大数据多尺度状态检测方法在磨损检测的应用  

Application of multi-scale state detection method based on big data in wear detection

作　　者：刘吉臻[1] 刘继伟[1] 曾德良[1] 柳玉[1] 

机构地区：[1]华北电力大学工业过程测控新技术与系统北京市重点实验室,北京102206

出　　处：《仪器仪表学报》2013年第1期180-186,共7页Chinese Journal of Scientific Instrument

基　　金：国家科技支撑计划(2011BAA04B03)资助项目

摘　　要：随着生产过程自动化、信息化、规模化发展,以大数据为基础的信息处理技术得到广泛应用。对大数据进行多尺度深层次的挖掘,可以为状态检测提供技术依据,有效提高设备的可靠性。针对数据量巨大、干扰源众多、信息密度低、复杂度高的对象提出一种基于大数据的状态检测方法,利用数据融合、信息粒化以及多尺度分析,提取对象状态参数。通过对某机组磨煤机磨辊磨损的状态检测的实例分析验证了算法的有效性。As the development of production process automation, informatization and scalization, the information pro- cessing technology based on big data has been widely used. Multi-scale deep-level mining on big data can provide technical basis for state detection and effectively improve equipment reliability. Aiming at the objects with huge amount of data, many interference sources, low information density and high complexity, a state detection method based on big data is proposed. Data fusion, information granularity and multi-scale analysis are used to extract the ob- ject state parameters. Example analysis of state detection of roller wear characteristics for a certain coal mill verifies the effectiveness of the proposed method.

关 键 词：大数据 信息融合 状态检测 多尺度分析 

分 类 号：TP274.5[自动化与计算机技术—检测技术与自动化装置] TH87[自动化与计算机技术—控制科学与工程]