基于故障重构预测的舰艇水轮发电机故障信号挖掘方法  

Ships hydro-generator fault signal mining method Based on the fault reconstruction predict

作　　者：汤敏丽[1] 

机构地区：[1]凯里学院,凯里556011

出　　处：《科学技术与工程》2014年第16期273-276,共4页Science Technology and Engineering

摘　　要：当前舰艇水轮发动机故障检测都是根据实际采集的故障特征信号以模式识别理论为基础,完成故障信号挖掘,受到真实空间弱信号和失真信号难以表达的影响,故障挖掘效果很差。提出运用虚拟空间信号重构理论进行舰艇水轮发电机故障挖掘,构建以信号群集合为元素的虚拟空间,通过故障重构预测技术对故障幅值进行预测,以预测值和真实幅值出发进行故障重构,重构结果融入规范的置信度以及支持度使其反映真实故障值特征,实现舰艇水轮发电机故障数据的准确挖掘。实验数据说明该方法可对舰艇水轮发电机故障数据进行高效、准确的挖掘。Current ships hydro-generator fault detection is based on the actual fault characteristic signal on pattern recognition theory, complete fault signal mining. Due to space weak signal and the signal distortion, the influence of fault mining effect is poor. The paper uses the virtual space mines ships turbine generator fault signal recon- struction theory, and builds the virtual space by signal collection of elements. Through the fault reconstruction prediction technology to forecast the fault amplitude, it proceeds to reconstruct fault by real amplitude and forecast amplitude, refactoring results into the specification of confidence and support to make it reflect the characteristics of the real breakdown value, realization of naval vessels hydro-generator accurate fault data mining. The experimental 4＇ result shows that the method is accurately and efficiently mining to ships hydro-generator fault data.

关 键 词：舰艇水轮发电机 虚拟空间 故障重构 数据挖掘 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]