鉴于快速存取记录器(QAR)数据是结构非常复杂和数据量大的时间序列数据,直接采用传统的符号聚合近似算法(SAX)对QAR数据进行描述、存储、检索等操作时,不能克服时间序列幅度值伸缩和时间轴漂移等方面的不足.提出了改进的符号聚合近似算法,将快速存取记录器数据划分为起飞、巡航和降落三个阶段,并利用此改进的算法对巡航阶段进行填补,对不同长度的故障模型序列进行有效的相似性搜索.通过实验以及其在飞机故障诊断项目中的应用,证明了其可行性和有效性,从而提高了飞机的排故效率.%As the Quick Access Recorder (QAR) data is a kind of time series data which is of very complex structure and large amount of data, it cannot avoid some of the shortcomings appeared in aspects such as time series amplitude flex and timeline drift, directly using the traditional Symbolic Aggregate Approximation (SAX) method to describe, store, retrieve QAR data. The paper put forward an improved SAX algorithm. As stated in the paper, the QAR data was divided into three steps as take-off, cruise and landing. With the improved algorithm, the cruise stage was filled as well, and the different lengths of fault model sequence were carried out effective similarity search. Through experiments and application in the aircraft fault diagnosis project, its feasibility and validity has been proven, the improved algorithm greatly improve the efficiency of troubleshooting.
展开▼
