Breadcrumbs Section. Click here to navigate to respective pages.
Chapter
Chapter
Integrating Neural Networks and Expert Systems for Fault Diagnosis of the MGC-20 Cyclotron
DOI link for Integrating Neural Networks and Expert Systems for Fault Diagnosis of the MGC-20 Cyclotron
Integrating Neural Networks and Expert Systems for Fault Diagnosis of the MGC-20 Cyclotron book
Integrating Neural Networks and Expert Systems for Fault Diagnosis of the MGC-20 Cyclotron
DOI link for Integrating Neural Networks and Expert Systems for Fault Diagnosis of the MGC-20 Cyclotron
Integrating Neural Networks and Expert Systems for Fault Diagnosis of the MGC-20 Cyclotron book
Click here to navigate to parent product.
ABSTRACT
A hybrid expert system has been designed and implemented for fault diagnosis of the ATOMKI MGC-20 cyclotron. Two artificial intelligence methodologies, multilayer feedforward neural network and the rule based expert system, are integrated to build the proposed hybrid expert system. The developed hybrid expert system consists of two levels. The first level is two feedforward neural networks and the second one is a rule based expert system. The two Neural networks are used for isolating the faulty parts of the cyclotron. The inputs of the networks are the indicators conditions of the cyclotron control panel, symptoms, where the outputs correspond to the status of the five main parts of the cyclotron. A rule based expert system is used for troubleshooting the faults inside the faulty part. It uses inputs and outputs of the neural networks and also use questions and answers from the user to define precisely the faults in the faulty part. The Performance evaluation of the developed hybrid expert system indicated that it has a high level of diagnostic performance compared with the diagnosis of a human professional expert.
