一种船用燃气轮机强跟踪扩展卡尔曼滤波器健康参数估计方法（英文）-豆柴文库

您所在位置：网站首页 / 一种船用燃气轮机强跟踪扩展卡尔曼滤波器健康参数估计方法（英文）.docx / 文档详情

在线预览结束，喜欢就下载吧，查找使用更方便

5 金币

下载文档

/ 2

下载提示

如果您无法下载资料，请参考说明：

1、部分资料下载需要金币，请确保您的账户上有足够的金币

2、已购买过的文档，再次下载不重复扣费

3、资料包下载后请先用软件解压，在使用对应软件打开

文本预览

一种船用燃气轮机强跟踪扩展卡尔曼滤波器健康参数估计方法（英文）
Abstract:
Gasturbinesplayanimportantroleintheoperationofships.Itiscrucialtomonitorthehealthstatusofgasturbinesinreal-timetoensuretheirsafeandefficientoperation.Inthispaper,weproposeastrongtrackingextendedKalmanfilter(STEKF)methodforhealthparameterestimationofgasturbines.
Introduction:
Gasturbinesarewidelyusedinthemarineindustryduetotheirhighpower-to-weightratio,lowemissions,andrapidstart-upcapabilities.Theperformanceandlongevityofgasturbinesareaffectedbyvariousfactorssuchastemperature,pressure,andcorrosion.Therefore,itisnecessarytomonitorthehealthstatusofgasturbinesinreal-time.
Inrecentyears,data-drivenapproacheshavebeenproposedforhealthparameterestimationofgasturbines.Thesemethodsusemachinelearningalgorithmstolearntherelationshipbetweendifferentvariablesandestimatethehealthparameters.However,thesemethodsrequirealargeamountoftrainingdataandarecomputationallyexpensive.
Inthispaper,weproposeastrongtrackingextendedKalmanfilter(STEKF)methodforhealthparameterestimationofgasturbines.Theproposedmethodusesthereal-timesensordatafromgasturbinesandestimatesthehealthparametersbycombiningthesensordatawithamathematicalmodelofthegasturbine.
Methodology:
TheproposedSTEKFmethodconsistsofthreemainsteps:1)modeldevelopment,2)filterinitialization,and3)filterupdate.
ModelDevelopment:
Wedevelopamathematicalmodelofthegasturbinethatdescribestherelationshipbetweentheinputvariables(e.g.,fuelflowrate,airflowrate,andtemperature)andtheoutputvariables(e.g.,poweroutput,exhausttemperature,andefficiency).Themodelisbasedonthefirstprincipleofthermodynamicsandiscalibratedusingdatafromthegasturbine.
FilterInitialization:
Thefilterisinitializedusingthemodeloutputsandsensormeasurementsattimet=0.Thefilterestimatesthestateofthegasturbine(e.g.,temperature,pressure,andefficiency)anditsparameters(e.g.,corrosionrate).
FilterUpdate:
Thefilterupdatesthestateandparameterestimatesinreal-timeusingthesensormeasurementsandthemodelpredictions.ThefilterusestheSTEKFalgorithm,whichisavarian