New energy automobile motor fault diagnosis method, device and equipment and storage medium

Abstract

The invention relates to the technical field of deep learning, and discloses a new energy automobile motor fault diagnosis method, device and equipment and a storage medium, and the method comprises the steps: obtaining a target motor state signal of a motor of a new energy automobile in a target time period, and determining target knowledge entry data corresponding to the target motor state signal; and inputting the target motor state signal and the target knowledge entry data into a trained motor fault diagnosis model to obtain a fault diagnosis result of the motor. By applying the technical scheme of the invention, the real-time equipment state and the knowledge items can be effectively combined, the accuracy and the interpretability of the diagnosis result are enhanced, and the fault diagnosis precision of the new energy automobile motor is optimized while the information fusion efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of deep learning technology, and specifically to a method, device, equipment and storage medium for diagnosing motor faults in new energy vehicles. Background Art

[0002] Globally, new energy vehicles, as a key direction for the future development of the automotive industry, have become an important force in promoting sustainable development and environmental protection. This transformation not only responds to environmental protection needs, but also represents the forefront of technological innovation. One of the core technologies of new energy vehicles is its motor system, which is directly related to the operating efficiency and safety of the vehicle.

[0003] As the core component of new energy vehicles, the design and performance of the motor have a decisive impact on the performance of the entire vehicle. Unlike the internal combustion engines of traditional vehicles, the motors of new energy vehicles mainly rely on electric drive, involving problems such as insulation damage of motor windings, wear of motor bearings, and failure of power electronic components. Due to the particularity of these motor components in design and materials, their failure modes and influencing factors are completely different from those of internal combustion engines.

[0004] However, new energy vehicles may face complex environmental factors such as large temperature changes and frequent vibrations during use, which are potential factors that may accelerate motor failure. For example, the insulation of motor windings is more easily damaged in high temperature environments, motor bearings wear faster under vibration conditions, and power electronic components may fail due to temperature fluctuations.

[0005] Effective fault diagnosis can not only prevent potential safety risks in advance and ensure the safe and stable operation of the vehicle, but is also the key to improving the performance of the vehicle and extending its service life. The motor system of new energy vehicles has higher requirements for fault diagnosis methods due to its complexity of structure and function. Existing motor fault diagnosis technologies mainly rely on the analysis of timing signals generated during motor operation, such as current and vibration signals. Although these methods have achieved some results in some specific scenarios, they still have major defects in the face of the diversity and complexity of modern motor systems, especially some fault conditions occur less frequently (fault data presents a long-tail distribution), which is difficult to identify simply by relying on data models.

[0006] Therefore, it is urgent to provide a technical solution to solve the above problems. Summary of the invention

[0007] In view of the above problems, the present application provides a new energy vehicle motor fault diagnosis method, device, equipment and storage medium to solve the problem in the prior art that the faults of new energy vehicle motors under different operating conditions cannot be accurately identified.

[0008] According to one aspect of an embodiment of the present application, a method for diagnosing a fault in a motor of a new energy vehicle is provided, the method comprising:

[0009] Acquire a target motor state signal of a motor of a new energy vehicle in a target period, and determine target knowledge item data corresponding to the target motor state signal;

[0010] The target motor state signal and the target knowledge item data are input into a trained motor fault diagnosis model to obtain a fault diagnosis result of the motor; wherein the motor fault diagnosis model comprises: a multi-scale state signal extraction module and a knowledge-data fusion feature extraction module; the multi-scale state signal extraction module is used to perform multi-scale feature extraction on the target motor state signal to obtain motor state information features; the knowledge-data fusion feature extraction module is used to: obtain fusion features based on the knowledge item embedding features corresponding to the motor state information features and the target knowledge items, perform linear mapping and softmax activation, and obtain the fault diagnosis result of the motor.

[0011] In an optional manner, the multi-scale state signal extraction module includes: a long short-term memory network, a 1D-Resnet network, a first KANs network, a second KANs network, a first feature cascade layer, and a second feature cascade layer; the target motor state signal includes: a motor high-frequency signal including a motor current signal, a motor vibration signal, and a motor speed signal, and a motor temperature signal;

[0012] The long short-term memory network is used to: dynamically extract long-distance dependency information of the motor high-frequency signal to obtain a first intermediate feature;

[0013] The 1D-Resnet network is used to: extract features of different levels from the motor high-frequency signal to obtain a second intermediate feature;

[0014] The first feature cascade layer is used to: perform feature fusion on the first intermediate feature and the second intermediate feature to obtain a third intermediate feature;

[0015] The first KANs network is used to: perform nonlinear mapping on the third intermediate feature to obtain a fourth intermediate feature;

[0016] The second KANs network is used to: perform nonlinear mapping on the motor temperature signal after Lagrange interpolation processing to obtain a fifth intermediate feature;

[0017] The second feature cascade layer is used to: perform feature fusion on the fourth intermediate feature and the fifth intermediate feature to obtain the motor state information feature.

[0018] In an optional manner, the knowledge-data fusion feature extraction module is a Transformer network including an encoder structure and a decoder structure;

[0019] The encoder structure is used to: obtain the fusion feature according to the motor state information feature and the knowledge item embedding feature;

[0020] The decoder structure is used to perform linear mapping and softmax activation on the fused features to obtain the fault diagnosis result of the motor.

[0021] In an optional manner, the method further includes:

[0022] Knowledge extraction is performed on the motor operation and maintenance external knowledge base of the motor to obtain a plurality of knowledge item data corresponding to the motor, and a mapping relationship between the motor state signal and the knowledge item data is constructed; wherein each motor state signal corresponds to one knowledge item data.

[0023] In an optional manner, the knowledge item data includes: phenomena, characteristics, causes, impacts and maintenance methods of specific faults.

[0024] According to another aspect of the embodiment of the present application, a new energy vehicle motor fault diagnosis device is provided, comprising:

[0025] A processing unit, used to obtain a target motor state signal of a motor of a new energy vehicle in a target period of time, and determine target knowledge item data corresponding to the target motor state signal;

[0026] An operating unit is used to input the target motor state signal and the target knowledge item data into a trained motor fault diagnosis model to obtain a fault diagnosis result of the motor; wherein the motor fault diagnosis model includes: a multi-scale state signal extraction module and a knowledge-data fusion feature extraction module; the multi-scale state signal extraction module is used to perform multi-scale feature extraction on the target motor state signal to obtain motor state information features; the knowledge-data fusion feature extraction module is used to: obtain fusion features based on the motor state information features and the knowledge item embedding features corresponding to the target knowledge items, perform linear mapping and softmax activation, and obtain the fault diagnosis result of the motor.

[0027] In an optional manner, the multi-scale state signal extraction module includes: a long short-term memory network, a 1D-Resnet network, a first KANs network, a second KANs network, a first feature cascade layer, and a second feature cascade layer; the target motor state signal includes: a motor high-frequency signal including a motor current signal, a motor vibration signal, and a motor speed signal, and a motor temperature signal;

[0028] The long short-term memory network is used to: dynamically extract long-distance dependency information of the motor high-frequency signal to obtain a first intermediate feature;

[0029] The 1D-Resnet network is used to: extract features of different levels from the motor high-frequency signal to obtain a second intermediate feature;

[0030] The first feature cascade layer is used to: perform feature fusion on the first intermediate feature and the second intermediate feature to obtain a third intermediate feature;

[0031] The first KANs network is used to: perform nonlinear mapping on the third intermediate feature to obtain a fourth intermediate feature;

[0032] The second KANs network is used to: perform nonlinear mapping on the motor temperature signal after Lagrangian interpolation processing to obtain a fifth intermediate feature;

[0033] The second feature cascade layer is used to: perform feature fusion on the fourth intermediate feature and the fifth intermediate feature to obtain the motor state information feature.

[0034] In an optional manner, the knowledge-data fusion feature extraction module is: a Transformer network including an encoder structure and a decoder structure;

[0035] The encoder structure is used to: obtain the fusion feature according to the motor state information feature and the knowledge item embedding feature;

[0036] The decoder structure is used to perform linear mapping and softmax activation on the fused features to obtain the fault diagnosis result of the motor.

[0037] According to another aspect of the embodiment of the present application, a new energy vehicle motor fault diagnosis device is provided, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other through the communication bus;

[0038] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations such as the new energy vehicle motor fault diagnosis method of the present invention.

[0039] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, in which at least one executable instruction is stored, and the executable instruction enables a new energy vehicle motor fault diagnosis device / equipment to perform operations such as the new energy vehicle motor fault diagnosis method of the present invention.

[0040] The embodiment of the present application obtains a target motor state signal of a motor of a new energy vehicle in a target time period, and determines target knowledge item data corresponding to the target motor state signal; the target motor state signal and the target knowledge item data are input into a trained motor fault diagnosis model to obtain a fault diagnosis result of the motor, which can effectively combine real-time device status with knowledge items, enhance the accuracy and interpretability of the diagnosis result, and optimize the accuracy of new energy vehicle motor fault diagnosis while improving information fusion efficiency.

[0041] The above description is only an overview of the technical solution of the embodiment of the present application. In order to more clearly understand the technical means of the embodiment of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the embodiment of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings are only used to illustrate the embodiments and are not to be considered as limiting the present application. In addition, the same reference symbols are used to represent the same components throughout the accompanying drawings. In the accompanying drawings:

[0043] Figure 1 A schematic diagram showing a flow chart of a first embodiment of a new energy vehicle motor fault diagnosis method provided by the present application;

[0044] Figure 2 The structural schematic diagram of the motor fault diagnosis model is shown;

[0045] Figure 3 shows a schematic diagram of the structure of a multi-scale state signal extraction module;

[0046] Figure 4 A schematic diagram showing a flow chart of a second embodiment of a new energy vehicle motor fault diagnosis method provided by the present application;

[0047] Figure 5 A schematic diagram showing the principle of knowledge extraction is shown;

[0048] Figure 6 A schematic diagram showing the structure of an embodiment of a new energy vehicle motor fault diagnosis device provided by the present application is shown;

[0049] Figure 7 A structural schematic diagram of an embodiment of a new energy vehicle motor fault diagnosis device provided in the present application is shown. DETAILED DESCRIPTION

[0050] Here, exemplary embodiments will be described in detail, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are only examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.

[0051] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0052] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0053] The term "multiple" as used in this application refers to two or more than two. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.

[0054] Existing motor fault diagnosis technologies mainly rely on the analysis of timing signals generated during motor operation, such as current and vibration signals. Although these methods have achieved some results in some specific scenarios, they still have major defects in the face of the diversity and complexity of modern motor systems, especially some faults occur less frequently (fault data presents a long-tail distribution), which are difficult to identify simply by relying on data models. Based on this:

[0055] Figure 1 The flowchart of the first embodiment of the new energy vehicle motor fault diagnosis method provided by the present application is shown, and the method is executed by the new energy vehicle motor fault diagnosis device. Figure 1 As shown, the method comprises the following steps:

[0056] Step S110: obtaining a target motor state signal of a motor of a new energy vehicle in a target period, and determining target knowledge item data corresponding to the target motor state signal.

[0057] Among them, the target time period defaults to the current time period, and the length of the target time period and the frequency of data collection can be set according to actual conditions, and there is no restriction here. The target motor state signal is the time series of the motor state signal in the target time period, and the motor state signal includes: current signal I(t), vibration signal V(t), speed signal S(t) and temperature signal T(t). The target knowledge item data is the knowledge item data corresponding to the target motor state signal. In this embodiment, different motor state signals correspond to one knowledge item data respectively.

[0058] Step S120: inputting the target motor state signal and the target knowledge item data into the trained motor fault diagnosis model to obtain the motor fault diagnosis result.

[0059] Among them, Figure 2 As shown in FIG. 1 , the motor fault diagnosis model includes: a multi-scale state signal extraction module and a knowledge-data fusion feature extraction module. The multi-scale state signal extraction module is used to extract multi-scale features of the target motor state signal to obtain motor state information features; the knowledge-data fusion feature extraction module is used to obtain fusion features based on the motor state information features and the knowledge item embedding features corresponding to the target knowledge items, and perform linear mapping and softmax activation to obtain the motor fault diagnosis result. The fault diagnosis result at least includes: the fault category of the motor. The target knowledge items are embedded using a pre-trained model such as BERT to obtain the knowledge item embedding feature X. K ={x k1 ,L,x kn}.

[0060] The technical solution of this embodiment can effectively combine the real-time device status with knowledge items, enhance the accuracy and interpretability of the diagnosis results, and optimize the accuracy of new energy vehicle motor fault diagnosis while improving the efficiency of information fusion.

[0061] In an alternative approach, Figure 3 As shown, the multi-scale state signal extraction module includes: a long short-term memory network, a 1D-Resnet network, a first KANs network, a second KANs network, a first feature cascade layer and a second feature cascade layer; the target motor state signal includes: a motor high-frequency signal including a motor current signal, a motor vibration signal and a motor speed signal, and a motor temperature signal.

[0062] Among them, the motor current signal I(t), the motor vibration signal V(t) and the motor speed signal S(t) are all high-frequency signals of the motor, and the motor temperature signal T(t) is a low-frequency signal of the motor. In this embodiment, the sampling frequency of the motor high-frequency signal is set to 1000Hz, and the sampling frequency of the motor low-frequency signal is set to 2Hz. A fixed-size window and a corresponding excerpt signal are generated by gradually moving the sliding time window on the motor high-frequency signal, and the excerpt signal is taken as a data sample. In this example, the sliding time window length is 2s, and the motor high-frequency signal is downsampled at the same time, and the downsampling ratio is 2. Let F(t) = {I(t), V(t), S(t)}, in this embodiment, batch_size is set to 8, and F(t) is the motor high-frequency signal.

[0063] The long short-term memory network is used to dynamically extract long-distance dependent information of the motor high-frequency signal to obtain the first intermediate feature.

[0064] Among them, the long short-term memory network can capture the long-distance dependency information in the time series signal excellently, and realize dynamic information capture through the gating mechanism. The long short-term memory network is set to 32 LSTM units by default. The input of the first LSTM unit of the long short-term memory network is the motor high-frequency signal F(t). The long short-term memory network can solve the long-term dependency problem in long sequence data. By introducing the gating mechanism: input gate, forget gate and output gate, it effectively controls the storage, update and output of information. The calculation method of each LSTM unit is the same. Taking the first LSTM unit as an example, its calculation formula is:

[0065] h1=f(W1g[h0,F(t)]+b1)

[0066] Among them, h1 is the output of the first time step (the first LSTM unit), h0 is the output of the previous time step, F(t) is the motor high-frequency signal, W1 is the weight matrix of the first LSTM unit, and b1 is the bias vector of the first LSTM unit. In the first time step, an initial hidden state h0 is usually initialized, which is usually set to a zero vector, and f(g) is the activation function, which is ReLU in this embodiment.

[0067] It should be noted that the output of the previous LSTM unit is used as the input of the next LSTM unit. The output obtained at the last time step is defined as the first intermediate feature h t-LSTM .

[0068] The 1D-Resnet network is used to extract features of different levels from the motor high-frequency signal to obtain the second intermediate features.

[0069] Among them, the 1D-ResNet network can capture features from different levels from shallow to deep through stacked residual blocks, and the multi-scale application provides the model with a more comprehensive ability to capture temporal features. The 1D-ResNet network is a deep neural network used to process sequence data. Its core feature is the use of residual blocks, each of which contains multiple convolutional layers and a skip connection. In this embodiment, the 1D-ResNet network includes three residual blocks, each of which includes two 1D convolutional layers, and each 1D convolutional layer is sequentially connected to batch normalization and ReLU activation function, and an identity mapping is set, and then the ReLU activation function is passed. The calculation formula for each residual block is:

[0070] y 1D =F 1D (x,W i )+W s x

[0071] Among them, y 1D represents the output of the current residual block, F 1D (g) is the residual mapping function to be learned, which is two 1D convolutional layers in this embodiment, x is the input of the current residual block, and W i is the comprehensive weight matrix of the current residual block, W s is the linear projection weight matrix.

[0072] It should be noted that the input of the first residual block in this embodiment is the motor high-frequency signal, the convolution kernel k=3, the step size s=1, and the padding p=1, then the output of the last residual block is defined as the second intermediate feature y 1D-out .

[0073] The first feature cascade layer is used to fuse the first intermediate feature with the second intermediate feature to obtain a third intermediate feature.

[0074] Among them, the superposition of multi-scale features is achieved through the first feature cascade layer, which can effectively retain all original feature information. The specific formula is:

[0075] y ms-out =y 1D-out +h t-LSTM

[0076] Among them, y ms-out It is the third intermediate feature.

[0077] The first KANs network is used to perform nonlinear mapping on the third intermediate feature to obtain a fourth intermediate feature.

[0078] Among them, the specific formula for the nonlinear mapping of the third intermediate feature by the first KANs network is:

[0079]

[0080] Among them, KAN(x) is the fourth intermediate feature, Φ is the external function, φ is the univariate function (B-spline curve function), vector x is the third intermediate feature, x p represents the pth element of vector x, and defines KAN(x) as the fourth intermediate feature y ms-kan .

[0081] The second KANs network is used to perform nonlinear mapping on the motor temperature signal after Lagrangian interpolation processing to obtain the fifth intermediate feature.

[0082] Among them, for the motor temperature signal T(t), Lagrange interpolation is performed on it, so that the motor temperature signal T(t) is transformed from several temperature value points into a signal T with a relatively high frequency. L (t). and make T L (t) After the second KANs network, nonlinear mapping is performed to obtain the fifth intermediate feature y TL .

[0083] The second feature cascade layer is used to: fuse the fourth intermediate feature with the fifth intermediate feature to obtain the motor state information feature.

[0084] Among them, the fourth intermediate feature and the fifth intermediate feature are cascaded (fused) to obtain the motor state information feature y d .

[0085] The above-mentioned further technical solution uses LSTM and 1D-ResNet to comprehensively capture data features in different dimensions through a multi-scale feature mining method: LSTM is used to realize the long-term dependency capture of multi-channel time series features, and 1D-ResNet is used to realize more refined time scale change characteristics. At the same time, it alleviates the contradiction between complex dependency processing and fast feature extraction in the cloud-edge fault diagnosis model, maintains the complexity of the model while having high computational efficiency; and optimizes the feature fusion process through the high nonlinearity and parameter efficiency of KANs; this method provides powerful multi-scale feature mining capabilities, providing a solid foundation for new energy vehicle motor fault diagnosis.

[0086] In an optional manner, the knowledge-data fusion feature extraction module is: a Transformer network including an encoder structure and a decoder structure.

[0087] The Transformer network is used to obtain fused features based on the motor state information features and knowledge item embedding features, and perform linear mapping and softmax activation on the fused features to obtain the motor fault diagnosis results.

[0088] Among them, the motor state information feature y d As the input of the encoder, the knowledge item embedding feature X K ={x k1 ,L,x kn}As the sample label input of the decoder structure, the encoder-decoder structure of the Transformer network enables the model to not only learn the patterns in the state signal, but also to accurately diagnose faults based on the information in the knowledge items. The Transformer network can better understand the semantics and contextual information behind the motor fault, thereby establishing a more effective connection between the two. In addition, by combining features from different domains (i.e., motor state and knowledge items), the Transformer network can more comprehensively learn and understand different types of data representations, enhancing its generalization ability on unknown data. Finally, the fused features are linearly mapped and softmax activated in the decoder structure to obtain the fault diagnosis results, realizing the multi-scale hybrid neural network new energy vehicle motor fault diagnosis with the introduction of external knowledge.

[0089] The above further technical solution uses the encoder-decoder structure of the Transformer network to capture the state signal data and knowledge entry data respectively, effectively improving the efficiency of information fusion, optimizing the accuracy of new energy vehicle motor fault diagnosis, and improving the generalization ability of the model. The use of the Transformer network can effectively combine the real-time motor status with text data such as maintenance history, enhance the accuracy and interpretability of the diagnosis results, not only promote the deep integration of data features, but also provide support for technological progress in the field of product operation and maintenance, and enhance the breadth and practical value of application.

[0090] Figure 4 The flowchart of the second embodiment of the new energy vehicle motor fault diagnosis method provided by the present application is shown, and the method is executed by the new energy vehicle motor fault diagnosis device. Figure 4 As shown, the method comprises the following steps:

[0091] Step S210: extracting knowledge from an external knowledge base of motor operation and maintenance of the motor, obtaining a plurality of knowledge item data corresponding to the motor, and constructing a mapping relationship between the motor state signal and the knowledge item data.

[0092] Among them, the external knowledge base of motor operation and maintenance includes: motor failure mechanism, historical failure data and maintenance records. Knowledge extraction is performed using frameworks such as DeepKE, OpenNRE, DeepDive and UIE to obtain multiple knowledge entry data corresponding to the motor. Figure 5As shown in the figure, the process of knowledge extraction includes entity recognition, relationship extraction, entity alignment and entity fusion. Each knowledge item data includes: phenomenon, characteristics, causes, impacts and maintenance methods of a specific fault. The knowledge item data is represented by graph-structured motor fault knowledge triples.

[0093] Each motor state signal corresponds to a piece of knowledge entry data. The decision tree algorithm is used to construct the mapping relationship between the motor state signal and the knowledge entry data. Each node represents a decision point for a feature, and the data is segmented according to the different values ​​of the feature; the leaf node is the end of the decision tree and represents a predicted output, that is, the knowledge entry data corresponding to the input feature (motor state signal).

[0094] It should be noted that the mapping relationship between the motor status signal and the knowledge entry data is an abstract relationship. In existing motor operation and maintenance cases, engineers or workers will make a comprehensive judgment on the motor and derive the current fault category of the motor. The fault category here is the knowledge side of the mapping relationship (the fault category of the specific fault corresponding to the knowledge entry data); the obtained motor status signal is the signal side of the mapping relationship.

[0095] Step S220: obtaining a target motor state signal of the motor of the new energy vehicle in a target period, and determining target knowledge item data corresponding to the target motor state signal.

[0096] Step S230: input the target motor state signal and the target knowledge item data into the trained motor fault diagnosis model to obtain the motor fault diagnosis result.

[0097] The technical solution of this embodiment further combines the mapping relationship between the motor status signal and the knowledge entry data, which can effectively combine the real-time device status with the knowledge entry, enhance the accuracy and interpretability of the diagnosis results, and optimize the accuracy of new energy vehicle motor fault diagnosis while improving the efficiency of information fusion.

[0098] Figure 6 The schematic diagram of the structure of the embodiment of the new energy vehicle motor fault diagnosis device provided by the present application is shown. Figure 6 As shown, the device 300 includes: a processing unit 310 and an operating unit 320.

[0099] The processing unit 310 is used to obtain a target motor state signal of a motor of a new energy vehicle in a target period of time, and determine target knowledge item data corresponding to the target motor state signal;

[0100] The operation unit 320 is used to input the target motor state signal and the target knowledge item data into the trained motor fault diagnosis model to obtain the motor fault diagnosis result; wherein, the motor fault diagnosis model includes: a multi-scale state signal extraction module and a knowledge-data fusion feature extraction module; the multi-scale state signal extraction module is used to perform multi-scale feature extraction on the target motor state signal to obtain the motor state information feature; the knowledge-data fusion feature extraction module is used to: obtain the fusion feature according to the knowledge item embedding feature corresponding to the motor state information feature and the target knowledge item, and perform linear mapping and softmax activation to obtain the motor fault diagnosis result.

[0101] In an optional manner, the multi-scale state signal extraction module includes: a long short-term memory network, a 1D-Resnet network, a first KANs network, a second KANs network, a first feature cascade layer, and a second feature cascade layer; the target motor state signal includes: a motor high-frequency signal including a motor current signal, a motor vibration signal, and a motor speed signal, and a motor temperature signal;

[0102] The long short-term memory network is used to: dynamically extract long-distance dependent information from the motor high-frequency signal to obtain the first intermediate feature;

[0103] The 1D-Resnet network is used to extract features at different levels of the motor high-frequency signal to obtain the second intermediate features;

[0104] The first feature cascade layer is used to: fuse the first intermediate feature with the second intermediate feature to obtain a third intermediate feature;

[0105] The first KANs network is used to: perform nonlinear mapping on the third intermediate feature to obtain the fourth intermediate feature;

[0106] The second KANs network is used to: perform nonlinear mapping on the motor temperature signal after Lagrangian interpolation processing to obtain the fifth intermediate feature;

[0107] The second feature cascade layer is used to: fuse the fourth intermediate feature with the fifth intermediate feature to obtain the motor state information feature.

[0108] In an optional manner, the knowledge-data fusion feature extraction module is a Transformer network including an encoder structure and a decoder structure;

[0109] The Transformer network is used to obtain fused features based on the motor state information features and knowledge item embedding features, and perform linear mapping and softmax activation on the fused features to obtain the motor fault diagnosis results.

[0110] In an optional manner, the apparatus 300 further includes: a construction unit; the construction unit is used to:

[0111] Knowledge extraction is performed on the motor operation and maintenance external knowledge base of the motor to obtain multiple knowledge item data corresponding to the motor, and a mapping relationship between the motor state signal and the knowledge item data is constructed; wherein each motor state signal corresponds to one knowledge item data.

[0112] In an optional manner, the knowledge item data includes: phenomena, characteristics, causes, impacts and repair methods of a specific fault.

[0113] The technical solution of this embodiment can effectively combine the real-time device status with knowledge items, enhance the accuracy and interpretability of the diagnosis results, and optimize the accuracy of new energy vehicle motor fault diagnosis while improving the efficiency of information fusion.

[0114] It should be noted that the new energy vehicle motor fault diagnosis device provided in the above embodiment and the new energy vehicle motor fault diagnosis method provided in the above embodiment belong to the same concept, and the specific way in which each module and unit performs the operation has been described in detail in the method embodiment and will not be repeated here.

[0115] Figure 7 A structural schematic diagram of an embodiment of a new energy vehicle motor fault diagnosis device provided in the present application is shown, which shows a structural schematic diagram of a computer system suitable for implementing the new energy vehicle motor fault diagnosis device of the embodiment of the present application. The specific embodiment of the present application does not limit the specific implementation of the new energy vehicle motor fault diagnosis device.

[0116] See also Figure 7 As shown, the new energy vehicle motor fault diagnosis device includes: a controller; a memory for storing one or more programs, and when the one or more programs are executed by the controller, the above-mentioned new energy vehicle motor fault diagnosis method is executed.

[0117] Please continue reading Figure 7As shown, the computer system 500 of the new energy vehicle motor fault diagnosis device includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 502 or the program loaded from the storage part 508 to the random access memory (RAM) 503, such as executing the method in the above embodiment. In RAM 503, various programs and data required for system operation are also stored. CPU 501, ROM 502 and RAM 503 are connected to each other through bus 504. Input / output (I / O) interface 505 is also connected to bus 504.

[0118] The following components are connected to the I / O interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed so that a computer program read therefrom is installed into the storage section 508 as needed.

[0119] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section 509, and / or installed from a removable medium 511. When the computer program is executed by a central processing unit (CPU) 501, various functions defined in the system of the present application are executed.

[0120] Another aspect of the present application also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the new energy vehicle motor fault diagnosis method described above is implemented. The computer-readable storage medium may be included in the new energy vehicle motor fault diagnosis device described in the above embodiment, or may exist independently without being assembled into the electronic device.

[0121] Another aspect of the present application also provides a computer program product or a computer program, which includes at least one executable instruction. When the executable instruction is run on a new energy vehicle motor fault diagnosis device / equipment, the new energy vehicle motor fault diagnosis device / equipment executes the new energy vehicle motor fault diagnosis method as described above.

[0122] The executable instructions can be specifically used to enable the new energy vehicle motor fault diagnosis device / apparatus to perform the following operations:

[0123] Obtaining a target motor state signal of a motor of a new energy vehicle in a target period of time, and determining target knowledge item data corresponding to the target motor state signal;

[0124] The target motor state signal and the target knowledge item data are input into the trained motor fault diagnosis model to obtain the motor fault diagnosis result; wherein, the motor fault diagnosis model includes: a multi-scale state signal extraction module and a knowledge-data fusion feature extraction module; the multi-scale state signal extraction module is used to perform multi-scale feature extraction on the target motor state signal to obtain the motor state information feature; the knowledge-data fusion feature extraction module is used to: obtain the fusion feature according to the knowledge item embedding feature corresponding to the motor state information feature and the target knowledge item, perform linear mapping and softmax activation, and obtain the motor fault diagnosis result.

[0125] In an optional manner, the multi-scale state signal extraction module includes: a long short-term memory network, a 1D-Resnet network, a first KANs network, a second KANs network, a first feature cascade layer, and a second feature cascade layer; the target motor state signal includes: a motor high-frequency signal including a motor current signal, a motor vibration signal, and a motor speed signal, and a motor temperature signal;

[0126] The long short-term memory network is used to: dynamically extract long-distance dependent information from the motor high-frequency signal to obtain the first intermediate feature;

[0127] The 1D-Resnet network is used to extract features at different levels of the motor high-frequency signal to obtain the second intermediate features;

[0128] The first feature cascade layer is used to: fuse the first intermediate feature with the second intermediate feature to obtain a third intermediate feature;

[0129] The first KANs network is used to: perform nonlinear mapping on the third intermediate feature to obtain the fourth intermediate feature;

[0130] The second KANs network is used to: perform nonlinear mapping on the motor temperature signal after Lagrangian interpolation processing to obtain the fifth intermediate feature;

[0131] The second feature cascade layer is used to: fuse the fourth intermediate feature with the fifth intermediate feature to obtain the motor state information feature.

[0132] In an optional manner, the knowledge-data fusion feature extraction module is a Transformer network including an encoder structure and a decoder structure;

[0133] The Transformer network is used to obtain fused features based on the motor state information features and knowledge item embedding features, and perform linear mapping and softmax activation on the fused features to obtain the motor fault diagnosis results.

[0134] In an optional manner, the method further includes:

[0135] Knowledge extraction is performed on the motor operation and maintenance external knowledge base of the motor to obtain multiple knowledge item data corresponding to the motor, and a mapping relationship between the motor state signal and the knowledge item data is constructed; wherein each motor state signal corresponds to one knowledge item data.

[0136] In an optional manner, the knowledge item data includes: phenomena, characteristics, causes, impacts and repair methods of a specific fault.

[0137] The technical solution of this embodiment can effectively combine the real-time device status with knowledge items, enhance the accuracy and interpretability of the diagnosis results, and optimize the accuracy of new energy vehicle motor fault diagnosis while improving the efficiency of information fusion.

[0138] It should be noted that the computer-readable medium shown in the embodiment of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it. In the present application, a computer-readable signal medium can include a data signal propagated in a baseband or as a part of a carrier wave, wherein a computer-readable computer program is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0139] The flowchart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present application. Wherein, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0140] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. The names of these units do not, in some cases, constitute limitations on the units themselves.

[0141] According to one aspect of an embodiment of the present application, a computer system is also provided, including a central processing unit (CPU), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) or a program loaded from a storage portion into a random access memory (RAM), such as executing the method in the above embodiment. In RAM, various programs and data required for system operation are also stored. CPU, ROM and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0142] The following components are connected to the I / O interface: an input part including a keyboard, a mouse, etc.; an output part including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage part including a hard disk, etc.; and a communication part including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication part performs communication processing via a network such as the Internet. A drive is also connected to the I / O interface as needed. Removable media, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., are installed on the drive as needed so that the computer program read therefrom is installed into the storage part as needed.

[0143] The above content is only a preferred exemplary embodiment of the present application and is not intended to limit the implementation scheme of the present application. A person skilled in the art can easily make corresponding changes or modifications based on the main concept and spirit of the present application. Therefore, the scope of protection of the present application shall be based on the scope of protection required by the claims.

Claims

1. A method for diagnosing faults of motors of new energy vehicles, characterized in that: The method comprises: Acquire a target motor state signal of a motor of a new energy vehicle in a target period, and determine target knowledge item data corresponding to the target motor state signal; The target motor state signal and the target knowledge item data are input into a trained motor fault diagnosis model to obtain a fault diagnosis result of the motor; wherein the motor fault diagnosis model comprises: a multi-scale state signal extraction module and a knowledge-data fusion feature extraction module; the multi-scale state signal extraction module is used to perform multi-scale feature extraction on the target motor state signal to obtain motor state information features; the knowledge-data fusion feature extraction module is used to: obtain fusion features based on the knowledge item embedding features corresponding to the motor state information features and the target knowledge items, perform linear mapping and softmax activation, and obtain the fault diagnosis result of the motor.

2. The method according to claim 1, characterized in that The multi-scale state signal extraction module includes: a long short-term memory network, a 1D-Resnet network, a first KANs network, a second KANs network, a first feature cascade layer, and a second feature cascade layer; the target motor state signal includes: a motor high-frequency signal including a motor current signal, a motor vibration signal, and a motor speed signal, and a motor temperature signal; The long short-term memory network is used to: dynamically extract long-distance dependency information of the motor high-frequency signal to obtain a first intermediate feature; The 1D-Resnet network is used to: extract features of different levels from the motor high-frequency signal to obtain a second intermediate feature; The first feature cascade layer is used to: perform feature fusion on the first intermediate feature and the second intermediate feature to obtain a third intermediate feature; The first KANs network is used to: perform nonlinear mapping on the third intermediate feature to obtain a fourth intermediate feature; The second KANs network is used to: perform nonlinear mapping on the motor temperature signal after Lagrange interpolation processing to obtain a fifth intermediate feature; The second feature cascade layer is used to: perform feature fusion on the fourth intermediate feature and the fifth intermediate feature to obtain the motor state information feature.

3. The method according to claim 2, characterized in that The knowledge-data fusion feature extraction module is a Transformer network including an encoder structure and a decoder structure; The Transformer network is used to obtain the fusion feature according to the motor state information feature and the knowledge item embedding feature, and perform linear mapping and softmax activation on the fusion feature to obtain the fault diagnosis result of the motor.

4. The method according to claim 1, characterized in that The method further comprises: Knowledge extraction is performed on the motor operation and maintenance external knowledge base of the motor to obtain a plurality of knowledge item data corresponding to the motor, and a mapping relationship between the motor state signal and the knowledge item data is constructed; wherein each motor state signal corresponds to one knowledge item data.

5. The method according to claim 4, characterized in that The knowledge item data includes: the phenomenon, characteristics, causes, impacts and maintenance methods of a specific fault.

6. A new energy vehicle motor fault diagnosis device, characterized in that: The device comprises: A processing unit, used to obtain a target motor state signal of a motor of a new energy vehicle in a target period of time, and determine target knowledge item data corresponding to the target motor state signal; An operating unit is used to input the target motor state signal and the target knowledge item data into a trained motor fault diagnosis model to obtain a fault diagnosis result of the motor; wherein the motor fault diagnosis model includes: a multi-scale state signal extraction module and a knowledge-data fusion feature extraction module; the multi-scale state signal extraction module is used to perform multi-scale feature extraction on the target motor state signal to obtain motor state information features; the knowledge-data fusion feature extraction module is used to: obtain fusion features based on the motor state information features and the knowledge item embedding features corresponding to the target knowledge items, perform linear mapping and softmax activation, and obtain the fault diagnosis result of the motor.

7. The device according to claim 6, characterized in that The multi-scale state signal extraction module includes: a long short-term memory network, a 1D-Resnet network, a first KANs network, a second KANs network, a first feature cascade layer, and a second feature cascade layer; the target motor state signal includes: a motor high-frequency signal including a motor current signal, a motor vibration signal, and a motor speed signal, and a motor temperature signal; The long short-term memory network is used to: dynamically extract long-distance dependency information of the motor high-frequency signal to obtain a first intermediate feature; The 1D-Resnet network is used to: extract features of different levels from the motor high-frequency signal to obtain a second intermediate feature; The first feature cascade layer is used to: perform feature fusion on the first intermediate feature and the second intermediate feature to obtain a third intermediate feature; The first KANs network is used to: perform nonlinear mapping on the third intermediate feature to obtain a fourth intermediate feature; The second KANs network is used to: perform nonlinear mapping on the motor temperature signal after Lagrange interpolation processing to obtain a fifth intermediate feature; The second feature cascade layer is used to: perform feature fusion on the fourth intermediate feature and the fifth intermediate feature to obtain the motor state information feature.

8. The device according to claim 7, characterized in that The knowledge-data fusion feature extraction module is a Transformer network including an encoder structure and a decoder structure; The Transformer network is used to obtain the fusion feature according to the motor state information feature and the knowledge item embedding feature, and perform linear mapping and softmax activation on the fusion feature to obtain the fault diagnosis result of the motor.

9. A new energy vehicle motor fault diagnosis device, characterized in that: include: Controller; A memory is used to store one or more programs. When the one or more programs are executed by the controller, the controller implements the new energy vehicle motor fault diagnosis method described in any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that: The storage medium stores at least one executable instruction. When the executable instruction is executed on the new energy vehicle motor fault diagnosis device / equipment, the new energy vehicle motor fault diagnosis device / equipment performs the operation of the new energy vehicle motor fault diagnosis method as described in any one of claims 1-5.

Images