degradation performance and remaining service life based on the mathematical representation of physical behavior during degradation. While this method can provide accurate predictions, it still requires a deep understanding of the bearing's physical characteristics, and the accuracy of the prediction largely depends on the accuracy of the physical model used. Hybrid prediction methods combine the advantages of physical models and data-driven approaches for remaining life prediction. This method can effectively simulate the degradation process of rolling bearings. However, this approach complicates the algorithm and is limited by the physical behavior of rolling bearings during degradation, leading to modeling difficulties.

[0081] The above methods cannot accurately assess the service condition of rolling bearings for two reasons: first, the degradation characteristics used cannot accurately characterize the degradation trajectory of rolling bearings; second, the degradation models used cannot map the failure mechanism of rolling bearings.

[0082] This invention discloses a method, system, terminal equipment, and medium for assessing the condition of rolling bearings. The method, based on the surface area defect law, extracts a set of bearing degradation features from full-life vibration signal data. These degradation features include local fluctuation characteristics and degradation trend characteristics. Firstly, relying on the surface area defect law allows for a more accurate description of the bearing degradation mechanism. Secondly, extracting these features from full-life vibration signal data overcomes the shortcomings of traditional methods that ignore historical operating data under normal service conditions. Furthermore, the integration of local fluctuation and degradation trend characteristics improves the interpretability of the degradation features, better reflects the physical characteristics of rolling bearings in actual operating conditions, and enhances the accuracy of the degradation features, thereby improving the accuracy of rolling bearing condition assessment. The remaining life prediction value is calculated based on the first arrival time (the moment when the state value first exceeds a threshold) and the predicted degradation amount. Finally, the service condition of the rolling bearing to be assessed is evaluated based on the life prediction value. This allows for flexible adjustment of the judgment threshold for the remaining life prediction to accommodate differences in the remaining life of bearings under different operating conditions, thereby improving the accuracy of rolling bearing condition assessment.

[0083] The rolling bearing condition assessment method disclosed in this invention will be described below.

[0084] like Figure 1 As shown, the rolling bearing condition assessment method disclosed in this invention includes the following steps:

[0085] Step 11: Obtain full-life vibration signal data of multiple bearings of the same model as the rolling bearing to be evaluated.

[0086] Specifically, multiple unused bearings of the same model as the rolling bearing to be evaluated are subjected to accelerated degradation experiments. In one embodiment of the invention, the vibration signal of the bearing is collected every 10 seconds, with a sampling duration of 1 second and a sampling frequency of 25.6 kHz, until the bearing completely fails, thus obtaining the full-life vibration signals of multiple bearings. It is worth noting that in the embodiments of the invention, the sampling interval and sampling duration are not fixed and can be adjusted according to the bearing's operating environment. If the bearing is in a high-speed, high-load operating environment, it is recommended to choose a shorter sampling interval to capture high-frequency vibration signals; if the bearing is operating in a low-speed, light-load environment, the sampling interval can be appropriately extended to reduce data redundancy.

[0087] The criterion for determining complete bearing failure is that when the amplitude of the vibration signal of the rolling bearing during operation exceeds 10 times the amplitude of its initial operating vibration signal, the rolling bearing is defined as having failed.

[0088] Furthermore, in other embodiments of the present invention, anomaly processing is performed on the collected full-life vibration signal data. For example, during the operation of a rolling bearing, environmental and operational issues may cause anomalies in the recorded full-life vibration signal data. For such anomalies, box plot analysis can be used; values ​​greater than or less than the upper and lower boundaries set by the box plot are considered anomalies. Then, the identified anomalies are corrected using the average of their upper and lower neighboring points. Meanwhile, due to the different scales of the input data, the numerical differences are significant. The main strategy for addressing this problem is to normalize the input parameter data. The most commonly used method for data normalization is min-max standardization, also known as deviation standardization, which is a linear transformation of the original data, mapping the resulting values ​​to the range [0, 1].

[0089] Step 12: Based on the surface area defect law, extract the bearing degradation feature set from the full-life vibration signal data.

[0090] The aforementioned degradation feature set is used to characterize the relationship between bearing service life and bearing service time. The degradation feature set includes at least one degradation feature, which includes local fluctuation features and degradation trend features.

[0091] For ease of explanation, the surface area defect law in this application will be explained first.

[0092] In an embodiment of the present invention, the surface area defect law is Paris's surface area defect law, which is specifically as follows: .in D The defect area is... N For the number of load cycles, n , C 0 is a defined material constant.

[0093] Based on the above surface area defect law, the expression for the degradation features in the degradation feature set is:

[0094]

[0095] in, Indicates the first The degradation characteristics of individual bearings Indicates the first The service life of each bearing. Indicates the first The first bearing Each service period, Indicates characteristics of degradation trend. A scaling factor to control the intensity of noise. It is a fixed parameter that is related to the properties of the material. This indicates noise and is used to describe local fluctuations during bearing degradation.

[0096] For example, in one embodiment of the present invention, the acquired set of degradation features is as follows: Figure 2 As shown. By Figure 2 It can be seen that, from Figure 2 It can be seen that the degradation index exhibits a clear trend of continuous increase with each sample point. Although there are staged changes characterized by a gradual increase in the early stage and an acceleration in the later stage, the degradation amount always maintains the same cumulative trend in each stage, laying a clear trend foundation for equipment life prediction. At the same time, the core upward trend of the index is not broken under local random fluctuations, demonstrating good robustness. Its signal-to-noise ratio is reasonable, which can effectively resist short-term interference, ensure the stability of the degradation model fitting, improve the reliability of life prediction results, and avoid misjudgment of degradation patterns caused by local fluctuations. Therefore, the constructed degradation index (degradation feature) has good trend and robustness.

[0097] It is worth mentioning that this invention, based on the surface area defect law, extracts the degradation feature set of bearings from full-life vibration signal data. The obtained degradation features include local fluctuation features and degradation trend features. On the one hand, relying on the surface area defect law can more accurately describe the degradation mechanism of bearings. On the other hand, extracting the degradation features of bearings from full-life vibration signal data overcomes the shortcomings of traditional methods that ignore historical operating data when rolling bearings are in normal service conditions. Furthermore, the degradation features integrate local fluctuation features and degradation trend features, improving the interpretability of the degradation features, making them more consistent with the physical characteristics of rolling bearings in actual working conditions, and improving the accuracy of the degradation features.

[0098] Step 13: Based on the degradation feature set, construct a bearing degradation model to predict the amount of bearing degradation, and train the bearing degradation model using full-life vibration signal data until the loss value of the bearing degradation model is less than the preset loss threshold, thus obtaining the trained bearing degradation model.

[0099] In an embodiment of the present invention, the expression for the bearing degradation model is:

[0100]

[0101] in, Indicates the bearing's service life The amount of degradation at that time This indicates the initial state of the bearing. The drift coefficient represents the difference between bearings and follows a normal distribution, i.e. , It is a fixed parameter that is related to the properties of the material. The diffusion coefficient is used to describe the degree of fluctuation during bearing degradation. This represents standard Brownian motion, used to represent the inherent variability of random degradation processes over time.

[0102] In an embodiment of the present invention, the bearing degradation model described above is a multi-scale convolutional neural network.

[0103] The training process of the bearing degradation model is explained below.

[0104] Specifically, at the start of training, the network's weights and bias parameters need to be randomly initialized. Common initialization methods include Xavier initialization and He initialization, which help avoid gradient vanishing or exploding problems.

[0105] Forward propagation: The input data is passed through each layer of the network, and the output of each layer is calculated. For convolutional layers, the input data is convolved with the convolutional kernel to extract features. Activation functions (such as ReLU, Sigmoid, etc.) are used to increase the non-linearity of the network. Finally, the degradation information mining results are output through fully connected layers.

[0106] Loss calculation: The loss function is determined based on the degradation mechanism of rolling bearings and cross-entropy. The error between the model output and the true label is calculated to determine the loss value and reflect the performance of the current model.

[0107] Backpropagation: The backpropagation algorithm is used to calculate the gradient of the loss function with respect to each parameter (weights and biases). Using the chain rule, the gradient propagates from the output layer back to the input layer.

[0108] Parameter Update: Optimization algorithms, such as Stochastic Gradient Descent (SGD) and Adam, are used to update the network parameters based on the calculated gradients. The choice of optimization algorithm and hyperparameter settings have a significant impact on training performance. Repeated Training: The above process is repeated multiple times on the entire training dataset. After each training round, the network readjusts its parameters, gradually improving its ability to fit the data.

[0109] Model evaluation: When the loss value of the bearing degradation model is less than the preset loss threshold, the bearing degradation model fit is determined, and the trained bearing degradation model is obtained.

[0110] Step 14: Acquire the vibration signal of the rolling bearing to be evaluated, input the vibration signal into the trained bearing degradation model to obtain the predicted degradation amount of the rolling bearing to be evaluated, and calculate the predicted remaining life of the rolling bearing to be evaluated based on the first arrival time and the predicted degradation amount. Determine the failure prediction according to industry standards or the relative method. Using the predicted degradation value and fitting method to analyze the parameters in the probability density function Make an estimate (where, express The mean of ; b is a fixed parameter related to material properties; c represents the diffusion coefficient), specifically, the bearing degradation model is used as the fitting function for fitting and solving. Parameters The maximum likelihood estimation method can be used to obtain ( express (variance). After estimating all the unknown parameters in the probability density function, the probability density function of the remaining lifetime can be solved, and then the remaining lifetime can be calculated.

[0111] The following describes the process of calculating the predicted life value of the rolling bearing to be evaluated based on the first arrival time and the predicted degradation value in step 14, including steps 14.1 to 14.2.

[0112] Step 14.1, using the calculation formula

[0113]

[0114] get Density function of remaining service life of rolling bearings ;in, This indicates that the bearing, calculated based on the first arrival time, has exceeded a given threshold from its current state. The time required , It is a value greater than 0, in order to ensure that the rolling bearing is in The failure point has not yet been reached. Thresholds are typically determined using industry standards or relative methods. express The state value at time t, where c represents the diffusion coefficient. express variance express The mean, , , It has no practical meaning; to simplify the expression above, we say: , , , b is a fixed parameter related to material properties; for example, in one embodiment of the present invention, the obtained remaining useful life density function is as follows: Figure 3 As shown. From Figure 3 As can be seen, with the increase of historical data, that is, with the increase of the number of condition monitoring points, the peak value of the probability density function increases and the opening narrows, proving that the probability density function is becoming more and more convergent, indicating that the reliability of the remaining lifetime prediction is also becoming higher and higher.

[0115] Step 14.2, using the calculation formula

[0116]

[0117] get Predicted remaining life of rolling bearings at all times .

[0118] In one embodiment of the present invention, 23 predictions were made for the target rolling bearing, and the prediction results at different monitoring points are as follows: Figure 4 As shown. From Figure 4 As can be seen, the predicted results at different monitoring points are close to the actual remaining life of the target rolling bearing.

[0119] It should be noted that, in the embodiments of the present invention, the parameters can be determined using predicted degradation values ​​and fitting methods. The model parameters are estimated; specifically, the bearing degradation model is used as the fitting function to solve the problem. Parameters It can be obtained using the maximum likelihood estimation method. Based on the properties of the Wiener process, the sample... Following a multivariate normal distribution, let Then its mean and variance are shown in the following formulas:

[0120]

[0121] The probability density function of the multivariate normal distribution is obtained from the above formula. Taking its logarithm yields the likelihood function containing the parameters. Then, the partial derivative of the likelihood function with respect to the parameters is used to obtain the solution parameters. The expression is as follows:

[0122]

[0123] Step 15: Based on the life prediction value, evaluate the service condition of the rolling bearing to be evaluated to obtain the condition evaluation result of the rolling bearing to be evaluated.

[0124] The remaining life prediction value is added to the service time of the rolling bearing to obtain the total life of the rolling bearing; the remaining life prediction value of the rolling bearing is divided by the total life of the rolling bearing to obtain the condition assessment index; the service condition of the rolling bearing to be assessed is evaluated based on the condition assessment index.

[0125] Specifically, through calculation formula

[0126]

[0127] in, Indicates the status assessment index, , This indicates the length of time the bearing has been in service;

[0128] according to Assess the service condition of the rolling bearing to be evaluated; where, when When this occurs, it indicates that the rolling bearing to be evaluated is unused and in excellent service condition. When this occurs, it indicates that the rolling bearing is damaged and cannot continue to be used.

[0129] Figure 5 The service status of the target rolling bearing is shown at different monitoring points. From Figure 5 As can be seen, the performance of the target rolling bearing gradually declines as its service time increases. This also demonstrates the effectiveness of this invention in evaluating the service condition of rolling bearings.

[0130] In summary, the rolling bearing condition assessment method disclosed in this invention is based on the surface area defect law. It extracts the bearing degradation feature set from full-life vibration signal data. The obtained degradation features include local fluctuation features and degradation trend features. On the one hand, relying on the surface area defect law can more accurately describe the bearing degradation mechanism. On the other hand, extracting the bearing degradation features from full-life vibration signal data overcomes the shortcomings of traditional methods that ignore historical operating data under normal service conditions of rolling bearings. Furthermore, the degradation features integrate local fluctuation features and degradation trend features, improving the interpretability of the degradation features and making them more consistent with the physical characteristics of rolling bearings in actual working conditions, thus improving the accuracy of degradation features and improving the accuracy of rolling bearing condition assessment. Based on the first arrival time and the predicted degradation amount, the remaining life prediction value is calculated. Finally, based on the life prediction value, the service condition of the rolling bearing to be assessed is evaluated. It can flexibly adjust the judgment threshold of the remaining life prediction according to the difference in the remaining life of the bearing under different working conditions, thereby improving the accuracy of rolling bearing condition assessment.

[0131] The rolling bearing condition assessment system disclosed in this invention will be described below.

[0132] like Figure 6 As shown, the rolling bearing condition assessment system 600 includes:

[0133] The degradation prediction module 601 is used to input the vibration signal of the rolling bearing to be evaluated into a pre-trained bearing degradation model to obtain a predicted degradation value of the rolling bearing to be evaluated; wherein, the bearing degradation model is trained based on a degradation feature set, which is extracted from the whole life vibration signal data based on the surface area defect law, and the degradation feature set includes at least one degradation feature, including local fluctuation features and degradation trend features; the life prediction module 602 is used to calculate the predicted remaining life of the rolling bearing to be evaluated based on the first arrival time and the predicted degradation value; wherein, the first arrival time is the moment when the state value first exceeds a threshold.

[0134] The condition assessment module 603 is used to assess the service condition of the rolling bearing to be assessed based on the remaining service life prediction value, and obtain the condition assessment result of the rolling bearing to be assessed.

[0135] The rolling bearing condition assessment system 600 also includes a model training module 604, which is used to train a bearing degradation model in the following manner:

[0136] The data acquisition module acquires full-life vibration signal data from multiple bearings of the same model as the rolling bearing to be evaluated.

[0137] Based on the surface area defect law, the feature extraction module extracts the bearing degradation feature set from the full-life vibration signal data. The degradation feature set is used to characterize the relationship between bearing service life and bearing service time. The degradation feature set includes at least one degradation feature, which includes local fluctuation features and degradation trend features.

[0138] The model training module constructs a bearing degradation model to predict the amount of bearing degradation based on the degradation feature set, and trains the bearing degradation model using full-life vibration signal data until the loss value of the bearing degradation model is less than a preset loss threshold, thus obtaining the trained bearing degradation model.

[0139] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. Their specific functions and technical effects can be found in the method embodiments section, and will not be repeated here. Those skilled in the art will understand that, for ease of description and brevity, the division of the above functional units and modules is only used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments 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. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0140] like Figure 7 As shown, embodiments of the present invention provide a terminal device, such as... Figure 7 As shown, the terminal device D10 of this embodiment includes: at least one processor D100 ( Figure 7The diagram shows only one processor, a memory D101, and a computer program D102 stored in the memory D101 and executable on the at least one processor D100, wherein the processor D100 executes the computer program D102 to implement the steps in any of the above method embodiments.

[0141] Specifically, when the processor D100 executes the computer program D102, it acquires full-life vibration signal data of multiple bearings of the same model as the rolling bearing to be evaluated; based on the surface area defect law, it extracts the bearing degradation feature set from the full-life vibration signal data; based on the degradation feature set, it constructs a bearing degradation model to predict the amount of bearing degradation, and trains the bearing degradation model using the full-life vibration signal data until the loss value of the bearing degradation model is less than a preset loss threshold, thus obtaining the trained bearing degradation model; it acquires the vibration signal of the rolling bearing to be evaluated, inputs the vibration signal into the trained bearing degradation model, obtains the predicted value of the degradation amount of the rolling bearing to be evaluated, and calculates the predicted value of the remaining life of the rolling bearing to be evaluated based on the first arrival time and the predicted value of the degradation amount; based on the life prediction value, it evaluates the service status of the rolling bearing to be evaluated, thus obtaining the status evaluation result of the rolling bearing to be evaluated. Based on the surface area defect law, a set of bearing degradation features is extracted from full-life vibration signal data. The obtained degradation features include local fluctuation features and degradation trend features. On the one hand, relying on the surface area defect law can more accurately describe the degradation mechanism of the bearing. On the other hand, extracting the bearing degradation features from full-life vibration signal data overcomes the shortcomings of traditional methods that ignore historical operating data under normal service conditions of rolling bearings. Furthermore, the degradation features integrate local fluctuation features and degradation trend features, improving the interpretability of the degradation features and making them more consistent with the physical characteristics of rolling bearings in actual working conditions, thus improving the accuracy of degradation features and thus improving the accuracy of rolling bearing condition assessment. Based on the first arrival time and the predicted value of degradation amount, the remaining life prediction value is calculated. Finally, based on the life prediction value, the service condition of the rolling bearing to be evaluated is assessed. The judgment threshold of the remaining life prediction can be flexibly adjusted according to the difference in the remaining life of the bearing under different working conditions, thereby improving the accuracy of rolling bearing condition assessment.

[0142] The processor D100 can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0143] In some embodiments, the memory D101 may be an internal storage unit of the terminal device D10, such as a hard disk or memory of the terminal device D10. In other embodiments, the memory D101 may be an external storage device of the terminal device D10, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the terminal device D10. Furthermore, the memory D101 may include both internal and external storage units of the terminal device D10. The memory D101 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory D101 can also be used to temporarily store data that has been output or will be output.

[0144] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0145] This application provides a computer program product that, when run on a terminal device, enables the terminal device to implement the steps described in the various method embodiments above.

[0146] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of protection of this application is limited to these examples; within the framework of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of one or more embodiments of this application as described above, which are not provided in detail for the sake of brevity.

[0147] One or more embodiments in this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments in this application should be included within the protection scope of this application.

Claims

1. A method for assessing the condition of a rolling bearing, characterized in that, include: The vibration signal of the rolling bearing to be evaluated is input into a pre-trained bearing degradation model to obtain the predicted value of the degradation amount of the rolling bearing to be evaluated; wherein, the bearing degradation model is trained based on a degradation feature set, the degradation feature set is extracted from the whole life vibration signal data based on the surface area defect law, and the degradation feature set includes at least one degradation feature, the degradation feature includes local fluctuation feature and degradation trend feature; Based on the first arrival time and the predicted degradation value, the predicted remaining life of the rolling bearing to be evaluated is calculated; wherein, the first arrival time is the moment when the state value first exceeds the threshold. Based on the remaining life prediction value, the service status of the rolling bearing to be evaluated is assessed to obtain the status assessment result of the rolling bearing to be evaluated; The expression for the degradation feature is: in, Indicates the first Degradation characteristics of individual bearings, Indicates the first The service life of each bearing. Indicates the first The first bearing Each service period, Indicates characteristics of degradation trend. A scaling factor is used to control the intensity of noise. It is a fixed parameter that is related to the properties of the material. This indicates noise and is used to describe local fluctuations during bearing degradation. The calculation of the predicted remaining life of the rolling bearing to be evaluated based on the first arrival time and the predicted degradation amount is specifically as follows: Determine the given threshold for failure prediction according to industry standards or relative methods. , This indicates that the bearing, calculated based on the first arrival time, has exceeded a given threshold from its current state. The required time is determined by estimating all unknown parameters in the probability density function using the predicted degradation value, and then using the remaining service life density function of the rolling bearing. Calculate the probability that the remaining life of the rolling bearing is in the range [0, +∞), using... and Integrating the product over [0, +∞) yields... Predicted remaining life of rolling bearings at all times .

2. The rolling bearing condition assessment method according to claim 1, characterized in that, The bearing degradation model is trained in the following manner: Acquire full-life vibration signal data from multiple bearings of the same model as the rolling bearing to be evaluated; Based on the surface area defect law, the degradation feature set of the bearing is extracted from the full-life vibration signal data; the degradation feature set is used to characterize the correlation between the bearing's service life and its service time. Based on the degradation feature set, a bearing degradation model is constructed to predict the amount of bearing degradation. The bearing degradation model is then trained using the full-life vibration signal data until the loss value of the bearing degradation model is less than a preset loss threshold, thus obtaining the trained bearing degradation model.

3. The rolling bearing condition assessment method according to claim 1, characterized in that, The expression for the bearing degradation model is: in, Indicates the bearing's service life. The amount of degradation at that time This indicates the initial state of the bearing. The drift coefficient represents the difference between bearings and follows a normal distribution, i.e. , It is a fixed parameter that is related to the properties of the material. The diffusion coefficient is used to describe the degree of fluctuation during bearing degradation. This represents standard Brownian motion, used to represent the inherent variability of random degradation processes over time.

4. The rolling bearing condition assessment method according to claim 1, characterized in that, The and The calculation is performed as follows: Through calculation formula get Density function of remaining service life of rolling bearings ;in, This indicates that the bearing, calculated based on the first arrival time, has exceeded a given threshold from its current state. The time required , It is a value greater than 0, in order to ensure that the rolling bearing is in The failure point has not yet been reached. This indicates a pre-defined threshold. express The state value at time t, where c represents the diffusion coefficient. express variance express The mean, , , , It is a fixed parameter that is related to the properties of the material; Through calculation formula get Predicted remaining life of rolling bearings at all times .

5. The rolling bearing condition assessment method according to claim 1, characterized in that, The step of evaluating the service condition of the rolling bearing to be evaluated based on the predicted life value specifically involves: The total lifespan of the rolling bearing is calculated by adding the predicted remaining lifespan value to the actual service time of the rolling bearing. The condition assessment index is obtained by dividing the predicted remaining life of the rolling bearing by the total life of the rolling bearing. The service condition of the rolling bearing to be evaluated is assessed based on the aforementioned condition assessment indicators.

6. The rolling bearing condition assessment method according to claim 5, characterized in that, The step of evaluating the service condition of the rolling bearing to be evaluated based on the condition evaluation index includes: When the status evaluation index is 0, the service status of the rolling bearing to be evaluated is determined to be unusable. When the condition evaluation index is 100%, the service condition of the rolling bearing to be evaluated is determined to be excellent.

7. The rolling bearing condition assessment method according to claim 5, characterized in that, The status assessment indicators It is obtained through the following calculation formula: in, Indicates the status assessment index, , This indicates the time the bearing has been in service.

8. The rolling bearing condition assessment method according to claim 1, characterized in that, The full-lifetime vibration signal data was obtained through accelerated degradation experiments.

9. A rolling bearing condition assessment system, characterized in that, include: The degradation prediction module is used to input the vibration signal of the rolling bearing to be evaluated into a pre-trained bearing degradation model to obtain the predicted value of the degradation of the rolling bearing to be evaluated; wherein, the bearing degradation model is trained based on a degradation feature set, the degradation feature set is extracted from the whole life vibration signal data based on the surface area defect law, and the degradation feature set includes at least one degradation feature, including local fluctuation feature and degradation trend feature; The life prediction module is used to calculate the remaining life prediction value of the rolling bearing to be evaluated based on the first arrival time and the predicted degradation value; wherein, the first arrival time is the moment when the state value first exceeds a threshold. The condition assessment module is used to assess the service condition of the rolling bearing to be assessed based on the remaining life prediction value, and obtain the condition assessment result of the rolling bearing to be assessed. The expression for the degradation feature is: in, Indicates the first Degradation characteristics of individual bearings, Indicates the first The service life of each bearing. Indicates the first The first bearing Each service period, Indicates characteristics of degradation trend. A scaling factor is used to control the intensity of noise. It is a fixed parameter that is related to the properties of the material. This indicates noise and is used to describe local fluctuations during bearing degradation. The calculation of the remaining life prediction value of the rolling bearing to be evaluated based on the first arrival time and the predicted degradation amount specifically involves: determining a given threshold for failure prediction according to industry standards or a relative method. , This indicates that the bearing, calculated based on the first arrival time, has exceeded a given threshold from its current state. The required time is determined by estimating all unknown parameters in the probability density function using the predicted degradation value, and then using the remaining service life density function of the rolling bearing. Calculate the probability that the remaining life of the rolling bearing is in the range [0, +∞). and Integrating the product over [0, +∞) yields... Predicted remaining life of rolling bearings at all times .

10. The rolling bearing condition assessment system according to claim 9, characterized in that, The system also includes a model training module, which is used to train the bearing degradation model in the following manner: The data acquisition module acquires full-life vibration signal data from multiple bearings of the same model as the rolling bearing to be evaluated. Based on the surface area defect law, the feature extraction module extracts the degradation feature set of the bearing from the full-life vibration signal data; the degradation feature set is used to characterize the correlation between the bearing's service life and its service time. The model training module constructs a bearing degradation model for predicting the amount of bearing degradation based on the degradation feature set, and trains the bearing degradation model using the full-life vibration signal data until the loss value of the bearing degradation model is less than a preset loss threshold, thus obtaining the trained bearing degradation model.

11. The rolling bearing condition assessment system according to claim 9, characterized in that, The expression for the bearing degradation model is: in, Indicates the bearing's service life. The amount of degradation at that time This indicates the initial state of the bearing. The drift coefficient represents the difference between bearings and follows a normal distribution, i.e. , It is a fixed parameter that is related to the properties of the material. The diffusion coefficient is used to describe the degree of fluctuation during bearing degradation. This represents standard Brownian motion, used to represent the inherent variability of random degradation processes over time.

12. The rolling bearing condition assessment system according to claim 9, characterized in that, The and The calculation is performed as follows: Through calculation formula get Density function of remaining service life of rolling bearings ;in, This indicates that the bearing, calculated based on the first arrival time, has exceeded a given threshold from its current state. The time required , It is a value greater than 0, in order to ensure that the rolling bearing is in The failure point has not yet been reached. This indicates a pre-defined threshold. express The state value at time t, where c represents the diffusion coefficient. express variance express The mean, , , , It is a fixed parameter that is related to the properties of the material; Through calculation formula get Predicted remaining life of rolling bearings at all times .

13. The rolling bearing condition assessment system according to claim 9, characterized in that, The step of evaluating the service status of the rolling bearing to be evaluated based on the predicted service life value specifically involves adding the predicted remaining service life value to the service time already served by the rolling bearing to obtain the total service life of the rolling bearing. The condition assessment index is obtained by dividing the predicted remaining life of the rolling bearing by the total life of the rolling bearing. The service condition of the rolling bearing to be evaluated is assessed based on the aforementioned condition assessment indicators.

14. The rolling bearing condition assessment system according to claim 13, characterized in that, The process of evaluating the service condition of the rolling bearing to be evaluated based on the aforementioned condition evaluation index specifically includes: When the status evaluation index is 0, the service status of the rolling bearing to be evaluated is determined to be unusable. When the condition evaluation index is 100%, the service condition of the rolling bearing to be evaluated is determined to be excellent.

15. The rolling bearing condition assessment system according to claim 13, characterized in that, The status assessment indicators The following calculation formula is used to obtain... in, Indicates the status assessment index, , This indicates the time the bearing has been in service.

16. The rolling bearing condition assessment system according to claim 9, characterized in that, The full-lifetime vibration signal data was obtained through accelerated degradation experiments.

17. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 8.

18. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 8.

Images