摘要
提出一种基于神经网络技术的数据映射方法,训练样本通过蒙特卡罗方法获得·该方法能将已知的一般工业实验数据转换成特定工厂应用的数据,实现了不同阶次冗余系统失效数据之间的映射,弥补了共因失效数据的不足,降低了共因失效分析中的不确定性,为系统可靠性分析提供可靠的数据来源·通过实例分析与对比,验证了神经网络是一种解决非线性问题的好方法,同时也说明应用神经网络技术进行共因失效数据分析的可行性·
Because of rarity of common cause failure events on a plant-and system-specific basis, some models are difficult to be applied. To solve the difficulty which is general in the common cause failure analysis, a mapping method based on neural network is presented, which is used for common cause failures data analysis. It can translate generic experience event data for plant-specific applications and map the failure data between the systems with different sizes, so it makes up the rarity of common cause failures data and reduce the uncertainties. It can provide reliable data for reliability analysis. At last this article provides an example to illustrate the application.
出处
《东北大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2004年第12期1183-1186,共4页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金资助项目(50275025)
