Font Size: a A A

Research On Alarm Triggered Intelligent Fault Diagnosis Expert System Of Nuclear Power Plant

Posted on:2016-04-16Degree:MasterType:Thesis
Country:ChinaCandidate:J B LuoFull Text:PDF
GTID:2272330479995341Subject:Motor and electrical appliances
Abstract/Summary:PDF Full Text Request
The security of nuclear power plants is always the focus. In order to minimize the possibility of a serious accident or a potential accident that is a threat for the safe operation of nuclear power plant, research and develop a fault diagnosis system of nuclear power plant is beneficial to improve the reliability, and has an important role and significance of security and economic operation of nuclear power plants.The paper proposes an intelligent diagnosis expert system of nuclear power plant based on alarm information. Firstly, depending on the operating mechanism analysis, nuclear power simulator and experience of experts to obtain the symptom of faults, while according to the characteristics of nuclear power systems, taking into account the diversity principle. And then build a belief rule base with an alarming unit, which is according to the representation in imprecise knowledge. Secondly, the multiple fault diagnosis uses the information fusion technology to improve the accuracy of diagnosis. Finally, we use the improved analytic hierarchy process to establish a reasonable inference control strategies to improve the rapidity of diagnosis.Secondly, the self-learning improved method of intelligent diagnostic expert system was first proposed. In the process of evidence obtained, we use BP neural network to establish mathematical model between condition or fault degree with the threshold and limited value of confidence function, which makes the system can carry out fault diagnosis in full condition, and improves the self-adaptability of the system.Finally, we use the network technology to obtain real-time data, and the database technology to achieve the connection and storage. Then according to modular design and VC++ and human factors engineering principles to exploit, we obtain diagnostics screen display in NetControl industrial configuration software, and provide operational guidance for the operator.The simulation results show that the alarm triggered fault diagnosis method is feasible, and it can fault diagnosis quickly and accurately, thereby improving the reliability of nuclear power plant operation.
Keywords/Search Tags:Fault diagnosis, alarm triggering, belief rule base, self-learning, human factors engineering
PDF Full Text Request
Related items