An Innovative Decentralized and Distributed Deep Learning Framework for Predictive Maintenance in the Industrial Internet of Things

dc.authoridHabbal, Adib/0000-0002-3939-2609
dc.authoridAlabadi, Montdher/0000-0001-7466-6575
dc.contributor.authorAlabadi, Montdher
dc.contributor.authorHabbal, Adib
dc.contributor.authorGuizani, Mohsen
dc.date.accessioned2024-09-29T16:04:29Z
dc.date.available2024-09-29T16:04:29Z
dc.date.issued2024
dc.departmentKarabük Üniversitesien_US
dc.description.abstractThe integration of predictive maintenance (PdM) with the Industrial Internet of Things (IIoT) represents a pivotal shift in equipment management, particularly with the incorporation of deep learning (DL) for processing time series data from IIoT devices. This combination offers a sophisticated approach to predictive analysis, harnessing DL's prowess in analyzing complex patterns in large data sets. However, it also presents notable challenges, including significant security risks associated with centralized organizations and the immense volume of time series data generated by IIoT. To address these issues, our study introduces an innovative decentralized framework thoughtfully segmented into device and edge levels. This framework leverages the strengths of blockchain technology and the interplanetary file system (IPFS). IPFS effectively manages the large-scale storage needs of time series data for DL applications in a decentralized manner, while blockchain provides a robust foundation for ensuring data security and maintaining consistent transactions. Furthermore, we conducted thorough performance analyses, examining aspects, such as accuracy, execution time, and computational cost, which validated the efficacy of our approach. Security considerations were also rigorously evaluated, focusing on potential attacker scenarios, the strengths of a decentralized architecture, and the immutable nature of smart contracts. The results highlight our framework's exceptional ability to ensure the highest level of security in DL, maintain data integrity, and preserve model accuracy. In conclusion, the seamless integration of DL, PdM, blockchain, and IPFS in our framework marks a significant advancement in contemporary industrial maintenance strategies. It successfully bridges the gap between advanced security needs and the handling of vast quantities of data, positioning our approach at the forefront of modern industrial maintenance solutions.en_US
dc.identifier.doi10.1109/JIOT.2024.3372375
dc.identifier.endpage20286en_US
dc.identifier.issn2327-4662
dc.identifier.issue11en_US
dc.identifier.scopus2-s2.0-85186972347en_US
dc.identifier.scopusqualityQ1en_US
dc.identifier.startpage20271en_US
dc.identifier.urihttps://doi.org/10.1109/JIOT.2024.3372375
dc.identifier.urihttps://hdl.handle.net/20.500.14619/6148
dc.identifier.volume11en_US
dc.identifier.wosWOS:001285460000042en_US
dc.identifier.wosqualityN/Aen_US
dc.indekslendigikaynakWeb of Scienceen_US
dc.indekslendigikaynakScopusen_US
dc.language.isoenen_US
dc.publisherIeee-Inst Electrical Electronics Engineers Incen_US
dc.relation.ispartofIeee Internet of Things Journalen_US
dc.relation.publicationcategoryMakale - Uluslararası Hakemli Dergi - Kurum Öğretim Elemanıen_US
dc.rightsinfo:eu-repo/semantics/closedAccessen_US
dc.subjectBlockchainsen_US
dc.subjectIndustrial Internet of Thingsen_US
dc.subjectSecurityen_US
dc.subjectInterPlanetary File Systemen_US
dc.subjectData modelsen_US
dc.subjectData privacyen_US
dc.subjectMedical servicesen_US
dc.subjectBlockchainen_US
dc.subjectdeep learning (DL)en_US
dc.subjectIndustrial Internet of Things (IIoT)en_US
dc.subjectinterplanetary file system (IPFS)en_US
dc.subjectpredictive maintenance (PdM)en_US
dc.titleAn Innovative Decentralized and Distributed Deep Learning Framework for Predictive Maintenance in the Industrial Internet of Thingsen_US
dc.typeArticleen_US

Dosyalar