資源描述:
《基于UKF的天然氣管道的泄漏檢測(cè)與定位》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。
1、摘要基于UKF的天然氣管道的泄漏檢測(cè)與定位摘要工業(yè)流體管道泄漏事故導(dǎo)致巨大的經(jīng)濟(jì)損失、資源浪費(fèi)和環(huán)境污染,有時(shí)甚至危及生命安全。針對(duì)緩慢發(fā)生的小泄漏的檢測(cè)和定位方法是基于擴(kuò)展卡爾曼濾波器-EKF(ExtendedKalmanFilter)的實(shí)時(shí)模型法。這種方法需要對(duì)管道模型線性化,但是,天然氣管道模型的非線性程度嚴(yán)重,線性化會(huì)使系統(tǒng)產(chǎn)生較大的誤差,影響檢測(cè)精度;每一次迭代中需要計(jì)算Jacob矩陣,影響其檢測(cè)速度。本文提出了一種基于無(wú)偏卡爾曼濾波器-UKF(UnscentedKalmanFilter)的泄漏檢測(cè)與定位方法,這種方
2、法在對(duì)泄漏量進(jìn)行狀態(tài)估計(jì)時(shí)不需要對(duì)管道模型進(jìn)行線性化,避免了線性化帶來(lái)的不精確問(wèn)題,而且在預(yù)測(cè)的階段不需要計(jì)算Jacob矩陣,只包含一些線性運(yùn)算,提高了檢測(cè)速度。在此基礎(chǔ)上,為了進(jìn)一步提高檢測(cè)精度,本文將迭代卡爾曼濾波器-IKF(IteratedKalmanFilter)的思想引入到UKF中,對(duì)UKF進(jìn)行了改進(jìn),得到了迭代無(wú)偏卡爾曼濾波器-IUKF(IteratedUnscentedKalmanFilter)?;贗UKF對(duì)管道泄漏進(jìn)行檢測(cè),可以保證在不增加系統(tǒng)維數(shù)的前提下,在狀態(tài)估計(jì)的更新過(guò)程中獲得更逼近真實(shí)值的系統(tǒng)的概率密
3、度特征,從而提高了管道檢測(cè)精度。本文對(duì)EKF、UKF和IUKF的檢測(cè)效果都進(jìn)行了仿真對(duì)比,仿真結(jié)果表明,UKF在檢測(cè)精度和速度上對(duì)EKF都有I摘要了很大程度的改進(jìn),而IUKF又在UKF的基礎(chǔ)上對(duì)檢測(cè)精度有了一定的提高,這對(duì)于長(zhǎng)距離天然氣管道的泄漏檢測(cè)與定位有重要的實(shí)際意義。關(guān)鍵詞:天然氣管道模型,泄漏檢測(cè)與定位,擴(kuò)展卡爾曼濾波器,無(wú)偏卡爾曼濾波器,迭代無(wú)偏卡爾曼濾波器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