基于特征事件序列的船用核动力系统故障诊断方法研究

Research on Fault Diagnosis Method for Marine Nuclear Power Plant Based on Characteristic Event Sequence

  • 摘要: 针对船用核动力系统工况多变、故障概率高、操纵员支持手段匮乏等问题,提出一种基于特征事件序列的故障诊断方法。在分析大量运行数据的基础上,通过定义特征事件序列来提取不同类型故障特征,并构建各种典型事故的标准特征事件序列谱。当系统运行发生故障时,按特定算法实时提取系统当前事件序列特征,将其与标准特征事件序列谱比对,通过计算相似度,辨识引起系统异常的初因事件。经试验验证,该方法可辨识初因事件的程度,并定位其相对位置,与传统数据驱动的方法相比,具有易追溯、可解释等优点,因而更具研究和推广价值。

     

    Abstract: Aiming at the problem of frequently changing running state, high fault probability and lack of operator support means for marine nuclear power plant (NPP), a fault diagnosis method based on characteristic event sequence was proposed. In the arithmetic, the characteristic event sequence was defined to identify fault characters, and a standard characteristic event sequence chart was built according to NPP record running data. When the system ran out of the way, a characteristic event sequence was constructed from current running state data, and the accident grades and rough positions were diagnosed by comparing current event sequence with standard characteristic event sequence. Validated by experiments, the arithmetic can fix accident grade and rough position, and it has lots of excellences comparing to the traditional data driving arithmetic, such as easily casting back, explicable, and therefore it is practically useful.

     

/

返回文章
返回