A bearing noise processing method for bearing fault diagnosis

Abstract

The present invention discloses a bearing noise processing method for bearing fault diagnosis, which relates to the technical field of data processing. The method includes: obtaining the target model of the target bearing, retrieving in the historical fault database based on the target model, and determining the historical bearing fault record set; determining a plurality of sets for distinguishing the historical bearing fault record set; obtaining a plurality of noise feature central values and a plurality of noise feature correlation bandwidths; obtaining a plurality of noise monitoring feature sets; performing cross-scale interaction fusion to obtain a target interaction fusion noise monitoring feature set, and using the target interaction fusion noise monitoring feature set as the bearing noise processing result. The present invention solves the technical problem in the prior art that the depth of bearing noise analysis is insufficient, resulting in low reliability of bearing fault diagnosis, and achieves the technical effect of deeply analyzing bearing noise data, accurately obtaining bearing noise data, and improving the accuracy of bearing fault diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a bearing noise processing method for bearing fault diagnosis. Background Art

[0002] Bearings are key components in many mechanical devices, and their fault diagnosis is crucial for the stable operation of the devices. Traditional bearing fault diagnosis methods mainly directly analyze the noise data, but the noise data is often interfered by the external environment, making it difficult to accurately extract the fault features. In recent years, bearing fault diagnosis methods based on neural network models have gradually been applied. Through the feature analysis of noise data, bearing faults can be detected earlier and more accurately. However, traditional noise processing methods usually use the same receptive field for feature extraction, ignoring the influence of different fault types on the noise signal features, resulting in low accuracy of fault diagnosis. Traditional bearing noise processing methods use single-scale feature extraction means, failing to fully consider the noise feature differences corresponding to different fault types, which easily causes the generation of redundant data and the loss of information.

[0003] The prior art has the technical problem that the depth of bearing noise analysis is insufficient, resulting in low reliability of bearing fault diagnosis. Summary of the Invention

[0004] The present application provides a bearing noise processing method for bearing fault diagnosis, which is used to solve the technical problem in the prior art that the depth of bearing noise analysis is insufficient, resulting in low reliability of bearing fault diagnosis.

[0005] In view of the above problems, the present application provides a bearing noise processing method for bearing fault diagnosis, and the method includes:

[0006] Obtain the target model of the target bearing, retrieve in the historical fault database based on the target model, and determine the historical bearing fault record set;

[0007] Perform heterogeneous discrimination on the historical bearing fault record set to determine multiple discriminated historical bearing fault record sets;

[0008] Traverse the multiple discriminated historical bearing fault record sets to perform centralized extraction of noise features, and obtain multiple noise feature centralized values and multiple noise feature correlation bandwidths;

[0009] Use the multiple noise feature correlation bandwidths as the multiple receptive fields of multiple noise processing network layers, and use the configured multiple noise processing network layers to perform multi-scale feature analysis on the noise monitoring data sequence of the target bearing within a preset monitoring window to obtain multiple noise monitoring feature sets;

[0010] Perform cross-scale interaction fusion on the multiple noise monitoring feature sets to obtain a target interaction fusion noise monitoring feature set, and use the target interaction fusion noise monitoring feature set as the bearing noise processing result.

[0011] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0012] In this application, the target model of the target bearing is obtained, retrieved based on the target model in the historical fault database to determine a set of historical bearing fault records, and then the set of historical bearing fault records is differentiated to determine multiple differentiated historical bearing fault record sets; traverse the multiple differentiated historical bearing fault record sets to extract noise features centrally, obtain multiple noise feature central values and multiple noise feature correlation bandwidths, and then use the multiple noise feature correlation bandwidths as the multiple receptive fields of multiple noise processing network layers, and use the configured multiple noise processing network layers to perform multi-scale feature analysis on the noise monitoring data sequence of the target bearing within a preset monitoring window to obtain multiple noise monitoring feature sets, and then perform cross-scale interaction fusion on the multiple noise monitoring feature sets to obtain a target interaction fusion noise monitoring feature set, and use the target interaction fusion noise monitoring feature set as the bearing noise processing result. It achieves the technical effects of ensuring the reliability of the bearing fault diagnosis result and improving the depth and accuracy of bearing noise data analysis. Description of the Drawings

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0014] Figure 1 It is a schematic flowchart of a bearing noise processing method for bearing fault diagnosis provided by an embodiment of this application;

[0015] Figure 2 It is a schematic flowchart of determining multiple noise feature central values in a bearing noise processing method for bearing fault diagnosis provided by an embodiment of this application. Detailed Embodiments

[0016] This application provides a bearing noise processing method for bearing fault diagnosis, which is used to solve the technical problem that the depth of bearing noise analysis in the prior art is insufficient, resulting in low reliability of bearing fault diagnosis.

[0017] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts shall fall within the protection scope of the present application.

[0018] It should be noted that the terms "include" and "have" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products, or devices.

[0019] Embodiments, such as Figure 1 As shown, the present application provides a bearing noise processing method for bearing fault diagnosis, wherein the method includes:

[0020] S100: Obtain the target model of the target bearing, retrieve in the historical fault database based on the target model, and determine the historical bearing fault record set;

[0021] In a possible embodiment, the target bearing refers to the bearing that needs to be fault diagnosed, usually the bearing used in a specific device. The target model refers to the specific model information of the target bearing, including the specifications, types, design parameters, etc. of the bearing. These information play an important role in retrieving historical fault data because different models of bearings may have different fault modes and fault characteristics. The historical fault database is a data set containing various bearing models and their historical fault records. Among them, the data in the database usually includes information such as bearing fault types, occurrence times, fault modes, noise data, etc.

[0022] Furthermore, using the target model as an index, retrieve in the historical fault database, and obtain the historical bearing fault record set according to the retrieval result. The historical bearing fault record combination refers to the fault records related to the bearings of this model found in the historical fault database. This set contains various fault instances that have occurred to the bearings of this model in the past. By obtaining the historical bearing fault record set, the technical effect of providing data support for in-depth analysis of bearing noise data in the follow-up is achieved.

[0023] S200: Perform heterogeneous differentiation on the historical bearing fault record set to determine multiple differentiated historical bearing fault record sets;

[0024] Further, perform outlier discrimination on the historical bearing fault record set to determine multiple discriminated historical bearing fault record sets. Step S200 of the embodiment of the present application further includes:

[0025] Randomly extract multiple historical bearing fault records from the historical bearing fault record set;

[0026] Enumerate the multiple historical bearing fault records pairwise to obtain multiple enumeration combinations;

[0027] Determine whether there is an enumeration combination among the multiple enumeration combinations whose record similarity is greater than the preset record similarity threshold. If not, use the multiple historical bearing fault records as multiple discrimination targets;

[0028] Based on the multiple discrimination targets, perform outlier discrimination on the historical bearing fault record set according to the preset record similarity threshold to obtain the multiple discriminated historical bearing fault record sets, where each discriminated historical bearing fault record set corresponds to a discrimination target.

[0029] In one embodiment, by classifying different fault data in the historical bearing fault record set, fault modes with significant differences are identified, and record sets belonging to different fault types are distinguished, thereby realizing outlier discrimination of the historical bearing fault record set. This process helps to ensure that during in-depth analysis of noise data, data of specific fault types can be focused on separately.

[0030] By performing outlier discrimination on the historical bearing fault record set, the multiple discriminated historical bearing fault record sets are obtained. Among them, each discriminated historical bearing fault record set refers to the record set obtained through outlier discrimination, and each set contains historical bearing fault records with similar fault characteristics, representing the bearing records of one fault type.

[0031] Optionally, randomly extract multiple historical bearing fault records from the historical fault record set and form multiple enumeration combinations by pairwise enumeration. Then, calculate the similarity between each pair of enumeration combinations to obtain the record similarity. Among them, the record similarity is a metric value used to measure the similarity between two historical fault record combinations and can be calculated through the cosine similarity calculation formula.

[0032] If the similarity of a pair of combinations exceeds a preset record similarity threshold (a value preset by those skilled in the art, used to determine whether two historical fault record combinations are similar enough to be classified into the same category), then it is considered that these records represent faults of the same type. At this time, multiple randomly selected historical bearing fault records cannot be used as discrimination targets, and multiple historical bearing fault records need to be randomly selected again. If not, multiple randomly selected historical bearing fault records belong to different types of faults respectively, and then these records are used as multiple discrimination targets respectively to discriminate the historical bearing fault record set.

[0033] Next, based on these discrimination targets, further outlier discrimination is performed on all historical fault records according to the preset record similarity threshold. That is to say, the historical bearing fault records in the historical bearing fault record set are respectively used to calculate the record similarity with the multiple discrimination targets by using the cosine similarity calculation formula, and added to the set corresponding to the discrimination target whose record similarity calculation result is greater than or equal to the preset record similarity threshold, so as to obtain the multiple discriminated historical bearing fault record sets. Optionally, when there is any historical bearing fault record whose record similarity calculation results with the records of more than two discrimination targets are all greater than or equal to the preset record similarity threshold, it is added to the set of the discrimination target corresponding to the maximum value of the record similarity calculation result. Each discriminated historical bearing fault record set in the multiple discriminated historical bearing fault record sets represents a different fault type or fault mode.

[0034] Through the outlier discrimination of historical fault records, the interference of noise data and redundant information can be effectively reduced, providing more accurate training data for subsequent noise feature analysis and fault diagnosis. Outlier discrimination helps to ensure that the data used is more consistent when analyzing a specific fault type, improving the accuracy and reliability of diagnosis.

[0035] S300: Traverse the multiple discriminated historical bearing fault record sets to perform centralized extraction of noise features, obtaining multiple noise feature central values and multiple noise feature associated bandwidths;

[0036] Furthermore, when traversing the multiple discriminated historical bearing fault record sets to perform centralized extraction of noise features, obtaining multiple noise feature central values and multiple noise feature associated bandwidths, step S300 of this application embodiment further includes:

[0037] Traverse the multiple discriminated historical bearing fault record sets to extract noise monitoring data, obtaining multiple sets of discriminated historical bearing noise monitoring data sequences;

[0038] Extract instantaneous noise monitoring data from each of the multiple sets of differentiated historical bearing noise monitoring data sequences to determine multiple sets of historical instantaneous noise monitoring data, where each historical instantaneous noise monitoring data is the monitoring data at the moment of the largest data change fluctuation in each differentiated historical bearing noise monitoring data sequence;

[0039] Extract features from the multiple sets of historical instantaneous noise monitoring data to obtain multiple sets of historical instantaneous noise features, and conduct centralized analysis on the multiple sets of historical instantaneous noise features to determine multiple noise feature centralized values;

[0040] Using the multiple sets of historical instantaneous noise features as indexes, conduct feature - associated bandwidth diffusion identification on the multiple sets of differentiated historical bearing noise monitoring data sequences to determine multiple sets of historical instantaneous noise feature - associated bandwidths, where each historical instantaneous noise feature - associated bandwidth reflects the duration of the data associated with the historical instantaneous noise features at the bearing fault occurrence moment in a differentiated historical bearing noise monitoring data sequence;

[0041] Calculate the mean values of the multiple sets of historical instantaneous noise feature - associated bandwidths respectively to obtain the multiple noise feature - associated bandwidths.

[0042] In the embodiments of the present application, conduct centralized extraction of noise features for each of the multiple sets of differentiated historical bearing fault records to determine the most representative noise feature centralized values, thereby obtaining the multiple noise feature centralized values. And conduct feature - associated analysis based on the data in the multiple sets of differentiated historical bearing fault records to determine the multiple noise feature - associated bandwidths. Among them, the noise feature - associated bandwidth is the time period associated with the noise features when a bearing fault occurs.

[0043] Optionally, extract noise monitoring data from multiple sets of differentiated historical bearing fault records to obtain the multiple sets of differentiated historical bearing noise monitoring data sequences. These data record the noise change conditions of the bearing in different states in a time - series manner. Furthermore, conduct monitoring data analysis on each noise monitoring data sequence to obtain the data with large instantaneous changes at these historical time points, thereby obtaining the multiple sets of historical instantaneous noise monitoring data. Where each historical instantaneous noise monitoring data is the monitoring data at the moment of the largest data change fluctuation in each differentiated historical bearing noise monitoring data sequence.

[0044] Feature extraction is performed on multiple historical instantaneous noise monitoring data sets, converting the noise data into a set of representative numerical values to capture the main change characteristics of the noise data (including time-domain characteristics, frequency-domain characteristics, instantaneous amplitude, instantaneous phase, and instantaneous frequency, etc.), thereby obtaining the multiple historical instantaneous noise feature sets. By centrally analyzing the multiple historical instantaneous noise feature sets, analyzing the representative feature central values of each historical instantaneous noise feature set, the multiple noise feature central values are obtained. Among them, the multiple noise feature central values represent the main characteristics of the noise corresponding to different fault types.

[0045] On this basis, using the historical instantaneous noise feature set as an index, further feature correlation bandwidth diffusion identification is performed on multiple sets of differentiated historical bearing noise monitoring data sequences. The goal of this process is to identify the diffusion range of the noise signal in different frequency bands, that is, to determine the noise feature correlation bandwidth, and obtain the multiple historical instantaneous noise feature correlation bandwidth sets. Among them, each historical instantaneous noise feature correlation bandwidth reflects the duration of the data associated with the noise characteristics at the fault occurrence moment in a set of differentiated historical bearing noise monitoring data sequences.

[0046] Furthermore, calculate the mean value of each historical instantaneous noise feature correlation bandwidth set, and finally obtain multiple noise feature correlation bandwidths. These bandwidth values can reflect the noise spectrum characteristics of the bearing under different working conditions, and thus provide valuable data support for the processing of bearing noise data.

[0047] This process is crucial for the entire bearing fault diagnosis system. Because by analyzing the noise feature central values and the noise feature correlation bandwidth, the system can understand the noise characteristics of the bearing under different fault states from multiple dimensions, helping to accurately distinguish different fault types, and thus providing a more reliable basis for subsequent diagnosis.

[0048] Furthermore, as Figure 2 shown, performing instantaneous feature extraction on the multiple historical instantaneous noise monitoring data sets to obtain multiple historical instantaneous noise feature sets, and centrally analyzing the multiple historical instantaneous noise feature sets to determine multiple noise feature central values, step S300 of the embodiment of the present application further includes:

[0049] Using the noise feature extraction network layer to perform instantaneous feature extraction on the multiple historical instantaneous noise monitoring data sets to obtain the multiple historical instantaneous noise feature sets;

[0050] Traverse the multiple historical instantaneous noise feature sets to calculate the feature means, and determine multiple historical instantaneous noise feature means;

[0051] Iterate over the multiple historical instantaneous noise feature means according to a preset iteration bandwidth in the multiple historical instantaneous noise feature sets to obtain multiple iterated historical instantaneous noise features;

[0052] When the aggregation amount of the multiple iterated historical instantaneous noise features is less than or equal to the aggregation amount of the multiple historical instantaneous noise feature means, take the multiple historical instantaneous noise feature means as the multiple noise feature central values.

[0053] Further, step S300 of the embodiment of the present application further includes:

[0054] When the aggregation amount of the multiple iterated historical instantaneous noise features is greater than the aggregation amount of the multiple historical instantaneous noise feature means, determine whether the difference in the aggregation amounts between the multiple iterated historical instantaneous noise features and the multiple historical instantaneous noise feature means is greater than or equal to a preset aggregation amount difference threshold. If so, continue to iterate based on the multiple iterated historical instantaneous noise features until the maximum number of iterations is satisfied, and take the multiple iterated historical instantaneous noise features obtained in the last iteration as the multiple noise feature central values;

[0055] If not, stop the iteration and take the multiple iterated historical instantaneous noise features as the multiple noise feature central values.

[0056] In a possible embodiment, obtain multiple sample noise monitoring data and multiple sample noise features as training data, and use the training data to perform supervised training on a framework constructed based on a feedforward neural network to learn the one-to-one mapping relationship between the noise monitoring data and the noise features until the training satisfies the maximum number of training times (such as 500 times, 600 times, etc.) to obtain the trained noise feature extraction network layer.

[0057] Optionally, use the noise feature extraction network layer to perform instantaneous feature analysis on the multiple historical instantaneous noise monitoring data sets to obtain the multiple historical instantaneous noise feature sets. Among them, each historical instantaneous noise feature reflects the instantaneous information of the corresponding historical instantaneous noise monitoring data change, including features such as instantaneous frequency, instantaneous amplitude, and instantaneous phase.

[0058] Furthermore, calculate the feature means for the multiple historical instantaneous noise feature sets respectively to determine multiple historical instantaneous noise feature means. Among them, the multiple historical instantaneous noise feature means reflect the average feature situation of the multiple historical instantaneous noise feature sets considering accidental values and error values.

[0059] Iterate over the multiple historical instantaneous noise feature means in the multiple historical instantaneous noise feature sets according to a preset iteration bandwidth. That is, obtain multiple iterated historical instantaneous noise features according to the moving amplitude set by the preset iteration bandwidth (the moving amplitude set by those skilled in the art during a single iteration, which is the difference between historical instantaneous noise features during a single move).

[0060] In one embodiment, construct multiple central regions centered on the multiple historical instantaneous noise feature means with the preset iteration bandwidth as the radius, and count the number of historical instantaneous noise features included in the multiple central regions to obtain the aggregation amount of the multiple historical instantaneous noise feature means. Among them, the aggregation amount of the multiple historical instantaneous noise feature means reflects the number of historical instantaneous noise features aggregated around the multiple historical instantaneous noise feature means. The more the number of aggregated features, the more representative the corresponding historical instantaneous noise feature mean is of the feature situation of the multiple historical instantaneous noise feature sets. Based on the same principle of obtaining the aggregation amount of the multiple historical instantaneous noise feature means, calculate the aggregation amount of the multiple iterated historical instantaneous noise features.

[0061] When the aggregation amount of the multiple iterated historical instantaneous noise features is less than or equal to the aggregation amount of the multiple historical instantaneous noise feature means, it indicates that the multiple historical instantaneous noise feature means are more representative of the multiple historical instantaneous noise feature sets. Therefore, use the multiple historical instantaneous noise feature means as the multiple noise feature central values.

[0062] When the aggregation amount of the multiple iterated historical instantaneous noise features is greater than the aggregation amount of the multiple historical instantaneous noise feature means, it indicates that the multiple iterated historical instantaneous noise features are more representative of the feature situation of the multiple historical instantaneous noise feature sets than the multiple historical instantaneous noise feature means. At this time, further judgment is required to determine whether to continue the iteration.

[0063] Optionally, determine whether the difference between the aggregation amounts of the multiple iterated historical instantaneous noise features and the multiple historical instantaneous noise feature means is greater than or equal to a preset aggregation amount difference threshold. If so, it indicates that although the multiple iterated historical instantaneous noise features are more representative of the feature situation of the multiple historical instantaneous noise feature sets than the multiple historical instantaneous noise feature means and to a greater extent, then continue to iterate based on the multiple iterated historical instantaneous noise features until the maximum number of iterations (set by those skilled in the art, such as 30 times, 45 times, etc.) is reached, and use the multiple iterated historical instantaneous noise features obtained in the last iteration as the multiple noise feature central values.

[0064] Otherwise, it indicates that although the multiple iterative historical instantaneous noise features can better represent the characteristics of the multiple historical instantaneous noise feature set compared to the mean of the multiple historical instantaneous noise features, the difference between the two is not significant, and they have reached a level where they can relatively well represent the multiple historical instantaneous noise feature set. At this time, the iteration is stopped, and the multiple iterative historical instantaneous noise features are used as the median of the multiple noise feature sets.

[0065] Further, using the multiple historical instantaneous noise feature set as an index, perform feature-associated bandwidth diffusion identification on the multiple differentiated historical bearing noise monitoring data sequence sets to determine multiple historical instantaneous noise feature-associated bandwidth sets. The step S300 of the embodiment of the present application further includes:

[0066] Use the noise feature extraction network layer to extract noise features from the multiple differentiated historical bearing noise monitoring data sequence sets to obtain multiple differentiated historical bearing noise feature sequence sets;

[0067] Randomly extract a historical instantaneous noise feature from the multiple historical instantaneous noise feature sets as the first historical instantaneous noise feature, and match the corresponding first differentiated historical bearing noise feature sequence from the multiple differentiated historical bearing noise feature sequence sets;

[0068] According to a preset approximate association scale, perform a nearest neighbor search for the first historical instantaneous noise feature in the first differentiated historical bearing noise feature sequence to obtain a first historical instantaneous noise feature neighborhood;

[0069] Statistically analyze the duration of the first historical instantaneous noise feature neighborhood, and use the statistical result as the first historical instantaneous noise feature-associated bandwidth;

[0070] According to a preset nearest neighbor association scale, perform feature-associated bandwidth diffusion identification on the multiple historical instantaneous noise feature sets in the corresponding multiple differentiated historical bearing noise monitoring data sequence sets to determine multiple historical instantaneous noise feature-associated bandwidth sets.

[0071] In an embodiment of the present application, first use the noise feature extraction network layer to analyze multiple differentiated historical bearing noise monitoring data sequences, extract the noise features of each data sequence, and obtain the multiple differentiated historical bearing noise feature sequence sets. Each differentiated historical bearing noise feature sequence contains the noise feature change situation of a bearing fault type generated during the operation of the bearing. Through the automatic feature learning ability of the deep learning network, the system can identify important noise features related to bearing faults.

[0072] First, randomly select a feature from multiple historical instantaneous noise feature sets as the first historical instantaneous noise feature, and search for the matching feature sequence in multiple distinguishable historical bearing noise feature sequence sets to obtain the first distinguishable historical bearing noise feature sequence.

[0073] In the first distinguishable historical bearing noise feature sequence found, the system will perform a nearest neighbor search on the first historical instantaneous noise feature according to a preset approximate association scale (the minimum similarity when features are associated, preset by those skilled in the art). Through the nearest neighbor search, the system can find multiple first distinguishable historical bearing noise features associated with the first historical instantaneous noise feature to obtain the neighborhood of the first historical instantaneous noise feature.

[0074] Optionally, calculate the similarity between the first historical instantaneous noise feature and the first distinguishable historical bearing noise features adjacent to it in the first distinguishable historical bearing noise feature sequence. When the calculation result meets the preset approximate association scale, add the adjacent first distinguishable historical bearing noise features to the neighborhood of the first historical instantaneous noise feature, and continue to analyze whether the similarity between the second adjacent first distinguishable historical bearing noise feature and the first historical instantaneous noise feature meets the preset nearest neighbor association scale until it does not meet, then stop the nearest neighbor search to obtain the neighborhood of the first historical instantaneous noise feature.

[0075] For the neighborhood of the first historical instantaneous noise feature found, count the time points at both ends of the neighborhood to obtain the duration of the neighborhood, and then use it as the association bandwidth of the first historical instantaneous noise feature.

[0076] Based on the same principle of obtaining the association bandwidth of the first historical instantaneous noise feature, perform feature association bandwidth diffusion identification on the multiple historical instantaneous noise feature sets according to the preset approximate association scale in the corresponding multiple distinguishable historical bearing noise monitoring data sequence sets to determine multiple historical instantaneous noise feature association bandwidth sets. This achieves the technical effect of providing a basis for subsequent multi-scale noise data processing.

[0077] S400: Use the multiple noise feature association bandwidths as the multiple receptive fields of multiple noise processing network layers, and use the configured multiple noise processing network layers to perform multi-scale feature analysis on the noise monitoring data sequence of the target bearing within a preset monitoring window to obtain multiple noise monitoring feature sets;

[0078] In one embodiment, multiple noise feature correlation bandwidths will be used as the receptive fields of multiple noise processing network layers. The receptive field refers to the input area that a neuron in a neural network can "perceive". It defines the data range that the network can process when extracting features. Using the noise feature correlation bandwidth as the receptive field can help the network extract the features of noise data at different scales. Each receptive field corresponds to a noise feature correlation bandwidth, meaning that the network will extract features for different noise features in different time periods or frequency bands.

[0079] The noise processing network layer is a hierarchical structure constructed by a deep learning network (such as a convolutional neural network), which is specifically used to process and analyze noise data. Each layer of the network analyzes the input noise data through different receptive fields and extracts features at different scales. The advantage of using multiple noise processing network layers is that the system can obtain different perspectives on the bearing state through feature extraction at different levels, such as short-term noise changes (high-frequency signals) and long-term changes (low-frequency signals).

[0080] Multi-scale feature analysis refers to analyzing the features of noise data at multiple scales (time scale or frequency scale). This is very important because the faults of bearings usually exhibit different noise patterns at different time scales. For example, high-frequency noise may indicate minor cracks or surface damage, while low-frequency noise may be related to the overall operating state of the bearing or large-scale damage. Through multi-scale feature analysis, the system can comprehensively consider signals at different time / frequency scales to more comprehensively diagnose bearing faults.

[0081] Within a preset monitoring window, a noise monitoring data sequence of the target bearing is collected, and the obtained noise monitoring data sequence will be input into multiple noise processing network layers. The input of the data sequence will be passed layer by layer, and feature extraction will be performed through each noise processing network layer. Each layer can process and identify noise patterns at different scales to obtain the multiple noise monitoring feature sets. These feature sets contain noise features extracted from different scales. Each noise monitoring feature set includes time-domain features such as amplitude, average value, peak value, root mean square value, etc.; frequency-domain features such as frequency components, spectral density, etc.; time-frequency domain features: features extracted by methods such as wavelet transform, which combine time-domain and frequency-domain information.

[0082] By using multiple noise processing network layers for multi-scale noise feature analysis, it is ensured that effective features of noise data are extracted at different scales and ranges. Thus, more comprehensive data analysis and processing of noise monitoring data are carried out, and the technical effect of improving the accuracy of bearing fault detection is achieved.

[0083] S500: Perform cross-scale interactive fusion on the multiple noise monitoring feature sets to obtain a target interactive fusion noise monitoring feature set, and use the target interactive fusion noise monitoring feature set as the bearing noise processing result.

[0084] Further, for performing cross-scale interactive fusion on the multiple noise monitoring feature sets to obtain a target interactive fusion noise monitoring feature set, step S500 of the embodiments of the present application further includes:

[0085] Randomly extract a first noise monitoring feature set and a second noise monitoring feature set from the multiple noise monitoring feature sets;

[0086] Calculate the similarity recognition between the first noise monitoring feature set and the second noise monitoring feature set to determine a first feature similarity set;

[0087] Perform normalization processing on the first feature similarity set to obtain a first feature similarity normalization value set;

[0088] Perform convolution calculation on the first feature similarity normalization value set and the second noise monitoring feature set to obtain a first interactive fusion noise monitoring feature set;

[0089] Randomly extract a third noise monitoring feature set from the multiple noise monitoring feature sets, and perform cross-scale interactive fusion on it and the first interactive fusion noise monitoring feature set to obtain a second interactive fusion noise monitoring feature set;

[0090] After multiple cross-scale interactive fusions until all the noise monitoring features in the multiple noise monitoring feature sets are fused, obtain the target interactive fusion noise monitoring feature set.

[0091] Further, step S500 of the embodiments of the present application further includes:

[0092] Obtain multiple sample feature similarity normalization value sets, multiple sample noise monitoring feature sets, and multiple sample interactive fusion noise monitoring feature sets as training data;

[0093] Use the training data to perform supervised training on the network layer constructed based on the convolutional neural network until the training converges to obtain a trained convolutional network layer;

[0094] Use the convolutional network layer to perform convolution calculation on the first feature similarity normalization value set and the second noise monitoring feature set to obtain the first interactive fusion noise monitoring feature set.

[0095] In one embodiment of the present application, multiple noise monitoring feature sets are cross-scale interactively fused to obtain a target interactively fused noise monitoring feature set. Cross-scale interactive fusion refers to fusing noise monitoring features from different scales (such as time scale, frequency scale, etc.). By fusing information from each scale, richer noise features can be obtained. This method can synthesize noise patterns at different scales and help the system make more accurate diagnoses.

[0096] In this step, multiple noise monitoring feature sets will be gradually fused into a target interactively fused noise monitoring feature set through operations such as similarity calculation, normalization, and convolution. Through multiple interactive fusions, features at different scales can be more comprehensively integrated, further improving the comprehensiveness and reliability of noise monitoring data processing and obtaining high-quality bearing noise processing results. By obtaining the bearing noise processing results, high-quality analysis data is provided for bearing fault diagnosis, thereby achieving the technical effects of improving the depth of bearing noise processing and providing data support for improving the reliability of bearing fault diagnosis.

[0097] In one embodiment, a first noise monitoring feature set and a second noise monitoring feature set are randomly extracted from the multiple noise monitoring feature sets. Furthermore, the cosine similarity calculation formula is used to calculate the similarity between the first noise monitoring feature set and the second noise monitoring feature set to determine a first feature similarity set. Among them, each first feature similarity reflects the similarity degree between the first noise monitoring feature and the second noise monitoring feature.

[0098] Optionally, the first feature similarity set is normalized by using the min-max normalization formula to obtain a first feature similarity normalization value set. The min-max normalization formula is X norm =(X-X min ) / (X max -X min ), where X norm is the normalized first feature similarity normalization value, X max is the maximum value in the first feature similarity set, and X min is the minimum value in the first feature similarity set.

[0099] Furthermore, a neural network is used to perform convolution calculation on the first feature similarity normalization value set and the second noise monitoring feature set to obtain a first interactively fused noise monitoring feature set. Convolution operation is a common technique in deep learning, which can extract local features in the input data and fuse the correlation information between different features.

[0100] Furthermore, a third noise monitoring feature set is randomly extracted from the multiple noise monitoring feature sets, and it is cross-scale interactively fused with the first interactively fused noise monitoring feature set to obtain a second interactively fused noise monitoring feature set, realizing the interactive fusion of features at different scales. After multiple cross-scale interactive fusions until all the noise monitoring features in the multiple noise monitoring feature sets are fused, the target interactively fused noise monitoring feature set is obtained. By repeatedly fusing different noise monitoring feature sets, the expression ability of the features is gradually enhanced. After each fusion, the generated new feature set will contain more comprehensive noise information.

[0101] For supervised learning, it is necessary to collect training data. Optionally, multiple sample feature similarity normalization value sets, multiple sample noise monitoring feature sets, and multiple sample interactively fused noise monitoring feature sets are obtained as training data.

[0102] Using the training data to perform supervised training on the network layer constructed based on the convolutional neural network, the complex relationships between different noise features can be learned, and the parameters of the convolutional network layer can be optimized until the network converges, obtaining an accurate feature fusion ability. Until the training converges, a trained convolutional network layer is obtained. Furthermore, using the trained convolutional network layer to perform convolutional calculations on the first feature similarity normalization value set and the second noise monitoring feature set, a more accurate interactively fused feature, that is, the first interactively fused noise monitoring feature set, is obtained. It achieves the technical effect of interactively fusing multi-scale noise monitoring features and deeply analyzing bearing noise data.

[0103] In summary, the embodiments of the present application at least have the following technical effects:

[0104] The present application obtains the target interactively fused noise monitoring feature set through heterogeneous discrimination and noise feature centralized extraction of the historical bearing fault record set, combined with multi-scale analysis and cross-scale interactive fusion, and uses the target interactively fused noise monitoring feature set as the bearing noise processing result. It achieves the technical effect of improving the accuracy of noise data analysis and providing reliable data for bearing fault diagnosis.

[0105] It should be noted that the above sequence of the embodiments of the present application is only for description and does not represent the advantages or disadvantages of the embodiments. And the above describes specific embodiments of this specification. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0106] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

[0107] This specification and the drawings are merely exemplary descriptions of the present application and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications therein.

Claims

1. A bearing noise processing method for bearing fault diagnosis, characterized in that: The method comprises: Obtaining a target model of a target bearing, searching a historical fault database based on the target model, and determining a set of historical bearing fault records; Performing heterogeneous differentiation on the historical bearing fault record set to determine a plurality of differentiated historical bearing fault record sets; Traversing the plurality of distinguished historical bearing fault record sets to perform concentrated noise feature extraction, and obtaining a plurality of noise feature concentrated values ​​and a plurality of noise feature associated bandwidths; The multiple noise feature association bandwidths are used as multiple receptive fields of multiple noise processing network layers, and the configured multiple noise processing network layers are used to perform multi-scale feature analysis on the noise monitoring data sequence of the target bearing within a preset monitoring window to obtain multiple noise monitoring feature sets; Performing cross-scale interactive fusion on the multiple noise monitoring feature sets to obtain a target interactive fusion noise monitoring feature set, and using the target interactive fusion noise monitoring feature set as a bearing noise processing result; The multiple sets of distinguished historical bearing fault records are traversed to extract noise features, and multiple noise feature concentration values ​​and multiple noise feature associated bandwidths are obtained, including: Traversing the plurality of distinguishing historical bearing fault record sets to extract noise monitoring data, and obtaining a plurality of distinguishing historical bearing noise monitoring data sequence sets; Extracting instantaneous noise monitoring data from the plurality of distinguished historical bearing noise monitoring data series respectively, and determining a plurality of historical instantaneous noise monitoring data sets, wherein each historical instantaneous noise monitoring data is monitoring data at a time when the maximum data change fluctuation occurs in each distinguished historical bearing noise monitoring data series; Extracting features from the plurality of historical instantaneous noise monitoring data sets to obtain a plurality of historical instantaneous noise feature sets, and centrally analyzing the plurality of historical instantaneous noise feature sets to determine a plurality of noise feature central values; Using the multiple historical instantaneous noise feature sets as indexes, feature-associated bandwidth diffusion identification is performed on the multiple sets of distinguished historical bearing noise monitoring data sequences to determine multiple sets of historical instantaneous noise feature-associated bandwidths, wherein each historical instantaneous noise feature-associated bandwidth reflects the duration of data associated with the historical instantaneous noise feature at the time of fault occurrence in a distinguished historical bearing noise monitoring data sequence; The mean values ​​of the plurality of historical instantaneous noise feature associated bandwidth sets are respectively calculated to obtain the plurality of noise feature associated bandwidths.

2. A bearing noise processing method for bearing fault diagnosis according to claim 1, characterized in that: Extracting instantaneous features from the plurality of historical instantaneous noise monitoring data sets to obtain a plurality of historical instantaneous noise feature sets, and centrally analyzing the plurality of historical instantaneous noise feature sets to determine a plurality of noise feature central values, including: Using a noise feature extraction network layer to extract instantaneous features from the plurality of historical instantaneous noise monitoring data sets, to obtain the plurality of historical instantaneous noise feature sets; Traversing multiple historical instantaneous noise feature sets to perform feature mean calculations and determine multiple historical instantaneous noise feature means; Iterating the mean values ​​of the multiple historical instantaneous noise features in the multiple historical instantaneous noise feature sets according to a preset iteration bandwidth to obtain multiple iterative historical instantaneous noise features; When the aggregation amount of the multiple iterative historical instantaneous noise features is less than or equal to the aggregation amount of the multiple historical instantaneous noise feature mean values, the multiple historical instantaneous noise feature mean values ​​are used as the centralized values ​​of the multiple noise features.

3. A bearing noise processing method for bearing fault diagnosis according to claim 2, characterized in that: include: When the aggregation amount of the multiple iterative historical instantaneous noise features is greater than the aggregation amount of the multiple historical instantaneous noise feature means, determine whether the aggregation amount difference between the multiple iterative historical instantaneous noise features and the multiple historical instantaneous noise feature means is greater than or equal to a preset aggregation amount difference threshold, and if so, continue to iterate based on the multiple iterative historical instantaneous noise features until the maximum number of iterations is met, and use the multiple iterative historical instantaneous noise features obtained in the last iteration as the multiple noise feature concentration values; If not, the iteration is stopped, and the multiple iterative historical instantaneous noise features are used as the centralized values ​​of the multiple noise features.

4. A bearing noise processing method for bearing fault diagnosis according to claim 3, characterized in that: Using the multiple historical instantaneous noise feature sets as indexes, performing feature-associated bandwidth diffusion identification on the multiple sets of distinguished historical bearing noise monitoring data sequences, and determining multiple historical instantaneous noise feature-associated bandwidth sets, including: Using the noise feature extraction network layer, the noise features of the plurality of distinguishing historical bearing noise monitoring data sequence sets are extracted to obtain a plurality of distinguishing historical bearing noise feature sequence sets; Randomly extracting a historical instantaneous noise feature from the plurality of historical instantaneous noise feature sets as a first historical instantaneous noise feature, and matching a corresponding first distinguishing historical bearing noise feature sequence from the plurality of distinguishing historical bearing noise feature sequence sets; According to a preset approximate correlation scale, a neighbor search is performed on the first historical instantaneous noise feature in the first distinguishing historical bearing noise feature sequence to obtain a first historical instantaneous noise feature neighborhood; Counting the duration of the first historical instantaneous noise feature neighborhood, and using the statistical result as the first historical instantaneous noise feature associated bandwidth; According to the preset nearest neighbor association scale, feature association bandwidth diffusion identification is performed on the multiple historical instantaneous noise feature sets in the corresponding multiple sets of distinguished historical bearing noise monitoring data sequences to determine multiple historical instantaneous noise feature association bandwidth sets.

5. A bearing noise processing method for bearing fault diagnosis according to claim 1, characterized in that: Performing cross-scale interactive fusion on the multiple noise monitoring feature sets to obtain a target interactive fusion noise monitoring feature set includes: randomly extracting a first noise monitoring feature set and a second noise monitoring feature set from the plurality of noise monitoring feature sets; Calculating the first noise monitoring feature set and the second noise monitoring feature set for similarity identification to determine a first feature similarity set; Normalizing the first feature similarity set to obtain a first feature similarity normalized value set; Performing convolution calculation on the first feature similarity normalization value set and the second noise monitoring feature set to obtain a first interactive fusion noise monitoring feature set; randomly extracting a third noise monitoring feature set from the multiple noise monitoring feature sets, and performing cross-scale interactive fusion on the third noise monitoring feature set with the first interactive fusion noise monitoring feature set to obtain a second interactive fusion noise monitoring feature set; After multiple cross-scale interactive fusions, until all noise monitoring features in the multiple noise monitoring feature sets are fused, the target interactive fusion noise monitoring feature set is obtained.

6. A bearing noise processing method for bearing fault diagnosis according to claim 5, characterized in that: include: Acquire multiple sample feature similarity normalized value sets, multiple sample noise monitoring feature sets, and multiple sample interactive fusion noise monitoring feature sets as training data; Use the training data to supervise the network layer built based on the convolutional neural network until the training converges to obtain a trained convolutional network layer; The convolutional network layer is used to perform convolution calculation on the first feature similarity normalization value set and the second noise monitoring feature set to obtain the first interactive fusion noise monitoring feature set.

7. A bearing noise processing method for bearing fault diagnosis according to claim 1, characterized in that: The historical bearing fault record set is heterogeneously distinguished to determine a plurality of distinguished historical bearing fault record sets, including: Randomly extracting a plurality of historical bearing fault records from the historical bearing fault record set; Enumerate the multiple historical bearing fault records in pairs to obtain multiple enumeration combinations; Determine whether there is an enumeration combination whose record similarity among the multiple enumeration combinations is greater than a preset record similarity threshold, and if not, use the multiple historical bearing fault records as multiple differentiation targets; Based on the multiple differentiation targets, the historical bearing fault record set is heterogeneously differentiated according to a preset record similarity threshold to obtain the multiple differentiated historical bearing fault record sets, wherein each differentiated historical bearing fault record set corresponds to a differentiation target.

Images