基于小波变换与D-S证据理论的热工目标模式识别  

Thermal Target Mode Recognition Based on Wavelet Transform and D-S Evidence Theory

作　　者：洪雨楠 刘鑫屏[1] 

机构地区：[1]华北电力大学控制与计算机工程学院,河北保定071003

出　　处：《广东电力》2017年第11期1-5,共5页Guangdong Electric Power

基　　金：国家重点基础研究发展计划(973计划)项目(2012CB215203)

摘　　要：受对象动态特性不同的影响,火电机组各参数在受到同一干扰源影响时出现变化的时间尺度存在差异。针对此问题,研究了一种小波变换与D-S证据理论相结合的目标模式识别方法:基于用小波变换的多分辨率特性,提取特征参数并在时域内动态对齐,再通过典型样本构造隶属度函数,最后运用D-S证据理论得到识别结果。实例分析表明,该方法可以有效地解决不同尺度的燃烧相关信号在时域内动态对齐问题,减小因各个信号出现变化的时间点不一致而导致的目标模式识别的不确定性。Due to differentdynamiccharacteristicsofobjects, therearedifferences invaried time scalesofparametersof the thermal power generating unit caused by influence of the same interference source. Therefore , the paper target mode recognition method combining wavelet transform and D-S evidence theory which is to acteristic of wavelet transform to extract characteristic parameters and dynamically align in time-domain typical sampling to construct subordinating degree function and finally apply D-S evidence theory to obtainsult Actual example analysis indicates this method can effectively solve the problemof dynamic aligning of combustion signals of different scales in time-domain and reduce uncertainty of target mode recognition caused by incpoints of signals.

关 键 词：热工信号 锅炉燃烧 小波变换 D-S证据理论 目标模式识别 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]