Best multiscale entropy can evaluate bearing degradation method, device, equipment and medium

Abstract

The application provides a best multi-scale entropy energy ratio bearing degradation evaluation method, device, equipment and medium, comprising: acquiring full life cycle data of a rolling bearing to be evaluated, performing feature extraction processing on the full life cycle data to generate a best multi-scale entropy energy ratio feature; calling a trained SVDD model to preprocess the best multi-scale entropy energy ratio feature to generate a degradation evaluation curve corresponding to the full life cycle data, wherein the historical full life cycle data of multiple groups of rolling bearings are used to construct a multi-scale weight entropy energy ratio, and the best weight parameter in the multi-scale weight entropy energy ratio is selected to train the SVDD model; and outputting a warning signal according to the degradation evaluation curve. The application aims to solve the problems that the bearing performance evaluation method based on data driving in the market cannot effectively reflect an early signal mutation, the stability in the early stage is insufficient, and the correlation change between characteristic values and output results of the degradation evaluation model under different data cannot be solved.

Description

Technical Field

[0001] This invention relates to the field of bearing degradation assessment technology, specifically to a method, apparatus, equipment, and medium for assessing bearing degradation with optimal multi-scale entropy-energy ratio. Background Technology

[0002] Rolling bearings are crucial components in mechanical engineering, widely used across various fields. During operation, external factors and shaft characteristics lead to various failure modes, many of which are not easily detected in their early stages. By the time these failures are identified, most rolling bearings have already suffered serious damage, causing further and more significant problems. Degradation assessment and early fault diagnosis of rolling bearings can effectively prevent these situations. In specific working environments, more innovative degradation assessment techniques can be employed to enhance the safety and performance of related equipment. The key to evaluating the effectiveness of a degradation assessment scheme lies in its ability to identify potential faults early, thereby effectively reducing losses. Currently, most rolling bearing degradation assessment methods are based on physical models and data-driven approaches. Due to the non-deterministic nature of rolling bearing states and the non-linearity of their changes in real-world environments, it is often difficult to establish accurate mathematical models. Therefore, data-driven degradation assessment methods have received more attention in recent years. Extensive research by scholars both domestically and internationally has shown that the effectiveness of degradation assessment varies depending on the specific degradation process. Therefore, finding a suitable method to extract appropriate features and combine them with a suitable degradation assessment model becomes crucial for rolling bearing degradation assessment.

[0003] A common data-driven bearing performance evaluation process in the market involves: first, collecting full lifecycle data of rolling bearings for feature extraction; then, constructing an effective degradation assessment model; inputting the features into the model for training and validation; and finally, selecting effective indicators to form degradation curves. Currently, time-domain and frequency-domain feature extraction are commonly used techniques in rolling bearings. However, simple single features carry limited information and cannot simultaneously account for early degradation and later fault progression. Furthermore, the extracted features cannot directly determine their correlation with the degradation assessment model's output, making it difficult to effectively combine them with the model to construct accurate degradation curves. Currently, there are methods applying Multi-Scale Permutation Entropy (MPE) to fault diagnosis. MPE can effectively address the problem of insufficient information carried by a single time-domain signal. However, MPE applied to fault diagnosis cannot effectively reflect early signal abrupt changes, and its application in degradation assessment leads to insufficient stability in the early stages. Another method proposes an entropy-energy ratio index based on permutation entropy (PE) and root mean square (RMS). This feature can effectively reflect the changes in faults in the early stages. However, since the proportional relationship between the magnitudes of RMS and PE values ​​cannot be determined, the RMS value has too much weight in the entropy-energy ratio feature, ignoring the information contained in the PE value. At the same time, this index still cannot solve the problem of the correlation changes between feature values ​​and the output results of degradation assessment models under different data.

[0004] In view of the above, this application is hereby submitted. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide an optimal multi-scale entropy-energy ratio bearing degradation assessment method, apparatus, equipment, and medium, which can effectively solve the problems of existing data-driven bearing performance assessment methods that cannot effectively reflect early signal abrupt changes, have insufficient stability in the early stage of assessment, and have an excessive weight of RMS value in the entropy-energy ratio feature due to the inability to determine the proportional relationship between RMS value and PE value, thus ignoring the information contained in PE value. At the same time, this index still cannot solve the problem of the correlation change between feature values ​​and degradation assessment model output results under different data.

[0006] This invention discloses an optimal multi-scale entropy-energy ratio bearing degradation evaluation method, comprising:

[0007] Acquire the full life cycle data of the rolling bearing to be evaluated, perform feature extraction processing on the full life cycle data, and generate the optimal multi-scale entropy-energy ratio feature;

[0008] The trained SVDD model is called to preprocess the optimal multi-scale entropy-energy ratio feature to generate a degradation evaluation curve corresponding to the full life cycle data. In this process, the multi-scale weighted entropy-energy ratio is constructed by acquiring historical full life cycle data of multiple sets of rolling bearings, and the optimal weight parameters in the multi-scale weighted entropy-energy ratio are selected to train the SVDD model.

[0009] Based on the degradation assessment curve, an early warning signal is output.

[0010] Preferably, before calling the trained SVDD model to preprocess the optimal multi-scale entropy-energy ratio feature, the method further includes:

[0011] Acquire historical full lifecycle data for multiple groups of rolling bearings;

[0012] Feature extraction processing is performed on the historical full life cycle data to generate multiple sets of multi-scale weighted entropy energy ratios;

[0013] Establish an original SVDD model, and train the multi-scale weight entropy ratio for each of the above-mentioned multi-scale weights to generate an SVDD training model.

[0014] The SVDD training model is invoked to test and train the historical full lifecycle data to generate optimal weight parameters.

[0015] Based on the optimal weight parameters and the SVDD training model, an SVDD model is generated.

[0016] Preferably, feature extraction processing is performed on the historical full lifecycle data to generate multiple sets of multi-scale weighted entropy-energy ratios, specifically:

[0017] Initialize the first parameter and the second parameter, wherein the value range of the first parameter is a = i / 10, i = 3, 4, ..., 30, and the value range of the second parameter is b = j / 10, j = 0, 1, ..., 30;

[0018] The historical full lifecycle data is coarse-grained to generate multiple sets of multi-scale sequences;

[0019] Calculate the RMS value and MPE value of each multi-scale sequence, and normalize the RMS value and MPE value respectively to generate multiple sets of RMS' values ​​and MPE' values.

[0020] The weighting form of the RMS value is rewritten according to the first parameter and the second parameter with different values, and each RMS' value. The formula for rewriting the weighting form of the RMS value is (RMS′+b)*a, where a is the first parameter and b is the second parameter.

[0021] Based on the RMS value after rewriting the weighted form and the MPE' value, multiple sets of multi-scale weighted entropy energy ratios are generated. The calculation formula for the multi-scale weighted entropy energy ratio is as follows: OMEER is the multi-scale weighted entropy-energy ratio.

[0022] Preferably, an SVDD original model is established, and the entropy ratio of each multi-scale weight is trained to generate an SVDD training model, specifically as follows:

[0023] Initialize the first parameter and the second parameter, wherein the value range of the first parameter is a = i / 10, i = 3, 4, ..., 30, and the value range of the second parameter is b = j / 10, j = 0, 1, ..., 30;

[0024] The multi-scale weight entropy ratios corresponding to the first and second parameters with different values ​​are respectively input into the original SVDD model for training. The distance from each multi-scale weight entropy ratio to the center of the hypersphere is calculated to generate the SVDD training model and the threshold line. The formula of the threshold line is Threshold = u + 3σ, where Threshold is the threshold line, u is the average distance from the multi-scale weight entropy ratio to the center of the hypersphere, and σ is the variance of the distance from the multi-scale weight entropy ratio to the center of the hypersphere.

[0025] Preferably, the SVDD training model is called to test and train the historical full lifecycle data to generate optimal weight parameters, specifically as follows:

[0026] The historical full lifecycle data is input into the SVDD training model for testing, and the number of points exceeding the threshold line is calculated.

[0027] When it is determined that the change of the number of values ​​exceeds the preset value and converges to a stable value, select all the first parameter and the second parameter that satisfy n = α, where n is the number of points that exceed the threshold line and α is the stable value obtained by convergence;

[0028] Based on the condition a,b = min(a+b), calculate all the first and second parameters that satisfy n = α to generate the optimal weight parameters, where min is the minimum value operation.

[0029] This invention also discloses an optimal multi-scale entropy-energy ratio bearing degradation evaluation device, comprising:

[0030] The data acquisition unit is used to acquire the full life cycle data of the rolling bearing to be evaluated, perform feature extraction processing on the full life cycle data, and generate the optimal multi-scale entropy-energy ratio feature.

[0031] The degradation assessment unit is used to call the trained SVDD model to preprocess the optimal multi-scale entropy-energy ratio feature and generate a degradation assessment curve corresponding to the full life cycle data. In this process, the multi-scale weighted entropy-energy ratio is constructed by acquiring historical full life cycle data of multiple sets of rolling bearings, and the optimal weight parameters in the multi-scale weighted entropy-energy ratio are selected to train the SVDD model.

[0032] The early warning unit is used to output an early warning signal based on the degradation assessment curve.

[0033] The present invention also discloses an optimal multiscale entropy-energy ratio bearing degradation assessment device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the optimal multiscale entropy-energy ratio bearing degradation assessment method as described above.

[0034] The present invention also discloses a readable storage medium, characterized in that it stores a computer program that can be executed by the processor of the device in which the storage medium is located, so as to implement the optimal multi-scale entropy ratio bearing degradation evaluation method as described in any of the above claims.

[0035] In summary, the optimal multi-scale entropy ratio bearing degradation assessment method, apparatus, equipment, and medium provided in this embodiment first extracts features from the original vibration signal to construct a multi-scale weighted entropy ratio. Then, the optimal weight parameters are calculated based on the multi-scale weighted entropy ratio. Based on the optimal parameters, optimal multi-scale entropy ratio fault features are formed. The optimal multi-scale entropy ratio is input into SVDD to train a model for degradation assessment. This multi-scale entropy ratio bearing degradation assessment method can identify rolling bearing performance degradation points early and provide timely warnings. Furthermore, this method effectively expresses the performance degradation assessment process. Simultaneously, this multi-scale entropy ratio bearing degradation assessment method proposes a new performance degradation characteristic index for rolling bearings, laying the foundation for the construction of performance degradation assessment models. This addresses the shortcomings of existing data-driven bearing performance evaluation methods, such as the inability to effectively reflect early signal mutations, insufficient stability in the early stages of evaluation, and the excessive weighting of the RMS value in the entropy-energy ratio feature due to the uncertain proportional relationship between the RMS and PE values, which neglects the information contained in the PE value. Furthermore, this indicator still cannot resolve the issue of changes in the correlation between feature values ​​and the output results of degradation evaluation models under different data. Attached Figure Description

[0036] Figure 1 This is a schematic flowchart of the optimal multi-scale entropy ratio bearing degradation evaluation method provided by the first aspect of the present invention.

[0037] Figure 2 This is a schematic flowchart of the optimal multi-scale entropy ratio bearing degradation evaluation method provided by the second aspect of the present invention.

[0038] Figure 3 This is a schematic flowchart of the optimal multi-scale entropy ratio bearing degradation evaluation method provided in the third aspect of the present invention.

[0039] Figure 4 This is a schematic diagram of the module of the optimal multi-scale entropy ratio bearing degradation evaluation device provided in the embodiments of the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to represent selected embodiments of the invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0042] This invention discloses a method, apparatus, equipment, and medium for evaluating bearing degradation with optimal multi-scale entropy energy ratio, which at least partially addresses the shortcomings of existing technologies.

[0043] Please see Figures 1 to 3 The first embodiment of the present invention provides an optimal multi-scale entropy-energy ratio bearing degradation assessment method, which can be executed by a bearing degradation assessment device (hereinafter referred to as the assessment device), specifically by one or more processors within the assessment device, to achieve the following steps:

[0044] In this embodiment, the evaluation device may be a user terminal device (such as a smartphone, smart computer, or other smart device), which can establish a communication connection with a server in the cloud to achieve data interaction.

[0045] S101, Obtain the full life cycle data of the rolling bearing to be evaluated, perform feature extraction processing on the full life cycle data, and generate the optimal multi-scale entropy-energy ratio feature;

[0046] S102, the trained SVDD model is called to preprocess the optimal multi-scale entropy-energy ratio feature to generate a degradation evaluation curve corresponding to the full life cycle data. In this process, the multi-scale weighted entropy-energy ratio is constructed by acquiring historical full life cycle data of multiple sets of rolling bearings, and the optimal weight parameters in the multi-scale weighted entropy-energy ratio are selected to train the SVDD model.

[0047] Specifically, in this embodiment, features are first extracted from historical full lifecycle data, i.e., the original vibration signal, to construct a multi-scale weighted entropy-energy ratio. The optimal weight parameters are then calculated based on the multi-scale weighted entropy-energy ratio, and the optimal multi-scale entropy-energy ratio fault features are formed based on these optimal parameters. The optimal multi-scale entropy-energy ratio is then input into SVDD to train the model for degradation assessment. This training process mainly consists of two parts: the construction of the multi-scale weighted entropy-energy ratio features and the selection of optimal weight parameters. The principle of constructing the multi-scale weighted entropy-energy ratio features is as follows: the multi-scale permutation entropy of the full lifecycle original vibration signal sequence is calculated, and the RMS value of the sequence is calculated based on the coarse-grained sequence. The RMS value is then converted into a weighted form, resulting in the multi-scale weighted entropy-energy ratio in the form of MPE / (RMS weight). At this point, the initial RMS weight is the original RMS value. The principle for selecting the optimal weight parameters is as follows: determine the range of weights in the multi-scale weight entropy ratio, set the step size within the determined range, obtain multiple multi-scale weight entropy ratios with different weights, use the multi-scale weight entropy ratio of normal data as the training data of SVDD, and at the same time use all normal data as test data. Calculate the number of sample points exceeding the threshold line through the test data, set the iteration stopping condition, and obtain the optimal multi-scale weight entropy ratio as the optimal multi-scale entropy ratio.

[0048] In one possible embodiment of the present invention, before calling the trained SVDD model to preprocess the optimal multi-scale entropy-energy ratio feature, the method further includes:

[0049] First, acquire historical full lifecycle data for multiple groups of rolling bearings;

[0050] Secondly, feature extraction processing is performed on the historical full lifecycle data to generate multiple sets of multi-scale weighted entropy-energy ratios, specifically:

[0051] Initialize the first parameter and the second parameter, wherein the value range of the first parameter is a = i / 10, i = 3, 4, ..., 30, and the value range of the second parameter is b = j / 10, j = 0, 1, ..., 30;

[0052] The historical full lifecycle data is coarse-grained to generate multiple sets of multi-scale sequences;

[0053] Calculate the RMS value and MPE value of each multi-scale sequence, and normalize the RMS value and MPE value respectively to generate multiple sets of RMS' values ​​and MPE' values.

[0054] The weighting form of the RMS value is rewritten according to the first parameter and the second parameter with different values, and each RMS' value. The formula for rewriting the weighting form of the RMS value is (RMS′+b)*a, where a is the first parameter and b is the second parameter.

[0055] Based on the RMS value after rewriting the weighted form and the MPE' value, multiple sets of multi-scale weighted entropy energy ratios are generated. The calculation formula for the multi-scale weighted entropy energy ratio is as follows: OMEER is the multi-scale weighted entropy-energy ratio.

[0056] Specifically, in this embodiment, the calculation process for constructing the multi-scale weighted entropy energy ratio feature is as follows: For a vibration signal time series x(i) of arbitrary length N, i = 1, 2, ..., N, the original vibration signal is first coarsened by setting a time scale factor τ, and the calculation process is as follows: Among them, y s () represents the coarsened sequence, s represents different time scales, and [] is the rounding symbol. The original sequence x(i), i = 1, 2, ..., N is coarsened into s new sequences with a length of [ / ]. After obtaining the multi-scale time series, its RMS and MPE values ​​at different time scales are calculated. After standardizing the RMS and MPE values ​​to 0-1, the multi-scale entropy-energy ratio is defined as follows: Where MEER is the multi-scale entropy-energy ratio, MPE' is the standardized MPE value, and RMS' is the standardized RMS value. To prevent the MPE from having too little effect due to a small RMS value, the RMS is modified to the following weighted form: (RMS′+b)*a. Since the RMS′*a form is simply multiplying the original MEER by a factor, it cannot effectively amplify signal abrupt changes. The RMS′+b form ignores the multiple relationship between the numerator and denominator. By setting reasonable parameters a and b, this linear term form can effectively amplify subtle changes in fault characteristics as the fault changes. Furthermore, compared to the polynomial form, the linear term form requires less computation time. The final multi-scale weighted entropy-energy ratio formula is shown below:

[0057] Next, an original SVDD model is established, and the entropy ratio of each multi-scale weight is trained to generate an SVDD training model, specifically as follows:

[0058] Initialize the first parameter and the second parameter, wherein the value range of the first parameter is a = i / 10, i = 3, 4, ..., 30, and the value range of the second parameter is b = j / 10, j = 0, 1, ..., 30;

[0059] The multi-scale weight entropy ratios corresponding to the first and second parameters with different values ​​are respectively input into the original SVDD model for training. The distance from each multi-scale weight entropy ratio to the center of the hypersphere is calculated to generate the SVDD training model and the threshold line. The formula of the threshold line is Threshold = u + 3σ, where Threshold is the threshold line, u is the average distance from the multi-scale weight entropy ratio to the center of the hypersphere, and σ is the variance of the distance from the multi-scale weight entropy ratio to the center of the hypersphere.

[0060] Specifically, in this embodiment, the initial conditions for parameters a and b are first set. When b = 0 and a = 1, the multi-scale weighted entropy ratio and the multi-scale entropy ratio are the same. Since RMS' is data after 0-1 standardization, its values ​​are all points within the interval [0,1]. If the value of b is too large, the influence of RMS in the denominator will be smaller. Similarly, if the value of a is too large, the weight of the numerator will be smaller, and if the value of a is too small, the weight of the numerator will be too large. The step size is set to 0.1. The value ranges of a and b are as follows: Next, parameter optimization is performed. The multi-scale weighted entropy ratio is calculated from the full life cycle data of the rolling bearing. The multi-scale weighted entropy ratio of all healthy samples is used as training data and input into the SVDD model. After calculation, the distance from each sample point to the center of the hypersphere is obtained. Based on the distance data, an early degradation threshold line is set as follows: Threshold = u + 3σ, where Threshold is the threshold line, u is the average distance from the multi-scale weighted entropy ratio to the center of the hypersphere, and σ is the variance of the distance from the multi-scale weighted entropy ratio to the center of the hypersphere.

[0061] Next, the SVDD training model is invoked to test and train the historical full-lifecycle data to generate optimal weight parameters, specifically as follows:

[0062] The historical full lifecycle data is input into the SVDD training model for testing, and the number of points exceeding the threshold line is calculated.

[0063] When it is determined that the change of the number of values ​​exceeds the preset value and converges to a stable value, select all the first parameter and the second parameter that satisfy n = α, where n is the number of points that exceed the threshold line and α is the stable value obtained by convergence;

[0064] Based on the condition a,b = min(a+b), calculate all the first and second parameters that satisfy n = α to generate the optimal weight parameters, where min is the minimum value operation.

[0065] Specifically, in this embodiment, healthy sample data is used as test data input into the trained SVDD model to calculate the number of sample points n exceeding the threshold. Finally, a termination condition is set based on the number of sample points exceeding the threshold. As the parameters increase, the influence of the denominator on the features becomes smaller and smaller, and the change in the value of n gradually stabilizes and converges to a stable value α. All a and b that satisfy n = α are selected, and the value conditions are set as follows: a,b = min(a + b), where min represents the minimum value operation. The a and b that satisfy the conditions are selected as the optimal parameters.

[0066] Finally, based on the optimal weight parameters and the SVDD training model, an SVDD model is generated.

[0067] S103, based on the degradation assessment curve, output an early warning signal.

[0068] In summary, the entire process of the multi-scale entropy ratio bearing degradation assessment method is as follows: collect full life cycle data; coarsely granulate the full life cycle data to construct a multi-scale sequence; calculate the RMS and MPE of the multi-scale sequence, and normalize them as RMS' value and MPE' value respectively; rewrite the RMS in weighted form; form a multi-scale weighted entropy ratio; initialize the conditions of the first parameter a and the second parameter b; input the multi-scale weighted entropy ratios of different values ​​of the first parameter a and the second parameter b under healthy data into the SVDD training model to obtain the hypersphere model and the threshold line; input the multi-scale weighted entropy ratios of different values ​​of the first parameter a and the second parameter b under healthy data into SVDD to test the number n of points exceeding the threshold line; when the change of the value of n gradually stabilizes, select the optimal parameter according to the set value conditions to form the optimal multi-scale entropy ratio; input the optimal multi-scale entropy ratio feature under the full life cycle into SVDD to calculate the distance between each point and the center of the hypersphere to form the degradation curve.

[0069] Unlike existing rolling bearing degradation assessment methods, the proposed multi-scale entropy-energy ratio bearing degradation assessment method is sensitive to impact failures based on permutation entropy, but its changes are not significant in early fault diagnosis. RMS, on the other hand, indicates degradation in early faults. As the fault severity increases, permutation entropy decreases, while the root mean square (RMS) value increases in the opposite direction. Combining permutation entropy with RMS to form the entropy-energy ratio index effectively solves the problem of insignificant changes in permutation entropy in early fault diagnosis. To address the insufficient effective information from single-dimensional features, a multi-scale calculation method is introduced, resulting in a new feature: the optimal multi-scale entropy-energy ratio. Due to the large difference between RMS and permutation entropy values, RMS dominates the entropy-energy ratio result. To address this issue, RMS is modified into a weighted form, resulting in a multi-scale weighted entropy-energy ratio. A degradation assessment model is then introduced to calculate the weight parameters to obtain the optimal multi-scale entropy-energy ratio, with appropriate adjustments made to the feature values ​​extracted for different datasets. This feature, applied to the SVDD degradation assessment model, effectively solves the problem of poor bearing degradation assessment results caused by feature selection.

[0070] Please see Figure 4 A second embodiment of the present invention provides an optimal multi-scale entropy-energy ratio bearing degradation evaluation device, comprising:

[0071] The data acquisition unit 201 is used to acquire the full life cycle data of the rolling bearing to be evaluated, perform feature extraction processing on the full life cycle data, and generate the optimal multi-scale entropy-energy ratio feature.

[0072] The degradation assessment unit 202 is used to call the trained SVDD model to preprocess the optimal multi-scale entropy-energy ratio feature and generate a degradation assessment curve corresponding to the full life cycle data. In this process, the multi-scale weighted entropy-energy ratio is constructed by acquiring historical full life cycle data of multiple sets of rolling bearings, and the optimal weight parameters in the multi-scale weighted entropy-energy ratio are selected to train the SVDD model.

[0073] The early warning unit 203 is used to output an early warning signal based on the degradation evaluation curve.

[0074] A third embodiment of the present invention provides an optimal multiscale entropy-energy ratio bearing degradation assessment device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the optimal multiscale entropy-energy ratio bearing degradation assessment method as described in any of the above embodiments.

[0075] A fourth embodiment of the present invention provides a readable storage medium, characterized in that it stores a computer program that can be executed by a processor of the device in which the storage medium is located, to implement the optimal multi-scale entropy ratio bearing degradation evaluation method as described in any of the above embodiments.

[0076] Exemplary examples show that the computer program described in the third and fourth embodiments of the present invention can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the device for evaluating the optimal multi-scale entropy-energy ratio bearing degradation. For example, the apparatus described in the second embodiment of the present invention.

[0077] The processor referred to can be a Central Processing Unit (CPU), or 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. The general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the optimal multi-scale entropy-energy ratio bearing degradation evaluation method, connecting various parts of the entire system through various interfaces and lines.

[0078] The memory can be used to store the computer program and / or modules. The processor, by running or executing the computer program and / or modules stored in the memory, and by calling the data stored in the memory, realizes various functions of the optimal multi-scale entropy energy ratio bearing degradation evaluation method. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, text conversion function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, text message data, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0079] If the implemented module is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0080] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0081] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions that fall within the scope of the present invention are within the scope of protection of the present invention.

Claims

1. An optimal multi-scale entropy-energy ratio bearing degradation assessment method, characterized in that, include: Acquire historical full lifecycle data for multiple groups of rolling bearings; Feature extraction processing is performed on the historical full lifecycle data to generate multiple sets of multi-scale weighted entropy-energy ratios, specifically: Initialize the first and second parameters, wherein the value range of the first parameter is: The range of values ​​for the second parameter is: ; The historical full lifecycle data is coarse-grained to generate multiple sets of multi-scale sequences; Calculate the RMS value and MPE value of each multi-scale sequence, and normalize the RMS value and MPE value respectively to generate multiple sets of RMS' values ​​and MPE' values. The weighted form of the RMS value is rewritten based on the first parameter and the second parameter with different values, and each RMS' value. The formula for rewriting the weighted form of the RMS value is as follows: , where a is the first parameter and b is the second parameter; Based on the RMS value after rewriting the weighted form and the MPE' value, multiple sets of multi-scale weighted entropy energy ratios are generated. The calculation formula for the multi-scale weighted entropy energy ratio is as follows: , The ratio of multi-scale weighted entropy to energy; Establish an original SVDD model, and train the multi-scale weight entropy ratio for each of the above-mentioned multi-scale weights to generate an SVDD training model. The SVDD training model is invoked to test and train the historical full lifecycle data, generating the optimal weight parameters in the multi-scale weighted entropy energy ratio. Generate an SVDD model based on the optimal weight parameters and the SVDD training model; Acquire the full life cycle data of the rolling bearing to be evaluated, perform feature extraction processing on the full life cycle data, and generate the best multi-scale entropy-energy ratio feature based on the optimal weight parameters; The trained SVDD model is invoked to preprocess the optimal multi-scale entropy-energy ratio feature to generate a degradation evaluation curve corresponding to the full life cycle data; Based on the degradation assessment curve, an early warning signal is output.

2. The optimal multi-scale entropy ratio bearing degradation evaluation method according to claim 1, characterized in that, A primitive SVDD model is established, and the entropy ratio of each multi-scale weight is trained to generate an SVDD training model, specifically as follows: The multi-scale weight entropy ratios corresponding to the first and second parameters with different values ​​are respectively input into the original SVDD model for training. The distance from each multi-scale weight entropy ratio to the center of the hypersphere is calculated, generating the SVDD training model and a threshold line. The formula for the threshold line is as follows: Threshold is the threshold line, and u is the average distance from the multi-scale weighted entropy energy ratio to the center of the hypersphere. Let V be the variance of the distance from the multi-scale weighted entropy energy ratio to the center of the hypersphere.

3. The optimal multi-scale entropy energy ratio bearing degradation evaluation method according to claim 2, characterized in that, The SVDD training model is invoked to test and train the historical full-lifecycle data to generate optimal weight parameters, specifically as follows: The historical full lifecycle data is input into the SVDD training model for testing, and the number of points exceeding the threshold line is calculated. When it is determined that the changes in the number of values ​​exceeding a certain threshold reach a preset value and converge to a stable value, all values ​​that satisfy this condition are selected. The first parameter and the second parameter, wherein, The number of points exceeding the threshold line. The stable value obtained upon convergence; According to the value conditions For all satisfactions The first parameter and the second parameter are calculated to generate the optimal weight parameters, where min is the minimum value operation.

4. An optimal multi-scale entropy-energy ratio bearing degradation evaluation device, characterized in that, include: Acquire historical full lifecycle data for multiple groups of rolling bearings; Feature extraction processing is performed on the historical full lifecycle data to generate multiple sets of multi-scale weighted entropy-energy ratios, specifically: Initialize the first and second parameters, wherein the value range of the first parameter is: The range of values ​​for the second parameter is: ; The historical full lifecycle data is coarse-grained to generate multiple sets of multi-scale sequences; Calculate the RMS value and MPE value of each multi-scale sequence, and normalize the RMS value and MPE value respectively to generate multiple sets of RMS' values ​​and MPE' values. The weighted form of the RMS value is rewritten based on the first parameter and the second parameter with different values, and each RMS' value. The formula for rewriting the weighted form of the RMS value is as follows: , where a is the first parameter and b is the second parameter; Based on the RMS value after rewriting the weighted form and the MPE' value, multiple sets of multi-scale weighted entropy energy ratios are generated. The calculation formula for the multi-scale weighted entropy energy ratio is as follows: , The ratio of multi-scale weighted entropy to energy; Establish an original SVDD model, and train the multi-scale weight entropy ratio for each of the above-mentioned multi-scale weights to generate an SVDD training model. The SVDD training model is invoked to test and train the historical full lifecycle data, generating the optimal weight parameters in the multi-scale weighted entropy energy ratio. Generate an SVDD model based on the optimal weight parameters and the SVDD training model; The data acquisition unit is used to acquire the full life cycle data of the rolling bearing to be evaluated, perform feature extraction processing on the full life cycle data, and generate the best multi-scale entropy-energy ratio feature according to the optimal weight parameters. The degradation assessment unit is used to call the trained SVDD model to preprocess the optimal multi-scale entropy-energy ratio feature and generate a degradation assessment curve corresponding to the full life cycle data. The early warning unit is used to output an early warning signal based on the degradation assessment curve.

5. An optimal multi-scale entropy-energy ratio bearing degradation assessment device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the optimal multi-scale entropy ratio bearing degradation assessment method as described in any one of claims 1 to 3.

6. A readable storage medium, characterized in that, The storage medium contains a computer program that can be executed by a processor of the device in which the storage medium resides, to implement the optimal multi-scale entropy ratio bearing degradation assessment method as described in any one of claims 1 to 3.

Images