الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 119 |  |  |
| | | from | 119 | | |
Abstract
The safety of nuclear power generation programs is affecting the availability and future of nuclear power plants. The objectives of nuclear safety consist in ensuring the siting and the plant conditions need to comply with adequate principles, such as the internationally accepted health, safety, and radioprotection principles. In particular, the plant at the chosen site shall guarantee that the health of the population and of the workers does not suffer adverse radiation consequences more severe than the established limits and that such effects be the lowest reasonably obtainable. where nuclear Engineers and developers around the world have a keen interest in using Artificial Intelligence for fault detection and diagnosis (FDD) models to improve the safety, reliability, and availability of nuclear power plants (NPP). Throughout this research work, Quantum Neural Networks (QNNs) are employed to classify faults of NPPs. During applying QNNs, the learning speed increases noticeably under fault conditions using Quantum Neural Algorithms. For a long time, fault detection and diagnosis (FDD) have been a significant feature of nuclear power control system design. A range of design techniques such as hardware excess, analytical excess, and specialist systems has been used to improve scheme performance. Artificial Neural Networks (ANN) have recently been highlighted for their potential feature (fault) recognition using QNNs. Due to their learning capabilities and their inherent parallel structures, ANNs are a promising method for fault-tolerant control system design. In this thesis, an approach based on QNNs and their models for detecting and diagnosing alarming failures in nuclear power reactors is presented. Quantum Perceptron Neural Networks and Multilayer of QNNs are used firstly for classification of plant sensors, and secondly for registering unique alarm variations from possible faults, and thirdly for creating a pattern classifier for the detection of faulty instruments. where the computational pucker of QNNs increases the safety of NPPs, which is needed to boost the speed and efficiency of the diagnosis process. QNNs are subsequently applied to develop many schemes for fault diagnosis. By adopting a supervised classification approach, the proposed algorithm provides high enough speed and efficiency to increase the safety of the reactor, reduce any adverse consequences, and prevent the occurrence of major accidents in NPPs. QNNs are used to identify faults of the Reactor Cooling Pump (RCP) of Kori-2 NPP. During this thesis, we compare the capabilities of quantum neural networks and classical neural networks in terms of their ability to classify with high accuracy and speed. Especially, Quantum Hamming Neural Networks, are able to classify and discover some problems and analyze them correctly, while classical networks are unable to implement such types of problems.