الفهرس 
chapter 1 IntroductionOnly 14 pages are availabe for public view
 |  | from | 134 |  |  |
| | | from | 134 | | |
Abstract
Digital Twin (DT) has a growth revolution by increasing Artificial Intelligence (AI) techniques and relative technologies as Internet of Things (IoT). They may be considered as the panacea for DT technology for various applications in the real world such as manufacturing, healthcare, and smart cities. The integration of DT and AI is a new avenue for open research in the upcoming days. However, for exploring the issues of developing Digital Twins, there are interesting in identifying challenges with standardization ensures future developments in this innovative theme.
This thesis contributes for constructing new intelligent approaches for fault diagnosis and prediction in industrial digital twins systems. Also, it provides potential research directions that would attract digital twins‟ researchers/engineers in the field. In addition, it discusses the incorporation of AI and DT for developing various IoT-based applications with exploring the challenges and opportunities in this innovative arena.
The DT presently supports a significant tool that can generate a huge dataset for fault prediction and diagnosis in a real-time scenario for critical industrial applications with the support of powerful AI. The physical assets of DT can produce system performance that is close to reality that delivers remarkable opportunities for machine fault diagnosis for effective measured fault conditions. For that purpose, this thesis proposes new intelligent approaches for fault classification and prediction in industrial digital twins systems. The first proposal presents an intelligent and efficient AI-based fault diagnosis framework using a new optimization technique, Flower Pollination Algorithm (FPA), and Machine Learning (ML) models for two industrial DT systems namely the triplex pump model and transmission system. The second proposal presents a hybrid framework for efficient fault detection, classification, and diagnosis by utilizing a combination of ML models and Modified Flower Pollination Algorithm (MFPA). The third proposal presents a new fault diagnosis approach based on a combination of Ensemble Learning techniques, and optimization techniques like Genetic Algorithm (GA), and ML classifiers like K- Nearest Neighbour (KNN), Decision Tree (DT), and Random Forest (RF). Finally, a new efficient fault classification framework is developed based on a combination of ML and Deep Learning (DL) models for industrial DT system namely centrifugal pump system.