A Transfer Learning-Based High Impedance Fault Detection Method Under a Cloud-Edge Collaboration Framework

Abstract

High impedance faults (HIFs) in distribution networks are hard to describe and be detected precisely because of the complexity and randomness of their features. Therefore, traditional feature analysis methods may lack sufficient reliability and generalization, which makes data-based methods a more appropriate option. However, according to previous statistical analyses, in practical scenarios, only a small quantity of historical HIF data (less than 20%) can be recorded and utilized. In this paper, a transfer learning-based HIF detection method is proposed under a cloud-edge collaboration framework of the Internet of Things, which can solve the problem of insufficient data by integrating historical data from multiple distribution networks. Through the cloud-edge collaboration framework, all features from different distribution networks are first integrated to form a basic cloud convolutional neural network model for HIF detection. The features are extracted and updated by edge computers based on the accurate synchronous measurements provided by distribution-level phasor measurement units. To uniform the data scales of the different distribution networks, principal component analysis is adopted during feature extraction. Specific to each distribution network, the target HIF detection model is transferred from the basic cloud model by fine-tuning. Furthermore, a data augmentation method based on locality sensitive hashing is proposed to improve the performance of the transferred model. The proposed HIF detection method can be operated in both online and offline modes. The performance was verified by seven different distribution networks in numerical simulations and one practical experimental distribution network. INDEX TERMS High impedance faults, cloud-edge collaboration, distribution-level phasor measurement units, transfer learning.

A Systematic Literature Review on Transfer Learning for Predictive Maintenance in Industry 4.0

“…Homogeneous Learning [148], [112], [28], [182], [29], [30], [31], [147], [32], [83], [153], [33], [84], [34], [183], [152], [36], [37], [38], [39] , [155], [40], [35] , , [41], [25], [42], [85], [43], [44], [60], [47], [48], [49], [65], [160], [86], [87], [89], [54], [91], [56], [81], [92], [58], [93], [162], [53], [94], [163], [61], …”

“…The main reason is the availability of public open-source fault data of bearings and gearboxes gathered from different machines in different working operations, which makes cross-domain studies easier to be executed. Only a few cross-domain PdM preliminary studies cope with different kinds of components or machines, such as Tennessee-Eastman (TE) process[104], Turbofan Engine[106], Chillers[80], Induction Motor[154], 3D Printer[32], Power Transmission Line Inspection[160], Fed-batch Penicillin Fermentation[182].…”

A cloud edge collaboration framework for fire detection model training

“…The cloud edge collaboration framework can be applied not only for model inference but also for model training. Zhang Y et al [19] proposed a migration learning based high impedance fault detection method under the IoT cloud edge collaboration framework to solve the problem of insufficient data by integrating historical data from multiple distribution networks. Ding C et al [20] proposed a cloud edge collaboration framework to provide persistent, fast response and high accuracy cognitive services, including two phases, initialization and update, using CloudCNN to assist in training EdgeCNN and improve the performance of EdgeCNN.…”

High-impedance Fault Detection Method Based on Feature Extraction and Synchronous Data Divergence Discrimination in Distribution Networks

“…For example, a semi-supervised learning method [8] has been proposed to detect HIFs in high signal-to-noise ratio (SNR) environments and can achieve fault detection using a combination of labeled and unlabeled data. In [11], a transferred learning-based method is presented to detect HIFs by combining principal component analysis (PCA) with a convolution neural network under a cloudedge computer framework. However, information communication factors such as channel noise and information coding methods affect HIF detection accuracy.…”