Method and device for identifying low-current single-phase earth fault of power distribution network based on cloud-side cooperation

Abstract

The embodiment of the invention provides a cloud-side collaborative power distribution network low-current single-phase earth fault identification method and device, and relates to the technical field of power distribution Internet of Things. The method comprises the following steps: acquiring a trained teacher model based on a neural network and used for small-current single-phase earth fault identification; performing knowledge distillation on the teacher model by adopting a multi-granularity feature distillation algorithm, and transmitting the knowledge distillation to a student model; and deploying the student model to an edge side of a cloud edge system, wherein the student model outputs a fault identification result based on data acquired by the edge side. The low-current single-phase earth fault identification method provided by the embodiment of the invention has the advantages of accurate identification and quick response.

Description

Technical Field

[0001] The present application relates to the technical field of power distribution Internet of Things, and specifically to a cloud-edge collaborative distribution network small current single-phase grounding fault identification method, a cloud-edge collaborative distribution network small current single-phase grounding fault identification device, an electronic device, a storage medium and a computer program product. Background Art

[0002] Low-current grounding systems are widely used in medium and low-voltage distribution networks. Single-phase grounding faults are prone to occur. While they can maintain power supply for a short period of time, persistent faults can increase the voltage in the non-fault phase, threatening equipment insulation and potentially causing serious accidents such as phase-to-phase short circuits. The overvoltage generated by arc grounding can also damage power equipment, impacting the safe and stable operation of the power grid. Accurate and rapid fault identification and resolution are essential to reduce accident losses and improve power supply reliability. Currently, deep learning models demonstrate high fault identification accuracy in the field of low-current single-phase grounding fault identification, leveraging their powerful feature extraction and pattern recognition capabilities. However, in the face of complex fault scenarios and massive amounts of data, generalization and computational efficiency still need to be improved. Furthermore, complex models are difficult to deploy on resource-limited edge devices in distribution networks. Therefore, research is urgently needed on accurate, rapid, and adaptable fault identification methods for distribution network edge devices.

[0003] Existing methods for identifying low-current single-phase grounding faults in distribution networks are primarily categorized into centralized cloud processing and pure edge computing. Traditional centralized cloud processing collects distribution network fault data and uploads it to the cloud, where it uses deep learning or expert systems for fault analysis. While this method achieves high-precision diagnosis, it relies on stable communication and suffers from high latency (typically >500ms) due to large data volumes, making it difficult to meet real-time requirements. Pure edge computing deploys lightweight algorithms on terminal devices, achieving millisecond-level responses (<50ms). However, due to local computing power limitations, it cannot handle complex faults (such as intermittent arc faults), lacks global topology analysis, and has a high false alarm rate. Summary of the Invention

[0004] The purpose of the embodiments of the present application is to provide a cloud-edge collaborative distribution network small current single-phase grounding fault identification method and device, and to train and deploy fault identification teacher and student models through cloud-edge collaboration to meet the requirements of accurate and fast identification of small current single-phase grounding faults in the distribution network and adapt to the computing resources of the distribution network edge devices, so as to at least solve some of the problems in the background technology.

[0005] In order to achieve the above-mentioned objectives, the present application provides a cloud-edge collaborative distribution network small current single-phase grounding fault identification method, including: a cloud-edge collaborative distribution network small current single-phase grounding fault identification method, the method comprising: obtaining a trained neural network-based teacher model for small current single-phase grounding fault identification; using a multi-granularity feature distillation algorithm to perform knowledge distillation on the teacher model and transfer it to a student model; deploying the student model to the edge side of the cloud-edge system, and the student model outputs a fault identification result based on the data obtained on the edge side.

[0006] Optionally, the trained neural network-based teacher model for identifying small current single-phase grounding faults is deployed on the cloud side of the cloud-edge system.

[0007] Optionally, the multi-granularity features include: original signal features, abstract semantic features and decision features; a multi-granularity feature distillation algorithm is used to perform knowledge distillation on the teacher model and transfer it to the student model, including: extracting multi-granularity features from the processing process of the input data by the teacher model and the student model; constructing a distillation loss function based on the multi-granularity features; and updating the parameters of the student model during the training process based on the distillation loss function.

[0008] Optionally, multi-granularity features are extracted from the processing process of the input data by the teacher model and the student model, including: extracting original signal features from the input layer of the teacher model and the input layer of the student model respectively; extracting abstract semantic features from the hidden layer of the teacher model and the hidden layer of the student model respectively; and extracting decision features from the output layer of the teacher model and the output layer of the student model respectively.

[0009] Optionally, a distillation loss function is constructed based on the multi-granularity features, including: constructing an input layer network loss function based on the morphological consistency between the original signal features extracted from the teacher model and the original signal features extracted from the student model; constructing a hidden layer network loss function based on the similarity measure between the abstract semantic features extracted from the teacher model and the abstract semantic features extracted from the student model; constructing an output layer network loss function based on the probability classification loss and probability distribution difference between the decision features extracted from the teacher model and the decision features extracted from the student model; and obtaining the distillation loss function by combining the input layer network loss function, the hidden layer network loss function and the output layer network loss function.

[0010] Optionally, an input layer network loss function is constructed based on the morphological consistency between the original signal features extracted from the teacher model and the original signal features extracted from the student model, including: using the mean square error to measure the morphological consistency of the original signal features; and obtaining the input layer network loss function based on the sum of the squares of the differences between corresponding elements of the input layer feature maps of the teacher model and the student model.

[0011] Optionally, a hidden layer network loss function is constructed based on a similarity measurement between the abstract semantic features extracted from the teacher model and the abstract semantic features extracted from the student model, including: using cosine similarity to measure the similarity of the abstract semantic features; and obtaining the input layer network loss function based on a cosine value of the angle between the vectors formed by corresponding elements of the hidden layer feature maps of the teacher model and the student model.

[0012] Optionally, an output layer network loss function is constructed based on the probability classification loss and probability distribution difference of the decision features extracted from the teacher model and the decision features extracted from the student model, including: using the cross entropy loss of the classification result as the probability classification loss; using the KL divergence loss of the classification result as the probability distribution difference; and using the weighted sum of the probability classification loss and the probability distribution difference as the output layer network loss function.

[0013] Optionally, the distillation loss function is obtained by integrating the input layer network loss function, the hidden layer network loss function and the output layer network loss function, including: obtaining the distillation loss function by weighted sum according to the respective weights of the input layer network loss function, the hidden layer network loss function and the output layer network loss function.

[0014] Optionally, updating the parameters of the student model during the training process according to the distillation loss function includes: adopting the distillation loss function as the loss function of the student model; and updating the parameters of the student model during the training process through the distillation loss function and the back propagation algorithm.

[0015] Optionally, after the student model outputs a fault identification result based on the data obtained from the edge side, the method further includes: uploading the fault identification result to the cloud; obtaining a response instruction from the cloud based on the fault identification result; and conveying a control instruction to the terminal based on the response instruction.

[0016] Optionally, after the student model outputs a fault identification result based on the data obtained on the edge side, the method further includes: caching the data obtained on the edge side, and regularly uploading the cached data to the cloud as new training samples.

[0017] Optionally, the method also includes: using the new training samples to train and update the neural network-based teacher model for small current single-phase grounding fault identification to obtain an updated teacher model; performing knowledge distillation based on the updated teacher model and transferring it to the student model to obtain an updated student model; and deploying the updated student model to the edge side of the cloud-edge system.

[0018] This application also provides a cloud-edge collaborative distribution network small current single-phase grounding fault identification device, which includes: a model acquisition module for obtaining a trained neural network-based teacher model for small current single-phase grounding fault identification; a knowledge distillation module for using a multi-granularity feature distillation algorithm to perform knowledge distillation on the teacher model and transfer it to a student model; and a model deployment module for deploying the student model to the edge side of the cloud-edge system, and the student model outputs a fault identification result based on the data obtained on the edge side.

[0019] This application also provides an electronic device, comprising: at least one processor; a memory connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the at least one processor implements the aforementioned cloud-edge collaborative distribution network small current single-phase grounding fault identification method by executing the instructions stored in the memory.

[0020] This application also provides a machine-readable storage medium, which stores instructions. When the instructions are executed by a processor, the processor is configured to execute the aforementioned cloud-edge collaborative distribution network small current single-phase grounding fault identification method.

[0021] A computer program product is also provided in the present application, including a computer program, which, when executed by a processor, implements the aforementioned cloud-edge collaborative distribution network small current single-phase grounding fault identification method.

[0022] The above technical solution has the following beneficial effects: (1) An innovative cloud-edge collaborative small current single-phase grounding fault identification model training and deployment method was proposed, which reduced the interactive delay of fault identification, improved the efficiency of massive data transmission, and achieved fault identification in seconds and global analysis optimization.

[0023] (2) An innovative multi-granularity feature distillation algorithm was proposed to improve the learning ability of the student model and ensure the recognition accuracy of the edge-side model.

[0024] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present application but do not constitute a limitation on the embodiments of the present application. In the accompanying drawings: Figure 1 The following schematically illustrates the steps of a method for identifying a small current single-phase grounding fault in a distribution network using cloud-edge collaboration according to an embodiment of the present application; Figure 2 Schematically shows an implementation diagram of the method for identifying small current single-phase grounding faults in a distribution network using cloud-edge collaboration according to an embodiment of the present application; Figure 3 The following schematically shows a schematic diagram of the structure of a cloud-edge-coordinated distribution network small current single-phase grounding fault identification device according to an embodiment of the present application; Figure 4 The internal structure of an electronic device according to an embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0026] The following describes the specific implementation of the embodiment of the present application in detail with reference to the accompanying drawings. It should be understood that the specific implementation described herein is only used to illustrate and explain the embodiment of the present application and is not intended to limit the embodiment of the present application.

[0027] Figure 1 The following schematically shows the steps of the method for identifying a small current single-phase grounding fault in a distribution network using cloud-edge collaboration according to an embodiment of the present application. Figure 1 As shown, a cloud-edge collaborative distribution network small current single-phase grounding fault identification method includes: S01. Obtain a trained neural network-based teacher model for identifying small current single-phase grounding faults; S02, using a multi-granularity feature distillation algorithm to perform knowledge distillation on the teacher model and transfer it to the student model; S03. Deploy the student model to the edge side of the cloud-edge system, and the student model outputs a fault identification result based on the data obtained from the edge side.

[0028] Through the above implementation method, a high-performance but complex deep neural network teacher model is first trained; then the large deep neural network model is distilled using a multi-granularity feature distillation algorithm to generate a student model with a simple structure and low computational complexity, and the recognition effect of the distilled small model is guaranteed to be infinitely close to that of the large model; finally, the distilled model is distributed and deployed on the edge side of the distribution Internet of Things with limited computing resources, and data processing and recognition are performed directly on the edge, which improves the efficiency of processing massive data for identifying small-current single-phase grounding faults in the distribution network and reduces interaction delays.

[0029] In some embodiments of the present application, the trained neural network-based teacher model for low-current single-phase grounding fault identification is preferably deployed on the cloud side of the cloud-edge system. Step S03 of deploying the student model to the edge side of the cloud-edge system includes: transmitting the cloud-trained student model to each smart feeder terminal on the edge side via wireless communication or other communication methods, thereby completing the model deployment on the edge side.

[0030] In some embodiments of the present application, the multi-granularity features include: original signal features, abstract semantic features, and decision features; a multi-granularity feature distillation algorithm is used to perform knowledge distillation on the teacher model and transfer it to the student model, including: extracting multi-granularity features from the processing process of the input data by the teacher model and the student model; constructing a distillation loss function based on the multi-granularity features; and updating the parameters of the student model during the training process based on the distillation loss function. Small current single-phase grounding fault data contains rich time domain, frequency domain, and time-frequency domain features. In order to make full use of this information, a multi-granularity feature distillation module is designed. Between different network layers of the teacher model and the student model, the original signal features of the input layer, the abstract semantic features of the hidden layer, and the decision features of the output layer are extracted respectively. This embodiment realizes the precise migration of features at different levels by constructing multiple groups of distillation loss functions.

[0031] In some embodiments of the present application, multi-granularity features are extracted from the processing of input data by the teacher model and the student model, including: extracting original signal features from the input layer of the teacher model and the input layer of the student model respectively; extracting abstract semantic features from the hidden layer of the teacher model and the hidden layer of the student model respectively; extracting decision features from the output layer of the teacher model and the output layer of the student model respectively. The extraction here includes the classification of the model function layers of the teacher model and the student model, and taking the output of a certain type of function layer as the corresponding original signal features, abstract semantic features and decision features. Specifically, the input layer features refer to the basic information retained in the original signal feature map extracted after the input convolution layers of the teacher model and the student model process the input fault data (pre-processed zero-sequence voltage signal), which helps the student model learn the basic form of the fault signal, such as amplitude change, rising and falling trends, etc. Suppose the feature map output by the shallow network of the teacher model is , the feature map output by the shallow network of the student model is The hidden layer features are the semantic features that are initially abstracted after the middle-layer networks of the teacher model and the student model further process the data. They can reflect the essential attributes of the fault, such as specific frequency components and waveform distortion patterns. Suppose the feature map output by the middle-layer network of the teacher model is , the feature map output by the middle network of the student model is The output layer features refer to the deep features related to decision-making extracted by the fully connected layers of the teacher model and the student model, which are closely related to the final decision of the model and guide the student model to imitate the decision logic of the teacher model. Let the feature representation of the deep network output of the teacher model be , the feature representation of the deep network output of the student model is .

[0032] In some embodiments of the present application, an input layer network loss function is constructed based on the morphological consistency between the original signal features extracted from the teacher model and the original signal features extracted from the student model. Specifically, it includes: using the mean square error to measure the morphological consistency of the original signal features; obtaining the input layer network loss function based on the sum of the squares of the differences between the corresponding elements of the input layer feature maps of the teacher model and the student model. Examples of input layer network loss functions are as follows: For shallow features, since the main focus is on the basic morphological consistency of the original signal, the mean square error (MSE) loss function is used, and the input layer network loss function The formula is:

[0033] Where N is the number of elements in the feature map, and The loss function minimizes the sum of squared differences between the input feature maps of the teacher and student models, thereby encouraging the student model to learn the basic waveform information of the original signal in the teacher model.

[0034] On the other hand, a hidden layer network loss function is constructed based on the similarity measurement between the abstract semantic features extracted from the teacher model and the abstract semantic features extracted from the student model. Specifically, it includes: using cosine similarity to measure the similarity of the abstract semantic features; obtaining the input layer network loss function based on the cosine value of the angle between the vectors formed by the corresponding elements of the hidden layer feature maps of the teacher model and the student model. Examples of hidden layer network loss functions are as follows: hidden layer features emphasize the similarity of abstract semantic features more, using cosine similarity loss function, hidden layer network loss function The formula is:

[0035] Where N is the number of elements in the feature map, and are the values ​​of the i-th element in the hidden layer feature maps of the teacher and student models, respectively. Cosine similarity measures the cosine of the angle between two vectors, ranging from -1 to 1. Values ​​closer to 1 indicate greater similarity. This loss function guides the student model to learn the abstract semantic representation of the teacher model's mid-level features, enabling it to capture the essential attributes of fault characteristics.

[0036] On the other hand, the output layer network loss function is constructed based on the probability classification loss and probability distribution difference between the decision features extracted from the teacher model and the decision features extracted from the student model. Specifically, the cross-entropy loss of the classification results is used as the probability classification loss; the KL divergence loss of the classification results is used as the probability distribution difference. KL divergence (Kullback-Leibler divergence) is a metric used to measure the difference between two probability distributions, also known as relative entropy or information divergence. Its theoretical significance lies in that a larger KL divergence indicates a greater difference between the two distributions; a KL divergence of 0 indicates that the two distributions are identical. KL divergence is commonly used in machine learning and statistics to help measure the amount of information lost when selecting an approximate distribution. The output layer network loss function is a weighted sum of the probability classification loss and the probability distribution difference. Output layer network loss function: The output layer features are closely related to the decision, and a combination of cross-entropy loss and KL divergence loss is required.

[0037] Among them, the cross entropy loss function formula is:

[0038] Where C is the number of fault types, is the probability of the c-th type of failure in the true label (usually 0 or 1), The probability of type c failure predicted by the model.

[0039] The KL divergence loss function formula is:

[0040] in, is the probability of the teacher model predicting the c-th type of failure, Predict the probability of type c failure for the student model.

[0041] Output layer network loss function The formula is:

[0042] in, and is a weight coefficient used to adjust the proportion of cross entropy loss and KL divergence loss in the total loss. This loss function encourages the student model to imitate the decision logic of the teacher model, improving the accuracy of fault identification.

[0043] In some embodiments of the present application, the distillation loss function is obtained by synthesizing the input layer network loss function, the hidden layer network loss function and the output layer network loss function, which may include: according to the respective weights of the input layer network loss function, the hidden layer network loss function and the output layer network loss function, the distillation loss function is obtained by adopting a weighted sum method. Distillation loss function for:

[0044] in, 、 and are the weight coefficients of input, hidden and output network losses respectively. Through the back propagation algorithm, according to the total distillation loss Update the parameters of the student model so that the student model gradually approaches the teacher model in terms of multi-level features, thus achieving effective distillation and transfer of knowledge. Updating the parameters of the student model may be by using the distillation loss function as the loss function of the student model. During the training process of the student model, the parameters of the student model are updated using the distillation loss function and a back propagation algorithm.

[0045] In the aforementioned implementation, the architectures of the teacher model and student model for fault identification can be the same or different. However, generally speaking, the teacher model has a more complex architecture, while the student model has a simpler structure and requires less computation. For example, a residual network (ResNet) can be used for both the teacher and student models, but the teacher model uses ResNet18 and the student model uses ResNet8.

[0046] In some embodiments of the present application, after the student model outputs a fault identification result based on data obtained by the edge side, the method further includes: uploading the fault identification result to the cloud; obtaining a response instruction from the cloud based on the fault identification result; and transmitting a control instruction to the terminal based on the response instruction. This embodiment provides a subsequent processing flow for the fault identification result. By uploading the fault identification result to the cloud through a cloud-edge collaboration mechanism, the edge side avoids storing a large amount of fault handling knowledge and treatment measures, allowing the edge side to focus more on fault identification.

[0047] In some embodiments of the present application, after the student model outputs a fault identification result based on the data obtained on the edge side, the method further includes: caching the data obtained on the edge side and periodically uploading the cached data to the cloud as new training samples. This embodiment provides a data collection function on the edge side, providing data samples for updating and iterating the teacher and student models.

[0048] In some implementations of this application, the new training samples are used to train and update a neural network-based teacher model for low-current single-phase ground fault identification, resulting in an updated teacher model. Knowledge distillation is then performed on the updated teacher model and transferred to a student model, resulting in an updated student model. The updated student model is then deployed to the edge of the cloud-edge system. This implementation provides an iterative update mechanism for the teacher and student models, improving the accuracy of online deployment.

[0049] Figure 2 The following schematically shows the implementation diagram of the cloud-edge collaborative distribution network small current single-phase grounding fault identification method according to the embodiment of the present application. Figure 2 As shown in the figure, the specific operation process of deploying a small current single-phase ground fault identification model based on cloud-edge collaboration mainly includes three parts: data storage and upload, model training and distribution, and task identification and response. The results include fault identification results, data includes historical fault data, models include student models, and instructions include issuing instructions based on fault identification results. The detailed steps are as follows: (1) The cloud platform collects data sets based on the history of small current single-phase grounding faults and provides training samples for the original teacher model and student model in the cloud.

[0050] (2) The cloud platform constructs a convolutional neural network, including the selected teacher network and student network. For example, the teacher model for fault identification uses ResNet18, and the student model uses ResNet8. The generated training samples are input into the teacher model, and the teacher model is trained first. The model training process mainly includes: the small current single-phase grounding fault identification problem can be modeled as a deep learning classification problem. The fault types include stable faults, arc faults, high-resistance arc faults, high-resistance discharge faults and gap discharge faults. The zero-sequence voltage transient signal historical data collected by the distribution network when the fault occurs is normalized, filtered and denoised in the edge-side intelligent feeder terminal, and then uploaded to the cloud. The data is then divided into training set, verification set and test set. First, the teacher model ResNet18 is trained, and then the multi-granularity feature distillation algorithm is used according to the training process parameters and results of the teacher model. The loss function guides the training of the student ResNet8 model, resulting in a concise, computationally efficient student model with recognition performance close to that of the teacher model. After training, the model is saved to a cloud server.

[0051] (3) The student model trained in the cloud is sent to each intelligent feeder terminal on the edge side through wireless communication to complete the model deployment on the edge.

[0052] (4) After receiving the terminal data, the intelligent feeder terminal at the edge node performs simple data processing on it, adjusts it into a data format that can be input into the model, and inputs it into the edge side model for fault identification. It also temporarily stores the data and regularly transmits the pre-processed historical fault data to the cloud.

[0053] (5) Operation and maintenance response is carried out according to the model judgment. The response process is as follows: the fault identification results are uploaded to the cloud, and then the cloud issues corresponding instructions, and then the edge side transmits the instructions to the terminal to control the device status.

[0054] (6) Based on the established teacher model, the teacher model and student model are trained and updated using the regularly updated historical fault data, and the updated student model is sent to the edge-side intelligent feeder terminal.

[0055] The above implementation demonstrates that this implementation proposes a multi-granularity feature distillation algorithm, which improves the learning capabilities of the student model and ensures the recognition accuracy of the edge model. Furthermore, the training and deployment method for a low-current single-phase grounding fault identification model based on this cloud-edge collaboration reduces fault identification interaction latency, improves the efficiency of massive data transmission, and achieves fault identification within seconds and global analysis optimization.

[0056] Based on the same inventive concept, this application also provides a cloud-edge collaborative distribution network small current single-phase grounding fault identification device, Figure 3 The schematic diagram shows the structure of the cloud-edge collaborative distribution network small current single-phase grounding fault identification device according to the embodiment of the present application. Figure 3 As shown, the device includes: a model acquisition module for acquiring a trained neural network-based teacher model for identifying small current single-phase grounding faults; a knowledge distillation module for performing knowledge distillation on the teacher model using a multi-granularity feature distillation algorithm and transferring the knowledge to a student model; and a model deployment module for deploying the student model to the edge side of the cloud-edge system, wherein the student model outputs a fault identification result based on the data acquired on the edge side.

[0057] In some optional implementations, the trained neural network-based teacher model for identifying small current single-phase grounding faults is deployed on the cloud side of the cloud-edge system.

[0058] In some optional embodiments, the multi-granularity features include: original signal features, abstract semantic features, and decision features; a multi-granularity feature distillation algorithm is used to perform knowledge distillation on the teacher model and transfer it to the student model, including: extracting multi-granularity features from the processing process of the input data by the teacher model and the student model; constructing a distillation loss function based on the multi-granularity features; and updating the parameters of the student model during the training process based on the distillation loss function.

[0059] In some optional embodiments, multi-granularity features are extracted from the processing of input data by the teacher model and the student model, including: extracting original signal features from the input layer of the teacher model and the input layer of the student model respectively; extracting abstract semantic features from the hidden layer of the teacher model and the hidden layer of the student model respectively; and extracting decision features from the output layer of the teacher model and the output layer of the student model respectively.

[0060] In some optional embodiments, a distillation loss function is constructed based on the multi-granularity features, including: constructing an input layer network loss function based on the morphological consistency between the original signal features extracted from the teacher model and the original signal features extracted from the student model; constructing a hidden layer network loss function based on the similarity measure between the abstract semantic features extracted from the teacher model and the abstract semantic features extracted from the student model; constructing an output layer network loss function based on the probability classification loss and probability distribution difference between the decision features extracted from the teacher model and the decision features extracted from the student model; and obtaining the distillation loss function by combining the input layer network loss function, the hidden layer network loss function and the output layer network loss function.

[0061] In some optional embodiments, an input layer network loss function is constructed based on the morphological consistency between the original signal features extracted from the teacher model and the original signal features extracted from the student model, including: using the mean square error to measure the morphological consistency of the original signal features; and obtaining the input layer network loss function based on the sum of the squares of the differences between corresponding elements of the input layer feature maps of the teacher model and the student model.

[0062] In some optional embodiments, a hidden layer network loss function is constructed based on a similarity measurement between the abstract semantic features extracted from the teacher model and the abstract semantic features extracted from the student model, including: using cosine similarity to measure the similarity of the abstract semantic features; and obtaining the input layer network loss function based on a cosine value of the angle between the vectors formed by corresponding elements of the hidden layer feature maps of the teacher model and the student model.

[0063] In some optional embodiments, an output layer network loss function is constructed based on the probability classification loss and probability distribution difference of the decision features extracted from the teacher model and the decision features extracted from the student model, including: using the cross entropy loss of the classification result as the probability classification loss; using the KL divergence loss of the classification result as the probability distribution difference; and using the weighted sum of the probability classification loss and the probability distribution difference as the output layer network loss function.

[0064] In some optional embodiments, the distillation loss function is obtained by integrating the input layer network loss function, the hidden layer network loss function and the output layer network loss function, including: obtaining the distillation loss function by weighted sum according to the respective weights of the input layer network loss function, the hidden layer network loss function and the output layer network loss function.

[0065] In some optional embodiments, the parameters of the student model are updated during the training process according to the distillation loss function, including: using the distillation loss function as the loss function of the student model; and updating the parameters of the student model during the training process through the distillation loss function and the back propagation algorithm.

[0066] In some optional embodiments, after the student model outputs a fault identification result based on the data obtained from the edge side, the method further includes: uploading the fault identification result to the cloud; obtaining a response instruction from the cloud based on the fault identification result; and conveying a control instruction to the terminal based on the response instruction.

[0067] In some optional embodiments, after the student model outputs a fault identification result based on the data obtained on the edge side, the method further includes: caching the data obtained on the edge side, and regularly uploading the cached data to the cloud as new training samples.

[0068] In some optional embodiments, the method further includes: using the new training samples to train and update a neural network-based teacher model for identifying small current single-phase grounding faults to obtain an updated teacher model; performing knowledge distillation based on the updated teacher model and transferring it to a student model to obtain an updated student model; and deploying the updated student model to the edge side of the cloud-edge system.

[0069] The specific definitions of the various functional modules in the above-mentioned cloud-edge collaborative distribution network small current single-phase grounding fault identification device can be found in the above-mentioned definition of the cloud-edge collaborative distribution network small current single-phase grounding fault identification method, which will not be repeated here. Each module in the above-mentioned system can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules. It also achieves the reduction of the interactive delay of fault identification, improves the efficiency of massive data transmission, and achieves the beneficial effects of second-level fault identification and global analysis optimization.

[0070] In some embodiments of the present application, an electronic device is further provided, comprising: at least one processor; a memory connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the at least one processor executes the aforementioned cloud-edge collaborative distribution network small current single-phase grounding fault identification method. Its internal structure diagram can be as follows Figure 4 shown. Figure 4 The internal structure diagram of an electronic device according to an embodiment of the present application is schematically shown. The electronic device includes a processor A01, a network interface A02, a memory (not shown in the figure) and a database (not shown in the figure) connected via a system bus. Among them, the processor A01 of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes an internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02 and a database (not shown in the figure). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 in the non-volatile storage medium A04. The network interface A02 of the electronic device is used to communicate with an external terminal through a network connection. When the computer program B02 is executed by the processor A01, a cloud-edge collaborative distribution network small current single-phase grounding fault identification method is implemented.

[0071] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the electronic device to which the solution of the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0072] In one embodiment provided in the present application, a machine-readable storage medium is provided, on which instructions are stored. When the instructions are executed by a processor, the processor is configured to execute the aforementioned cloud-edge collaborative distribution network small current single-phase grounding fault identification method.

[0073] In one embodiment provided in the present application, a computer program product is provided, including a computer program, which, when executed by a processor, implements the aforementioned cloud-edge collaborative distribution network small current single-phase grounding fault identification method.

[0074] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0075] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0076] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0077] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0078] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0079] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0080] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can be implemented using any method or technology to store information. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change RAM (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0081] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0082] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A cloud-edge collaborative distribution network small current single-phase grounding fault identification method, characterized in that: The method comprises: Obtain a trained neural network-based teacher model for low-current single-phase grounding fault identification; A multi-granularity feature distillation algorithm is used to perform knowledge distillation on the teacher model and transfer the knowledge to the student model; The student model is deployed to the edge side of the cloud-edge system, and the student model outputs a fault identification result based on the data obtained from the edge side.

2. The method according to claim 1, characterized in that The trained neural network-based teacher model for identifying small current single-phase grounding faults is deployed on the cloud side of the cloud-edge system.

3. The method according to claim 1, characterized in that The multi-granularity features include: original signal features, abstract semantic features and decision features; A multi-granularity feature distillation algorithm is used to distill the knowledge of the teacher model and transfer it to the student model, including: Extracting multi-granularity features from the processing of input data by the teacher model and the student model; Constructing a distillation loss function based on the multi-granularity features; The parameters of the student model are updated during the training process according to the distillation loss function.

4. The method according to claim 3, characterized in that Extracting multi-granularity features from the processing of input data by the teacher model and the student model includes: Extracting original signal features from the input layer of the teacher model and the input layer of the student model respectively; Extracting abstract semantic features from the hidden layer of the teacher model and the hidden layer of the student model respectively; Decision features are extracted from the output layer of the teacher model and the output layer of the student model respectively.

5. The method according to claim 3, characterized in that A distillation loss function is constructed based on the multi-granularity features, including: Constructing an input layer network loss function according to the morphological consistency of the original signal features extracted from the teacher model and the original signal features extracted from the student model; Constructing a hidden layer network loss function according to a similarity measure between the abstract semantic features extracted from the teacher model and the abstract semantic features extracted from the student model; Constructing an output layer network loss function according to the probability classification loss and probability distribution difference of the decision features extracted from the teacher model and the decision features extracted from the student model; The distillation loss function is obtained by integrating the input layer network loss function, the hidden layer network loss function and the output layer network loss function.

6. The method according to claim 5, characterized in that An input layer network loss function is constructed according to the morphological consistency between the original signal features extracted from the teacher model and the original signal features extracted from the student model, including: The mean square error is used to measure the morphological consistency of the original signal features; The input layer network loss function is obtained according to the sum of square differences of corresponding elements of the input layer feature maps of the teacher model and the student model.

7. The method according to claim 5, characterized in that Constructing a hidden layer network loss function according to a similarity measure between the abstract semantic features extracted from the teacher model and the abstract semantic features extracted from the student model, including: Using cosine similarity to measure the similarity of the abstract semantic features; The input layer network loss function is obtained according to the cosine value of the angle between the vectors formed by the corresponding elements of the hidden layer feature maps of the teacher model and the student model.

8. The method according to claim 5, characterized in that An output layer network loss function is constructed according to the probability classification loss and probability distribution difference between the decision features extracted from the teacher model and the decision features extracted from the student model, including: The cross entropy loss of the classification result is used as the probability classification loss; The KL divergence loss of the classification result is used as the probability distribution difference; The weighted sum of the probability classification loss and the probability distribution difference is used as the output layer network loss function.

9. The method according to claim 5, characterized in that The distillation loss function is obtained by integrating the input layer network loss function, the hidden layer network loss function and the output layer network loss function, including: According to the respective weights of the input layer network loss function, the hidden layer network loss function and the output layer network loss function, the distillation loss function is obtained by a weighted sum method.

10. The method according to claim 3, characterized in that Updating the parameters of the student model during the training process according to the distillation loss function includes: Using the distillation loss function as the loss function of the student model; During the training process, the parameters of the student model are updated through the distillation loss function and the back propagation algorithm.

11. The method according to claim 1, wherein After the student model outputs a fault identification result based on the data acquired on the edge side, the method further includes: Uploading the fault identification result to the cloud; Obtaining a response instruction from the cloud based on the fault identification result; A control instruction is conveyed to the terminal based on the response instruction.

12. The method according to claim 1, characterized in that After the student model outputs a fault identification result based on the data acquired on the edge side, the method further includes: The data obtained by the edge side is cached and the cached data is regularly uploaded to the cloud as new training samples.

13. The method according to claim 12, characterized in that The method further includes: using the new training sample to train and update a neural network-based teacher model for identifying small current single-phase grounding faults to obtain an updated teacher model; Perform knowledge distillation based on the updated teacher model and transfer it to the student model to obtain an updated student model; The updated student model is deployed to the edge side of the cloud-edge system.

14. A cloud-edge collaborative distribution network small current single-phase grounding fault identification device, characterized in that: The device comprises: A model acquisition module is used to obtain a trained neural network-based teacher model for identifying small current single-phase grounding faults; A knowledge distillation module is used to distill the knowledge of the teacher model and transfer it to the student model using a multi-granularity feature distillation algorithm; and A model deployment module is used to deploy the student model to the edge side of the cloud-edge system, and the student model outputs a fault identification result based on the data obtained from the edge side.

15. An electronic device, characterized in that: include: at least one processor; a memory connected to the at least one processor; In which, the memory stores instructions that can be executed by the at least one processor, and the at least one processor implements the cloud-edge collaborative distribution network small current single-phase grounding fault identification method as described in any one of claims 1 to 13 by executing the instructions stored in the memory.

16. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by a processor, it implements the cloud-edge collaborative distribution network small current single-phase grounding fault identification method described in any one of claims 1 to 13.

17. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, it implements the cloud-edge collaborative distribution network small current single-phase grounding fault identification method described in any one of claims 1 to 13.

Images