A device failure determination method and apparatus

Abstract

The application provides a device fault determination method and device, and relates to the technical field of fault detection. The method performs open set fault recognition based on rotor position information of a magnetic suspension bearing in a device, can recognize fault types appearing in a training process, marks fault types not appearing in the training process as unknown faults, mines the similarity between unknown fault samples and known fault types, and improves the accuracy and practicability of a magnetic suspension bearing fault diagnosis system.

Description

Technical Field

[0001] This application relates to the field of fault detection technology, and more specifically, to a method and apparatus for determining equipment faults. Background Technology

[0002] In the fault diagnosis mechanism of large equipment, traditional methods usually focus on key components (such as bearings, gearboxes, etc.) and the monitoring sensors are placed near the component to be diagnosed. Such faults are often caused by component wear and fatigue damage due to long-term operation.

[0003] For large equipment equipped with magnetic bearings, due to their inherent frictionless and non-contact characteristics, the failure probability of the magnetic bearings themselves is extremely low. Failures are mostly caused by system-level faults in the large equipment containing the magnetic bearings or by faults in other components, resulting in a wide variety of fault types. Existing fault identification schemes struggle to collect samples of all types of faults for training, making effective fault identification difficult for large equipment equipped with magnetic bearings. Summary of the Invention

[0004] In view of this, this application provides a method and apparatus for determining equipment faults, which can be used to effectively identify faults in large equipment equipped with magnetic levitation bearings.

[0005] Specifically, this application is implemented through the following technical solution:

[0006] In a first aspect, embodiments of this application provide a method for determining equipment faults, including:

[0007] Obtain the rotor position information of the magnetic levitation bearing in the target device within the target time period;

[0008] Based on the rotor position information, open set fault identification is performed to determine a first target fault type that matches the rotor position information from multiple fault types; the multiple fault types include multiple preset fault types, as well as unknown fault types that represent types other than the preset fault types;

[0009] When the first target fault type is an unknown fault type, a second target fault type is determined from the multiple preset fault types based on the attribute characteristics of the rotor position information; the second target fault type is used to provide a reference for determining the actual type corresponding to the unknown fault type.

[0010] Optionally, the step of performing open-set fault identification based on the rotor position information, and determining a first target fault type matching the rotor position information from multiple fault types, includes:

[0011] The rotor position information is input into the trained fault classification model to obtain the first target fault type output by the fault classification model; the fault classification model is a neural network with an OpenMax layer for open set recognition.

[0012] Optionally, the fault classification model is determined through the following steps:

[0013] Construct the initial neural network;

[0014] The initial neural network is trained using the first rotor position information sample and the truth label corresponding to the first rotor position information sample to obtain a trained neural network.

[0015] The normalization layer of the trained neural network is replaced with the OpenMax layer for open set recognition to obtain the fault classification model.

[0016] Optionally, determining the second target fault type from the multiple preset fault types based on the attribute features of the rotor position information includes:

[0017] Determine the attribute features of the rotor position information and the feature distances between them and the attribute features of multiple second rotor position information samples;

[0018] Based on the feature distance and the preset approximation threshold, the target rotor position information sample is determined from the plurality of second rotor position information samples;

[0019] The truth label corresponding to the target rotor position information sample is determined to be the second target fault type.

[0020] Optionally, the attribute features of the rotor position information include statistical attribute features;

[0021] The statistical attribute features are determined through the following steps:

[0022] Determine the statistical values ​​of the rotor position information across multiple statistical dimensions;

[0023] For any given statistical dimension, based on the statistical values ​​under that statistical dimension and the corresponding judgment threshold, the statistical attribute features under that statistical dimension are determined.

[0024] Optionally, the attribute features include self-learning attribute features;

[0025] The self-learning attribute features are determined through the following steps:

[0026] The rotor position information is input into the trained feature extraction model to obtain the feature coefficient vector output by the feature extraction model, and the feature coefficient vector is used as the self-learning attribute feature.

[0027] Optionally, the feature extraction model can be trained through the following steps:

[0028] An initial feature extraction model is constructed; the initial feature extraction model is used to encode the rotor position information to obtain a first feature vector; the first feature vector is reconstructed using a feature matrix and a feature coefficient vector to obtain a second feature vector; the second feature vector is decoded to obtain the reconstructed information corresponding to the rotor position information;

[0029] The initial feature extraction model is trained based on the first loss function, the second loss function, and the third loss function to obtain a trained feature extraction model.

[0030] The trained feature extraction model has a feature matrix that satisfies the first loss function, the second loss function, and the third loss function;

[0031] The first loss function is used to minimize the difference between the first feature vector and the second feature vector; the second loss function is used to minimize the L0 norm of the feature coefficient vector; and the third loss function is used to minimize the difference between the rotor position information and the reconstructed information.

[0032] Optionally, when the attribute features include statistical attribute features and self-learning attribute features, the feature distance between the attribute features used to determine the rotor position information and the attribute features of the plurality of second rotor position information samples includes:

[0033] For any second rotor position information sample, determine the first difference information between the self-learning attribute features in the rotor position information and the self-learning attribute features in the attribute features of the second rotor position information sample; and determine the second difference information between the statistical attribute features in the rotor position information and the statistical attribute features in the attribute features of the second rotor position information sample.

[0034] Based on the first difference information, the second difference information, and the weight corresponding to the second difference information, the feature distance between the attribute features of the rotor position information and the attribute features of the second rotor position information sample is determined.

[0035] Secondly, embodiments of this application also provide a device for determining equipment faults, comprising:

[0036] The acquisition module is used to acquire the rotor position information of the magnetic levitation bearing in the target device within the target time period;

[0037] The first identification module is used to perform open set fault identification based on the rotor position information, and determine a first target fault type that matches the rotor position information from multiple fault types; the multiple fault types include multiple preset fault types, and unknown fault types that represent types other than the preset fault types;

[0038] The second identification module is used to determine a second target fault type from the multiple preset fault types based on the attribute characteristics of the rotor position information when the first target fault type is an unknown fault type; the second target fault type is used to provide a reference for determining the actual type corresponding to the unknown fault type.

[0039] Optionally, the first identification module is specifically used for:

[0040] The rotor position information is input into the trained fault classification model to obtain the first target fault type output by the fault classification model; the fault classification model is a neural network with an OpenMax layer for open set recognition.

[0041] Optionally, the device further includes a first training module for:

[0042] Construct the initial neural network;

[0043] The initial neural network is trained using the first rotor position information sample and the truth label corresponding to the first rotor position information sample to obtain a trained neural network.

[0044] The normalization layer of the trained neural network is replaced with the OpenMax layer for open set recognition to obtain the fault classification model.

[0045] Optionally, the second identification module is specifically used for:

[0046] Determine the attribute features of the rotor position information and the feature distances between them and the attribute features of multiple second rotor position information samples;

[0047] Based on the feature distance and the preset approximation threshold, the target rotor position information sample is determined from the plurality of second rotor position information samples;

[0048] The truth label corresponding to the target rotor position information sample is determined to be the second target fault type.

[0049] Optionally, the attribute features of the rotor position information include statistical attribute features; the second identification module is specifically used for:

[0050] Determine the statistical values ​​of the rotor position information across multiple statistical dimensions;

[0051] For any given statistical dimension, based on the statistical values ​​under that statistical dimension and the corresponding judgment threshold, the statistical attribute features under that statistical dimension are determined.

[0052] Optionally, the attribute features include self-learning attribute features; the second recognition module is specifically used for:

[0053] The rotor position information is input into the trained feature extraction model to obtain the feature coefficient vector output by the feature extraction model, and the feature coefficient vector is used as the self-learning attribute feature.

[0054] Optionally, the device further includes a second training module for:

[0055] An initial feature extraction model is constructed; the initial feature extraction model is used to encode the rotor position information to obtain a first feature vector; the first feature vector is reconstructed using a feature matrix and a feature coefficient vector to obtain a second feature vector; the second feature vector is decoded to obtain the reconstructed information corresponding to the rotor position information;

[0056] The initial feature extraction model is trained based on the first loss function, the second loss function, and the third loss function to obtain a trained feature extraction model.

[0057] The trained feature extraction model has a feature matrix that satisfies the first loss function, the second loss function, and the third loss function;

[0058] The first loss function is used to minimize the difference between the first feature vector and the second feature vector; the second loss function is used to minimize the L0 norm of the feature coefficient vector; and the third loss function is used to minimize the difference between the rotor position information and the reconstructed information.

[0059] Optionally, when the attribute features include statistical attribute features and self-learning attribute features, the second identification module is specifically used for:

[0060] For any second rotor position information sample, determine the first difference information between the self-learning attribute features in the rotor position information and the self-learning attribute features in the attribute features of the second rotor position information sample; and determine the second difference information between the statistical attribute features in the rotor position information and the statistical attribute features in the attribute features of the second rotor position information sample.

[0061] Based on the first difference information, the second difference information, and the weight corresponding to the second difference information, the feature distance between the attribute features of the rotor position information and the attribute features of the second rotor position information sample is determined.

[0062] Thirdly, an optional implementation of this application also provides a computer device, a processor, and a memory, wherein the memory stores machine-readable instructions executable by the processor, and the processor is used to execute the machine-readable instructions stored in the memory. When the machine-readable instructions are executed by the processor, they perform the steps of the first aspect above, or any possible implementation of the first aspect.

[0063] Fourthly, an optional implementation of this application also provides a computer-readable storage medium storing a computer program that, when run, performs the steps of the first aspect or any possible implementation of the first aspect.

[0064] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of this application.

[0065] The equipment fault determination method and apparatus provided in this application embodiment perform open set fault identification based on the rotor position information of the magnetic levitation bearing in the equipment. It can identify the fault types that appear during the training process, mark the fault types that do not appear during the training process as unknown faults, and mine the similarity between unknown fault samples and known fault types, thereby improving the accuracy and practicality of the magnetic levitation bearing fault diagnosis system.

[0066] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0067] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly described below. These drawings are incorporated in and constitute a part of this specification. They illustrate embodiments conforming to this application and, together with the specification, serve to explain the technical solutions of this application. It should be understood that the following drawings only show some embodiments of this application and should not be considered as limiting the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0068] Figure 1 A flowchart illustrating a method for determining equipment faults provided in this application embodiment;

[0069] Figure 2A schematic diagram of the feature extraction model provided in the embodiments of this application;

[0070] Figure 3 This is a schematic diagram of a device for determining equipment faults provided in an embodiment of this application;

[0071] Figure 4 This is a schematic diagram of the computer device structure provided in an embodiment of this application. Detailed Implementation

[0072] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0073] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0074] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0075] In this document, the term "and / or" merely describes a relationship, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Furthermore, the term "at least one" in this document means any combination of at least two of any one or more elements. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0076] Research has revealed that existing fault diagnosis mechanisms for large equipment typically focus on critical components such as bearings, with monitoring sensors positioned near the components being diagnosed. These faults are often caused by wear and fatigue damage resulting from long-term operation. However, for large equipment equipped with magnetic bearings, due to their inherent frictionless and non-contact characteristics, the probability of the magnetic bearings themselves failing is extremely low. Faults are more often caused by system-level failures within the large equipment or by failures in other components, resulting in a wide variety of fault types. Existing fault identification schemes struggle to collect samples of all types of faults for training, making effective fault identification for large equipment equipped with magnetic bearings difficult.

[0077] In view of this, this application provides a method and apparatus for determining equipment faults, which performs open set fault identification based on the rotor position information of the magnetic levitation bearing in the equipment. It can identify the fault types that appear during the training process, mark the fault types that do not appear during the training process as unknown faults, and mine the similarity between the unknown fault type samples and the known fault types, thereby improving the accuracy and practicality of the magnetic levitation bearing fault diagnosis system.

[0078] The deficiencies of the existing technical solutions are the result of the inventors' practice and careful research. Therefore, the discovery process of the above problems and the solutions proposed in this application below are the inventors' contributions to this application.

[0079] To facilitate understanding of this embodiment, a detailed description of the device fault determination method disclosed in this application embodiment will be provided first. The execution subject of the device fault determination method provided in this application embodiment is generally a computer device with a certain computing capability, such as a terminal device, a server, or other processing device. In some possible implementations, the device fault determination method can be implemented by a processor calling computer-readable instructions stored in memory.

[0080] See Figure 1 The diagram shown is a flowchart of a device fault determination method provided in an embodiment of this application. The method includes:

[0081] S101. Obtain the rotor position information of the magnetic levitation bearing in the target device within the target time period.

[0082] In this step, the target equipment can be a device equipped with magnetic levitation bearings. Magnetic levitation bearings are an advanced bearing technology that uses magnetic force to suspend the rotor (rotating component) in the stator (stationary component) without physical contact. It completely eliminates the physical contact and mechanical friction present in traditional mechanical bearings (such as ball bearings and sliding bearings). Magnetic levitation bearings are widely used in applications requiring high speed, high precision, long lifespan, zero pollution, low maintenance, or extreme environments, such as high-speed rotating machinery, the energy sector, medical equipment, aerospace, and semiconductor manufacturing.

[0083] In a magnetic bearing-rotor system, the magnetic bearing is typically equipped with a displacement sensor to monitor its operating status in real time. This step utilizes the displacement offset information collected by the displacement sensor over a specific time period to identify faults.

[0084] For example, a magnetic levitation bearing may be equipped with multiple displacement sensors, such as lateral and longitudinal displacement sensors at both ends of the bearing, as well as an axial displacement sensor. These sensors can capture the displacement changes of the magnetic levitation bearing in different directions, forming multidimensional time series data.

[0085] The acquired raw displacement data undergoes preprocessing, including denoising, filtering, and normalization, to eliminate interference signals and standardize the data format. The preprocessed displacement data is then segmented according to a predefined time window (target time period) to obtain a series of rotor position information samples.

[0086] In practical applications, the frequency and time window length for collecting rotor position information should be reasonably set according to the operating characteristics and fault features of the equipment to ensure that the collected data can fully reflect the operating status and potential fault features of the magnetic levitation bearing.

[0087] S102. Based on the rotor position information, perform open set fault identification and determine a first target fault type that matches the rotor position information from multiple fault types; the multiple fault types include multiple preset fault types and unknown fault types that represent types other than the preset fault types.

[0088] In this step, open-set identification technology is used to identify the fault type from the acquired rotor position information. Unlike traditional closed-set identification, open-set fault identification can identify fault types that have not been seen during training and mark them as "unknown faults," which is of great significance for fault diagnosis in practical engineering applications.

[0089] In some possible implementations, the rotor position information can be input into a trained fault classification model to obtain the first target fault type output by the fault classification model; the fault classification model is a neural network with an OpenMax layer for open set recognition.

[0090] The fault classification model can be used to identify faults in magnetic levitation bearings. The fault classification model can be trained using the following steps:

[0091] Construct the initial neural network;

[0092] The initial neural network is trained using the first rotor position information sample and the truth label corresponding to the first rotor position information sample to obtain a trained neural network.

[0093] The normalization layer of the trained neural network is replaced with the OpenMax layer for open set recognition to obtain the fault classification model.

[0094] The initial neural network can be a Convolutional Neural Network (CNN) or a Multilayer Perceptron (MLP), with the appropriate network structure selected based on the characteristics of the rotor position information. The initial neural network can serve as a basic classification model; after training the basic classification model, it can be modified into a model capable of open-set fault identification.

[0095] The first rotor position information sample is a sample of known fault types, used to train the neural network to identify known fault types. During training, the network parameters can be optimized using the backpropagation algorithm to enable the network to accurately identify known fault types. The first rotor position information sample can be used as the training sample for the initial neural network. By inputting the first rotor position information sample into the initial neural network, the identification result of the initial neural network on the first rotor position information can be obtained.

[0096] In this step, the first rotor position information sample can be represented as X = {x1, x2, ..., x...} n The corresponding truth label can be represented as Y∈[1,2,3,...,C], where n represents the number of dimensions of displacement information in the first rotor position information sample, and C represents the number of known fault types. Then, the trained initial neural network f:X→Y can be obtained.

[0097] For example, the initial neural network described above may include multiple convolutional layers, pooling layers, and fully connected layers. Convolutional layers are responsible for extracting local features from the rotor position information, such as frequency domain features and time domain features; pooling layers are used for dimensionality reduction and extraction of key features; and fully connected layers map the extracted features to a fault type space. During network design, factors such as data dimensionality, feature complexity, and computational resource limitations need to be considered, and parameters such as network depth, width, and activation function should be set appropriately. Furthermore, a normalized SoftMax layer needs to be added to the network's output layer to convert the network's output into probability distributions for various fault types. This initial neural network will serve as the foundation for the open-set fault identification model and will be trained and modified in subsequent steps.

[0098] Training the neural network is a crucial step in achieving accurate fault identification. This stage requires preparing a large number of rotor position information samples of known fault types as training data. These samples should cover as many known fault types and operating conditions as possible to improve the model's generalization ability. Each sample has a corresponding ground truth label indicating the fault type it belongs to. The training process employs supervised learning methods, optimizing network parameters through backpropagation and gradient descent.

[0099] For example, the rotor position information samples are first input into the initial neural network to obtain the network's predicted output. Then, the loss between the predicted output and the ground truth label is calculated; commonly used loss functions include the cross-entropy loss function. Next, the gradient of the loss function with respect to the network parameters is calculated using the backpropagation algorithm, and the network parameters are updated using gradient descent to reduce the loss. This process is repeated multiple times (multiple training epochs) until the network converges or reaches a preset number of training epochs. To prevent overfitting, regularization techniques such as L1 / L2 regularization and Dropout are typically used. Furthermore, techniques such as cross-validation can be used to evaluate the model's performance and adjust the network structure and hyperparameters. After sufficient training, the neural network can accurately identify known fault types in the training set, preparing for subsequent open-set fault identification.

[0100] A well-trained neural network typically uses a SoftMax layer as its output layer to map features to a probability distribution of fault types. Replacing the SoftMax layer with an OpenMax layer enables open-set fault identification, allowing the recognition of fault types not encountered during training. The OpenMax layer calculates the distance between the input sample and known fault types to determine if the sample belongs to a known fault type; if not, it is marked as an unknown fault type.

[0101] Traditional neural networks typically use a SoftMax layer as the final normalization layer, which transforms the network's output into probability distributions for each class, with the sum of these probabilities being 1. This design performs well in closed-set recognition problems, but has limitations in open-set recognition problems because it always classifies input samples into a known class, even if the sample does not actually belong to any known class. This makes it difficult to perform effective fault identification in the specific scenario of magnetic levitation bearings. To address this issue, this application replaces the SoftMax layer with an OpenMax layer.

[0102] The OpenMax layer works by calculating the distance between an input sample and each known category based on the statistical properties of known categories, and introducing an additional "unknown" category. In its implementation, a statistical model is first built for each known fault category, typically based on the distribution of training samples in the feature space. Then, for a new input sample, its distance or similarity to the statistical models of each known category is calculated. If the distance between a sample and all known categories exceeds a certain threshold, it is identified as the "unknown" category (output 0); otherwise, it is identified as the closest known category. This method effectively distinguishes between known and unknown fault types, improving the accuracy and reliability of the fault diagnosis system in practical applications. Through this modification, we obtain a fault classification model with open-set recognition capabilities, which can accurately identify not only known fault types but also unknown fault types not seen during training.

[0103] Thus, we can obtain the trained fault classification model f1: X→Y′, where Y′∈[0,1,2,3,...,C]. When the output is 0, it indicates that the sample does not belong to any known fault type, that is, it is an unknown fault.

[0104] S103. When the first target fault type is an unknown fault type, a second target fault type is determined from the multiple preset fault types based on the attribute characteristics of the rotor position information; the second target fault type is used to provide a reference for determining the actual type corresponding to the unknown fault type.

[0105] In this step, when the first target fault type is identified as an unknown fault type (i.e., the output is 0), the known fault type most similar to the unknown fault can be found by analyzing the attribute characteristics of the rotor position information, providing a reference for fault cause analysis.

[0106] In some possible implementations, the attribute features of the rotor position information can be determined, and the feature distances between the attribute features and the attribute features of a plurality of second rotor position information samples can be determined. Then, based on the feature distances and a preset approximate threshold, a target rotor position information sample can be determined from the plurality of second rotor position information samples. The truth label corresponding to the target rotor position information sample is determined to be the second target fault type.

[0107] The second rotor position information sample consists of samples with known fault types, used for feature comparison with samples of unknown fault types. Feature distance represents the distance between two samples in the feature space, and can be measured using methods such as Euclidean distance or cosine similarity. A preset approximate threshold is used to determine whether two samples are similar; if the feature distance is less than the threshold, the two samples are considered similar.

[0108] In some possible implementations, the attribute characteristics of the rotor position information include statistical attribute characteristics; the statistical attribute characteristics are determined through the following steps:

[0109] Determine the statistical values ​​of the rotor position information under multiple statistical dimensions; for any statistical dimension, based on the statistical values ​​under that statistical dimension and the judgment threshold corresponding to that statistical dimension, determine the statistical attribute features under that statistical dimension.

[0110] Statistical attributes can be statistical features extracted from rotor position information, including statistical quantities such as mean, variance, kurtosis, and skewness. Multiple statistical dimensions can include time-domain and frequency-domain statistical dimensions. Time-domain statistical dimensions include mean, variance, kurtosis, and skewness; frequency-domain statistical dimensions include typical components in the spectrum, such as second and third harmonics. In one possible implementation, statistical attributes can include features derived from rules based on expert experience, where each rule can be considered a statistical dimension.

[0111] For each statistical dimension, the statistical attribute features of that dimension are determined based on the statistical value and the corresponding judgment threshold. For example, if the statistical value of a certain statistical dimension exceeds the judgment threshold, the feature value of that dimension is 1; otherwise, it is 0. In this way, a binary feature vector can be obtained, representing the statistical features of the rotor position information.

[0112] In magnetic levitation bearing fault diagnosis, statistical characteristics are crucial indicators reflecting the system's operating status. After obtaining rotor position information, it's necessary to extract statistical values ​​across multiple dimensions. These values ​​describe the characteristics of the rotor position information from different perspectives. Specifically, statistical dimensions can be broadly categorized into time-domain and frequency-domain statistical dimensions. Time-domain statistical dimensions primarily focus on the statistical characteristics of rotor position information over time, including mean, standard deviation, variance, kurtosis, skewness, peak-to-peak value, waveform factor, impulse factor, and margin factor. The mean reflects the average level of pose deviation, the standard deviation and variance reflect the degree of fluctuation in pose deviation, kurtosis reflects the sharpness of the distribution, and skewness reflects the asymmetry of the distribution. These indicators are significant for identifying different types of faults. For example, bearing loosening typically leads to an increase in the standard deviation of pose deviation, while imbalance faults may cause an increase in the amplitude of pose deviation at a specific frequency. Frequency domain statistics focus on the characteristics of rotor position information in the frequency domain. This is typically achieved by performing a Fourier transform on the time-domain signal to obtain the spectrum, and then extracting features from the spectrum, such as the amplitude of specific frequency components (e.g., the first and second harmonics of the rotor speed), spectral energy distribution, and spectral kurtosis. These frequency domain features are particularly important for identifying periodic faults (such as imbalance or misalignment). By calculating these multi-dimensional statistical values, we can comprehensively capture the characteristics of rotor position information, providing strong support for subsequent fault identification.

[0113] After obtaining the statistical values ​​for each statistical dimension, these continuous statistical values ​​can be converted into discrete attribute features to facilitate subsequent feature comparison and fault identification. This conversion process can be completed based on preset judgment thresholds. For each statistical dimension, there is a corresponding judgment threshold, which is usually determined based on a large amount of historical data and expert experience. When the statistical value under that statistical dimension exceeds or falls below the judgment threshold, it can be considered that there is an anomaly under that dimension, and the corresponding feature value is set to 1; otherwise, the feature value is set to 0.

[0114] For example, if the kurtosis value of the pose offset signal exceeds a preset kurtosis threshold, then the feature value under the "kurtosis" statistical dimension is 1, indicating the presence of a kurtosis anomaly. Similarly, if the second harmonic amplitude of the pose offset signal exceeds a preset second harmonic threshold, then the feature value under the "second harmonic" statistical dimension is also 1, indicating the presence of a second harmonic anomaly. In this way, continuous statistical values ​​can be converted into binary feature vectors, with each element representing the presence or absence of anomalies under a specific statistical dimension. This binary representation simplifies the feature comparison process while retaining the key information required for fault diagnosis. It is important to note that the setting of the judgment threshold has a significant impact on the feature extraction effect. A threshold that is too high may cause some fault features to be ignored, while a threshold that is too low may cause normal fluctuations to be misjudged as fault features. Therefore, in practical applications, it is necessary to reasonably set the judgment thresholds for each statistical dimension according to the specific conditions of the equipment and the operating environment to obtain appropriate fault diagnosis results.

[0115] In some possible implementations, the attribute features include self-learning attribute features; the self-learning attribute features are determined through the following steps:

[0116] The rotor position information is input into the trained feature extraction model to obtain the feature coefficient vector output by the feature extraction model, and the feature coefficient vector is used as the self-learning attribute feature.

[0117] In this step, deep features of rotor position information can be extracted through automatic learning. These features may be difficult to identify or define under normal circumstances. To address this, a feature extraction model can be designed, employing an autoencoder structure, capable of automatically learning the feature representation of rotor position information.

[0118] See Figure 2 The diagram shown illustrates the feature extraction model provided in this embodiment. The model comprises three parts: an encoder, a feature reconstruction module, and a decoder. The encoder encodes the input rotor position information X into a feature vector Z, which captures the key features of the rotor position information. The feature reconstruction module reconstructs the feature vector using a pre-learned feature matrix A and feature coefficient vector α to obtain the reconstructed feature vector Z'. The feature matrix A = [A1, A2, ..., A...]. M It contains a series of basic features, and the feature coefficient vector α = [α1, α2, ..., α] M The weights of these basic features are represented by ]. The reconstructed feature vector can be represented as z. i =α1A1 + α2A2 + ... + α M A M This is the weighted sum of the basic features. Finally, the decoder decodes the reconstructed feature vector into reconstructed information. This reconstructed information is as close as possible to the original input X.

[0119] During model training, three optimization objectives are adopted: (1) minimizing the difference between the feature vector Z and the reconstructed feature vector Z' to ensure the accuracy of feature reconstruction; (2) minimizing the L0 norm of the feature coefficient vector α to make the feature representation as sparse as possible and improve the interpretability of the features; (3) minimizing the difference between the original input X and the reconstructed information. The differences between them ensure the accuracy of the entire self-encoding process.

[0120] The feature extraction model trained in this way can map rotor position information into a sparse feature coefficient vector. This vector captures the key features of the rotor position information and can be used as a self-learning attribute feature for subsequent fault identification. Compared with statistical features, self-learning features can capture more complex and abstract fault features, improving the accuracy of fault identification.

[0121] In some possible implementations, the feature extraction model can be trained using the following steps:

[0122] An initial feature extraction model is constructed; the initial feature extraction model is used to encode rotor position information to obtain a first feature vector; the first feature vector is reconstructed using a feature matrix and a feature coefficient vector to obtain a second feature vector; the second feature vector is decoded to obtain the reconstructed information corresponding to the rotor position information; the initial feature extraction model is trained based on a first loss function, a second loss function, and a third loss function to obtain a trained feature extraction model; the trained feature extraction model has a feature matrix that satisfies the first loss function, the second loss function, and the third loss function; the first loss function is used to minimize the difference between the first feature vector and the second feature vector; the second loss function is used to minimize the L0 norm of the feature coefficient vector; the third loss function is used to minimize the difference between the rotor position information and the reconstructed information.

[0123] The initial feature extraction model described above can be an autoencoder structure, consisting of an encoder and a decoder. The encoder encodes the rotor position information into feature vectors, and the decoder decodes the feature vectors into reconstructed rotor position information. The feature reconstruction module is located between the encoder and decoder and is used to reconstruct the feature vectors using the feature matrix and feature coefficient vectors.

[0124] During training, the feature extraction model can be trained by optimizing three loss functions to achieve the above three optimization objectives: the first loss function is used to minimize the difference between the first feature vector (the feature vector output by the encoder) and the second feature vector (the reconstructed feature vector) to ensure the accuracy of feature reconstruction; the second loss function is used to minimize the L0 norm of the feature coefficient vector to make the feature coefficient vector as sparse as possible and improve the interpretability of the features; the third loss function is used to minimize the difference between the rotor position information and the reconstructed information to ensure the reconstruction accuracy of the entire autoencoder.

[0125] The first loss function can be expressed as min||z′ i -z i The second loss function can be expressed as min||α||0, and the third loss function can be expressed as...

[0126] The feature extraction model trained in this way can automatically learn the features of rotor position information and output a sparse feature coefficient vector as self-learning attribute features.

[0127] In some possible implementations, when the attribute features include statistical attribute features and self-learning attribute features, the feature distance can be determined through the following steps:

[0128] For any second rotor position information sample, determine the first difference information between the self-learning attribute features in the rotor position information and the self-learning attribute features in the attribute features of the second rotor position information sample; and determine the second difference information between the statistical attribute features in the rotor position information and the statistical attribute features in the attribute features of the second rotor position information sample; based on the first difference information, the second difference information, and the weights corresponding to the second difference information, determine the feature distance between the attribute features of the rotor position information and the attribute features of the second rotor position information sample.

[0129] The first difference information represents the difference between the two samples in the self-learning feature space, which can be calculated using metrics such as Euclidean distance and cosine similarity. The second difference information represents the difference between the two samples in the statistical feature space, which, since statistical features are binary vectors, can be calculated using metrics such as Hamming distance.

[0130] Weights are used to balance the importance of self-learning features and statistical features in feature distance calculation. By adjusting the weights, the contribution ratio of the two features in the feature distance calculation can be controlled. Feature distance can be expressed as: Feature distance = First difference information + Weight * Second difference information.

[0131] For example, the feature distance can be represented as di , where is the feature distance between the rotor position information and the i-th second rotor position information sample; the first difference information can be represented as ||α|| i -α test ‖, where α i Let α be the self-learned attribute feature in the attribute features of the i-th second rotor position information sample. test The self-learning attribute features in the rotor position information; the second difference information can be represented as ||t|| i -t test ‖, where t i t represents the statistical attribute features among the attribute features of the i-th second rotor position information sample. test These are the statistical attribute features in the rotor position information.

[0132] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0133] Corresponding to the embodiments of the aforementioned equipment fault determination method, this application also provides embodiments of the equipment fault determination apparatus.

[0134] The embodiments of the equipment fault determination device of this application can be applied to equipment equipped with magnetic levitation bearings. The device embodiments can be implemented through software, hardware, or a combination of both. Taking software implementation as an example, as a logical device, it is formed by the processor of the device loading the corresponding computer program instructions from the non-volatile memory into memory for execution.

[0135] See Figure 3 The diagram shown is a schematic of a device for determining equipment faults provided in an embodiment of this application.

[0136] The device includes:

[0137] The acquisition module 310 is used to acquire the rotor position information of the magnetic levitation bearing in the target device within the target time period;

[0138] The first identification module 320 is used to perform open set fault identification based on the rotor position information, and determine a first target fault type that matches the rotor position information from a variety of fault types; the variety of fault types includes a variety of preset fault types, and an unknown fault type representing a type other than the preset fault types;

[0139] The second identification module 330 is used to determine a second target fault type from the multiple preset fault types based on the attribute characteristics of the rotor position information when the first target fault type is an unknown fault type; the second target fault type is used to provide a reference for determining the actual type corresponding to the unknown fault type.

[0140] Optionally, the first identification module 320 is specifically used for:

[0141] The rotor position information is input into the trained fault classification model to obtain the first target fault type output by the fault classification model; the fault classification model is a neural network with an OpenMax layer for open set recognition.

[0142] Optionally, the device further includes a first training module 340, used for:

[0143] Construct the initial neural network;

[0144] The initial neural network is trained using the first rotor position information sample and the truth label corresponding to the first rotor position information sample to obtain a trained neural network.

[0145] The normalization layer of the trained neural network is replaced with the OpenMax layer for open set recognition to obtain the fault classification model.

[0146] Optionally, the second identification module 330 is specifically used for:

[0147] Determine the attribute features of the rotor position information and the feature distances between them and the attribute features of multiple second rotor position information samples;

[0148] Based on the feature distance and the preset approximation threshold, the target rotor position information sample is determined from the plurality of second rotor position information samples;

[0149] The truth label corresponding to the target rotor position information sample is determined to be the second target fault type.

[0150] Optionally, the attribute features of the rotor position information include statistical attribute features; the second identification module 330 is specifically used for:

[0151] Determine the statistical values ​​of the rotor position information across multiple statistical dimensions;

[0152] For any given statistical dimension, based on the statistical values ​​under that statistical dimension and the corresponding judgment threshold, the statistical attribute features under that statistical dimension are determined.

[0153] Optionally, the attribute features include self-learning attribute features; the second recognition module 330 is specifically used for:

[0154] The rotor position information is input into the trained feature extraction model to obtain the feature coefficient vector output by the feature extraction model, and the feature coefficient vector is used as the self-learning attribute feature.

[0155] Optionally, the device further includes a second training module 350, used for:

[0156] An initial feature extraction model is constructed; the initial feature extraction model is used to encode the rotor position information to obtain a first feature vector; the first feature vector is reconstructed using a feature matrix and a feature coefficient vector to obtain a second feature vector; the second feature vector is decoded to obtain the reconstructed information corresponding to the rotor position information;

[0157] The initial feature extraction model is trained based on the first loss function, the second loss function, and the third loss function to obtain a trained feature extraction model.

[0158] The trained feature extraction model has a feature matrix that satisfies the first loss function, the second loss function, and the third loss function;

[0159] The first loss function is used to minimize the difference between the first feature vector and the second feature vector; the second loss function is used to minimize the L0 norm of the feature coefficient vector; and the third loss function is used to minimize the difference between the rotor position information and the reconstructed information.

[0160] Optionally, when the attribute features include statistical attribute features and self-learning attribute features, the second identification module 330 is specifically used for:

[0161] For any second rotor position information sample, determine the first difference information between the self-learning attribute features in the rotor position information and the self-learning attribute features in the attribute features of the second rotor position information sample; and determine the second difference information between the statistical attribute features in the rotor position information and the statistical attribute features in the attribute features of the second rotor position information sample.

[0162] Based on the first difference information, the second difference information, and the weight corresponding to the second difference information, the feature distance between the attribute features of the rotor position information and the attribute features of the second rotor position information sample is determined.

[0163] The specific implementation process of the functions and roles of each unit in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0164] This application also provides a computer device, such as... Figure 4 The diagram shown is a schematic representation of a computer device structure provided in an embodiment of this application, including:

[0165] Processor 41 and memory 42; the memory 42 stores machine-readable instructions executable by the processor 41, and the processor 41 executes the machine-readable instructions stored in the memory 42. When the machine-readable instructions are executed by the processor 41, the processor 41 performs the following steps:

[0166] Obtain the rotor position information of the magnetic levitation bearing in the target device within the target time period;

[0167] Based on the rotor position information, open set fault identification is performed to determine a first target fault type that matches the rotor position information from multiple fault types; the multiple fault types include multiple preset fault types, as well as unknown fault types that represent types other than the preset fault types;

[0168] When the first target fault type is an unknown fault type, a second target fault type is determined from the multiple preset fault types based on the attribute characteristics of the rotor position information; the second target fault type is used to provide a reference for determining the actual type corresponding to the unknown fault type.

[0169] The aforementioned memory 42 includes a main memory 421 and an external memory 422; the main memory 421, also known as internal memory, is used to temporarily store the computational data in the processor 41, as well as the data exchanged with external memory 422 such as a hard disk. The processor 41 exchanges data with the external memory 422 through the main memory 421.

[0170] The specific execution process of the above instructions can be referred to the steps of the equipment fault determination method described in the embodiments of this application, and will not be repeated here.

[0171] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. 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 application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0172] This application also provides a computer-readable storage medium storing a computer program. When a processor runs the computer program, it executes the steps of the device fault determination method described in the above-described method embodiments. The storage medium can be a volatile or non-volatile computer-readable storage medium.

[0173] This application also provides a computer program product, including a computer program / instruction, which, when executed by the computer program / instruction processor, implements the device fault determination method provided in the various embodiments of this application.

[0174] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium; in another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0175] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. In the several embodiments provided in this application, it can be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division; in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0176] The units described as separate components may or may not be physically separate. 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 units can be selected to achieve the purpose of this embodiment according to actual needs.

[0177] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0178] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0179] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the technical scope disclosed in this application. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

[0180] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for determining equipment faults, characterized in that, The method includes: Obtain the rotor position information of the magnetic levitation bearing in the target device within the target time period; Based on the rotor position information, open set fault identification is performed to determine a first target fault type that matches the rotor position information from multiple fault types; the multiple fault types include multiple preset fault types, as well as unknown fault types that represent types other than the preset fault types; If the first target fault type is an unknown fault type, a second target fault type is determined from the multiple preset fault types based on the attribute characteristics of the rotor position information; the second target fault type is used to provide a reference for determining the actual type corresponding to the unknown fault type. The step of determining a second target fault type from multiple preset fault types based on the attribute features of the rotor position information includes: determining the feature distances between the attribute features of the rotor position information and the attribute features of multiple second rotor position information samples; determining a target rotor position information sample from the multiple second rotor position information samples based on the feature distances and a preset approximation threshold; and determining the truth label corresponding to the target rotor position information sample as the second target fault type. The attribute features of the rotor position information include statistical attribute features; the statistical attribute features are determined by the following steps: determining the statistical values ​​of the rotor position information under multiple statistical dimensions; for any statistical dimension, based on the statistical values ​​under that statistical dimension and the judgment threshold corresponding to that statistical dimension, determining the statistical attribute features under that statistical dimension. The attribute features include self-learning attribute features; the self-learning attribute features are determined by the following steps: inputting the rotor position information into a trained feature extraction model to obtain the feature coefficient vector output by the feature extraction model, and using the feature coefficient vector as the self-learning attribute features; The feature extraction model is trained through the following steps: Constructing an initial feature extraction model; the initial feature extraction model encodes rotor position information to obtain a first feature vector; reconstructing the first feature vector using a feature matrix and feature coefficient vectors to obtain a second feature vector; decoding the second feature vector to obtain reconstructed information corresponding to the rotor position information; training the initial feature extraction model based on a first loss function, a second loss function, and a third loss function to obtain a trained feature extraction model; the trained feature extraction model has a feature matrix that satisfies the first loss function, the second loss function, and the third loss function; the first loss function minimizes the difference between the first feature vector and the second feature vector; the second loss function minimizes the difference between the feature coefficient vectors. Norm; the third loss function is used to minimize the difference between the rotor position information and the reconstructed information; When the attribute features include statistical attribute features and self-learning attribute features, determining the feature distance between the attribute features of the rotor position information and the attribute features of multiple second rotor position information samples includes: for any second rotor position information sample, determining a first difference information between the self-learning attribute features in the rotor position information and the self-learning attribute features in the attribute features of the second rotor position information sample; and determining a second difference information between the statistical attribute features in the rotor position information and the statistical attribute features in the attribute features of the second rotor position information sample; and determining the feature distance between the attribute features of the rotor position information and the attribute features of the second rotor position information sample based on the first difference information, the second difference information, and the weight corresponding to the second difference information.

2. The method according to claim 1, characterized in that, The step of performing open-set fault identification based on the rotor position information, and determining a first target fault type that matches the rotor position information from multiple fault types, includes: The rotor position information is input into the trained fault classification model to obtain the first target fault type output by the fault classification model; the fault classification model is a neural network with an OpenMax layer for open set recognition.

3. The method according to claim 2, characterized in that, The fault classification model is determined through the following steps: Construct the initial neural network; The initial neural network is trained using the first rotor position information sample and the truth label corresponding to the first rotor position information sample to obtain a trained neural network. The normalized SoftMax layer of the trained neural network is replaced with the OpenMax layer of the open set recognition to obtain the fault classification model.

4. A device for determining equipment faults, characterized in that, The device includes: The acquisition module is used to acquire the rotor position information of the magnetic levitation bearing in the target device within the target time period; The first identification module is used to perform open set fault identification based on the rotor position information, and determine a first target fault type that matches the rotor position information from multiple fault types; the multiple fault types include multiple preset fault types, and unknown fault types that represent types other than the preset fault types; The second identification module is used to determine a second target fault type from the multiple preset fault types based on the attribute features of the rotor position information when the first target fault type is an unknown fault type; the second target fault type is used to provide a reference for determining the actual type corresponding to the unknown fault type. The second identification module is specifically used to: determine the feature distance between the attribute features of the rotor position information and the attribute features of a plurality of second rotor position information samples; determine the target rotor position information sample from the plurality of second rotor position information samples based on the feature distance and a preset approximate threshold; and determine the truth label corresponding to the target rotor position information sample as the second target fault type. The attribute features of the rotor position information include statistical attribute features; the second identification module is specifically used to: determine the statistical values ​​of the rotor position information under multiple statistical dimensions; for any statistical dimension, based on the statistical values ​​under that statistical dimension and the judgment threshold corresponding to that statistical dimension, determine the statistical attribute features under that statistical dimension; The attribute features include self-learning attribute features; the second recognition module is specifically used to: input the rotor position information into the trained feature extraction model to obtain the feature coefficient vector output by the feature extraction model, and use the feature coefficient vector as the self-learning attribute features; The device further includes a second training module, configured to: construct an initial feature extraction model; the initial feature extraction model is used to encode rotor position information to obtain a first feature vector; reconstruct the first feature vector using a feature matrix and a feature coefficient vector to obtain a second feature vector; decode the second feature vector to obtain reconstructed information corresponding to the rotor position information; train the initial feature extraction model based on a first loss function, a second loss function, and a third loss function to obtain a trained feature extraction model; the trained feature extraction model has a feature matrix that satisfies the first loss function, the second loss function, and the third loss function; the first loss function is used to minimize the difference between the first feature vector and the second feature vector; the second loss function is used to minimize the difference between the feature coefficient vector and the first feature vector. Norm; the third loss function is used to minimize the difference between the rotor position information and the reconstructed information; When the attribute features include statistical attribute features and self-learning attribute features, the second identification module is specifically configured to: for any second rotor position information sample, determine a first difference information between the self-learning attribute features in the rotor position information and the self-learning attribute features in the attribute features of the second rotor position information sample; and determine a second difference information between the statistical attribute features in the rotor position information and the statistical attribute features in the attribute features of the second rotor position information sample; and determine the feature distance between the attribute features of the rotor position information and the attribute features of the second rotor position information sample based on the first difference information, the second difference information, and the weight corresponding to the second difference information.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 3.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method as described in any one of claims 1 to 3.

Images