欢迎来到天天文库
浏览记录
ID:36790484
大小:1.17 MB
页数:75页
时间:2019-05-15
《基于UKF的天然气管道的泄漏检测与定位》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、摘要基于UKF的天然气管道的泄漏检测与定位摘要工业流体管道泄漏事故导致巨大的经济损失、资源浪费和环境污染,有时甚至危及生命安全。针对缓慢发生的小泄漏的检测和定位方法是基于扩展卡尔曼滤波器-EKF(ExtendedKalmanFilter)的实时模型法。这种方法需要对管道模型线性化,但是,天然气管道模型的非线性程度严重,线性化会使系统产生较大的误差,影响检测精度;每一次迭代中需要计算Jacob矩阵,影响其检测速度。本文提出了一种基于无偏卡尔曼滤波器-UKF(UnscentedKalmanFilter)的泄漏检测与定位方法,这种方
2、法在对泄漏量进行状态估计时不需要对管道模型进行线性化,避免了线性化带来的不精确问题,而且在预测的阶段不需要计算Jacob矩阵,只包含一些线性运算,提高了检测速度。在此基础上,为了进一步提高检测精度,本文将迭代卡尔曼滤波器-IKF(IteratedKalmanFilter)的思想引入到UKF中,对UKF进行了改进,得到了迭代无偏卡尔曼滤波器-IUKF(IteratedUnscentedKalmanFilter)。基于IUKF对管道泄漏进行检测,可以保证在不增加系统维数的前提下,在状态估计的更新过程中获得更逼近真实值的系统的概率密
3、度特征,从而提高了管道检测精度。本文对EKF、UKF和IUKF的检测效果都进行了仿真对比,仿真结果表明,UKF在检测精度和速度上对EKF都有I摘要了很大程度的改进,而IUKF又在UKF的基础上对检测精度有了一定的提高,这对于长距离天然气管道的泄漏检测与定位有重要的实际意义。关键词:天然气管道模型,泄漏检测与定位,扩展卡尔曼滤波器,无偏卡尔曼滤波器,迭代无偏卡尔曼滤波器IIABSTRACTLEAKDETECTIONANDLOCATIONOFGASPIPELINESBASEDONUNSCENTEDKALMANFILTERABSTR
4、ACTPipelineaccidentsandsubsequentleakscausehugeeconomicloss,andenvironmentalpollutionsthatcanevenendangerlifesafety.Economics,safety,andenvironmentalissuesassociatedwithpipelineleaksareagrowingconcernaroundtheworld.TherealtimemodelmethodbasedonEKFiswidelyusedtodetec
5、tandlocatesmallleaks.TheextendedKalmanfiltermustlinearizethenonlinearmodelbeforeestimatingthesatesofthismodel.Andthenonlineardegreeofgaspipelinemodelisserious,whichgeneratesgreaterrorduringlinearization;ineachpredictionstepoftheestimation,itneedstocomputetheJacobmat
6、rix,whichslowstheconvergence.ThemethodpresentedinthispaperisbasedonunscentedKalmanfilter(UKF),whichdoesnotneedtolinearizenonlineargaspipelinemodel.Soitcanimprovetheprecision.Anditonlyincludessomesimplelinearcomputationintheestimation,sothedetectionspeedisgreatlyenha
7、nced.InordertofurtherimprovethedetectionprecisionbasedonUKF,thispapermodifiestheUKFwithintroducingIKFintoit-calledIUKF.IUKFcanobtainmoreaccurateprobabilitycharacteristicsofstatesintheprocessofIIIABSTRACTupdating,whichimprovestheestimationprecisionwithoutincreasingth
8、edimensions.AndthesimulationresultsprovethattheIUKFmakesabetterperformancecomparingwithEKFandUKF,whichmakesgreatsenseinleakdetectionandloc
此文档下载收益归作者所有