Recently updated on September 20th, 2024 at 10:16 am<\/p>","protected":false},"excerpt":{"rendered":"
TThe research focused on developing an automated application called QuEEN PyService for extracting events from the QuEEN monitoring system database and performing advanced Power Quality analysis. The application facilitated the integration of DELFI (DEep Learning for False voltage dips Identification) classifiers, enabling intensive testing on a large number of voltage sags. This testing revealed that the DELFI classifiers outperformed the currently used criteria. <\/p>\n","protected":false},"author":93,"featured_media":0,"comment_status":"open","ping_status":"closed","template":"","meta":{"_acf_changed":false,"footnotes":""},"tags":[1324,1627,1338,1407],"targets":[1317],"rapporti_tipologie":[762],"class_list":["post-192449","rapporti","type-rapporti","status-publish","hentry","tag-electricity-network-en","tag-geo-database-en","tag-power-quality-en","tag-quality-of-supply","targets-research","rapporti_tipologie-report-en"],"acf":{"projects":{"ID":191051,"post_author":"464","post_date":"2024-07-10 16:51:29","post_date_gmt":"2024-07-10 14:51:29","post_content":"","post_title":"Application of Information Technology, Internet of Things, Peer-to-peer, to the electrical system in order to improve security and resilience","post_excerpt":"The project develops methods, tools, demonstrators and experiments to evaluate the application potential of the latest information and communication technologies for the benefit of the electro-energy system.","post_status":"publish","comment_status":"open","ping_status":"closed","post_password":"","post_name":"application-of-information-technology-internet-of-things-peer-to-peer-to-the-electrical-system-in-order-to-improve-security-and-resilience","to_ping":"","pinged":"","post_modified":"2024-08-09 15:59:29","post_modified_gmt":"2024-08-09 13:59:29","post_content_filtered":"","post_parent":0,"guid":"https:\/\/www.rse-web.it\/progetti\/application-of-information-technology-internet-of-things-peer-to-peer-to-the-electrical-system-in-order-to-improve-security-and-resilience\/","menu_order":0,"post_type":"progetti","post_mime_type":"","comment_count":"0","filter":"raw"},"order_posts":"","dont_show_search":false,"related_posts":false,"dont_show_hompage":false,"show_on_slider":false,"single_post_data":{"titolo_spot":"","post_content":"
The report describes the development of the QuEEN PyService application, which integrates advanced Power Quality techniques and innovative classifiers by directly accessing data from the QuEEN monitoring system. The application, fully developed in Python, automates the following functions: (i) reading all voltage sags recorded by the QuEEN system’s measurement devices; (ii) generating databases containing waveform images and effective values associated with each event, and integrating classifiers for validating and identifying the AT\/MT origin of voltage sags based on Deep Learning techniques developed by RSE in previous research activities; (iii) providing advanced Power Quality analysis services. Thanks to this application, it was possible to analyze voltage sags recorded in 2019 by 24 measurement devices, resulting in a statistical sample of 3,355 events.<\/p>\n
Each of these events was characterized using the DELFI (DEep Learning for False voltage dips Identification) application, which analyzes images representing both waveforms and effective values. This comprehensive analysis was performed to assess the performance of classifiers developed by RSE. The evaluation focused on events below the Class 3 immunity curve and compared the performance of Deep Learning-based classifiers with the currently implemented criterion in QuEEN, using the judgment of an “expert operator.” The DELFI classifier, which reads images of effective value trends, achieved the best performance, correctly identifying 98% of events below the Class 3 immunity curve, compared to 75.5% achieved by the existing QuEEN criterion. Additionally, a classifier for origin recognition, also based on Deep Learning techniques, was validated. The performance of this classifier, compared to RSE’s previously implemented “global method” for origin attribution, showed a 73% accuracy rate, which was considered unsatisfactory. Future developments will integrate Machine Learning techniques already developed by RSE into the QuEEN PyService application to undergo a comprehensive verification process.<\/p>\n","link_estreno":false,"scarica_file":[{"download_option":"download","file_name":"Download Report","download":{"ID":166557,"id":166557,"title":"20010470","filename":"20010470.pdf","filesize":2431883,"url":"https:\/\/www.rse-web.it\/wp-content\/uploads\/2022\/05\/20010470.pdf","link":"https:\/\/www.rse-web.it\/en\/rapporti\/messa-a-punto-del-modello-di-deep-learning-per-la-caratterizzazione-dei-buchi-di-tensione-e-valutazione-delle-sue-prestazioni\/attachment\/20010470\/","alt":"","author":"93","description":"","caption":"","name":"20010470","status":"inherit","uploaded_to":166556,"date":"2022-05-12 13:01:18","modified":"2022-05-12 13:01:18","menu_order":0,"mime_type":"application\/pdf","type":"application","subtype":"pdf","icon":"https:\/\/www.rse-web.it\/wp-includes\/images\/media\/document.png"}}],"button":{"text":"","link":""},"referente_group":false,"data_emissione":"2020-12-30","autori":"M. Zanoni, R. G. Maria Chiumeo, L. Tenti, M. Volta (RSE S.PA.)","rapporto":"","rif_rse":"20010470"},"satellite_post_url":""},"yoast_head":"\n