Utilizing different types of deep learning models for classification of series arc in photovoltaics systems

Omran, Alaa Hamza; Said, Dalila Mat; Hussin, Siti Maherah; Abdulhussain, Sadiq H.; Samet, Haidar

doi:10.1016/j.compeleceng.2021.107478

Cited by 7 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the need for artificial intelligence in various one-dimensional tasks has provoked the popularity of adapting classical multidimensional CNNs to onedimensional ones, or creating them ad-hoc. Such numerous tasks include detection of series arc faults [6], diagnosis of cardiological diseases [7], mosquito classification [8], transmission line fault detection [9], etc. Dealing with one-dimensional signals is what unites these dissimilar problems in contexts of convolutional neural networks.…”

Section: Introductionmentioning

confidence: 99%

“…Dealing with one-dimensional signals is what unites these dissimilar problems in contexts of convolutional neural networks. In [6], the authors use 1D voltage vectors for detection and classification of series arc faults in photovoltaic systems applying a 1D CNN. In [7], the input dataset consists of electrocardiogram (ECG) signals, which are essentially 1D timing sequences.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Convolutional neural network for subharmonics detection in Hardware-In-the-Loop

Yushkova,

Zambreno,

Sanchez

et al. 2023

Preprint

View full text Add to dashboard Cite

Hardware-In-the-Loop (HIL) has become a popular technique for testing real-time complex systems.One of its possible applications is power converter modeling in order to test digital controllers.When a switching converter model is used, the output of the controller can be read incorrectly by the model due to the lack of synchronization of their clocks.Therefore, undesirable oscillations (subharmonics) can appear in the output due to the aliasing effects, which prevent the HIL model from its normal functionality.While there are various solutions present in academic and industrial research in order to address this arising problem, there is no automatic algorithm to detect it when a reference model is not available. In this paper, we implement a one-dimensional convolutional neural network which allows detecting the aliasing distortions.The 1D network was chosen due to the kind of the signal to be classified - an inductor current of the converter model, which is in effect a 1D time sequence.The proposed CNN architecture is shallow and consists of 27 layers only.Despite its simplicity, it shows remarkable performance of more than 99% for both validation and testing datasets, and for the post-training characteristics.Moreover, it makes the architecture potentially implementable in embedded systems, like a real-time HIL system, what can be a good objective for future research.

show abstract