摘要
概述了时序模式发现的基本理论知识、目前的研究现状。同时对时序数据序列中发现模式问题进行描述,提出利用线段的斜率反正切值作为模式识别的样本,从而在分类时忽略模式的畸变,并成功建立故障模式,以充分发挥存储在汽轮机历史数据库中数据对电力生产的指导作用。
The basic theory knowledge and research status of time-series pattern discovery were summarized and the problems in time-series data sequence were described, The arctg slope of line was used as the sample of pattern recognition to ignore the aberrance of pattern in the classification. Malfunction patterns were successfully set up to direct electric power production.
出处
《湖南电力》
2007年第2期20-23,共4页
Hunan Electric Power
关键词
时序
模式发现
汽轮机
监控
time-series
discovery patterns
turbine
monitoring
