A Parallel and Hierarchical Neural Networks Model for Acoustic Signature Analysis

用于声学特征分析的并行分层神经网络模型

• 批准号: 9415004
• 负责人: shahla keyvan
• 金额: $ 8.35万
• 依托单位: Missouri University of Science and Technology
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1994
• 资助国家: 美国
• 起止时间: 1994-09-15 至 1997-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9415004&HistoricalAwards=false
• 关键词: Parallel; Hierarchical; Neural; Networks; Model

9415004 Keyvan According to the NRC Regulatory Guide 1.133, any reactor licensed since 1978 must have systems to detect loose parts and components with reactor vessels and primary coolant systems. Early detection of loose parts in light water reactors is important from both safety and economic point of view. Early detection of loose parts can prevent extended outages, reduce downtime for repair of damage caused by undetected parts, and could provide an indication of degradation of reactor internals. The result of this research will provide a methodology to better utilize acoustic signature for diagnostic purposes. In case of nuclear plants, this will have significant positive impact on safety, economic, and plant life extension for aging plants. The result of this research will have two major positive impacts: 1) Effective acoustic signature analysis for general diagnostic monitoring purpose, 2) Advancement of the loose part monitoring process, which will in turn minimize operator error. ***

根据NRC监管指南1.133，自1978年以来，任何获得许可的反应堆都必须有检测反应堆容器和主冷却剂系统松动部件和组件的系统。从安全和经济的角度来看，早期发现轻水反应堆的松动部件是非常重要的。早期发现松动的部件可以防止长时间的停机，减少因未检测到的部件造成的损坏的维修时间，并可以提供反应堆内部退化的指示。本研究的结果将提供一种方法，以更好地利用声学特征的诊断目的。就核电站而言，这将对老化核电站的安全、经济和延长寿命产生重大的积极影响。本研究的结果将有两个主要的积极影响：1)有效的声学特征分析用于一般诊断监测目的；2)推进松散部件监测过程，从而最大限度地减少操作员的错误。＊＊＊