基于柴油機(jī)循環(huán)平穩(wěn)特性的氣體壓力重構(gòu)方法研究.pdf

1、武漢理工大學(xué)碩士學(xué)位論文基于柴油機(jī)循環(huán)平穩(wěn)特性的氣體壓力重構(gòu)方法研究姓名：陳彪申請學(xué)位級別：碩士專業(yè)：輪機(jī)工程指導(dǎo)教師：向陽20090501武漢理]：人學(xué)碩十學(xué)位論文AbstractCylinderpressureisoneofimportantindexforcharacterizingdieseloperationstatesThereforeunderthesituationofbeingunabletogaincylinderp

2、ressuredirectlyfromtheminitypehighspeeddieselengine，indirectidentificationofcylinderpressurebythecylinderheads’vibrationrespondsignalshasbecomemainresearchdirectionofdieselenginefaultdiagnosisTheauthorcarriedoutstudieson

3、theissueofidentifmgdieselenginecylinderpressurebycylinderheadsurfacevibrationsignalsStudycontentis鶴follows：(1)Thepaperhasstudiedgenerationmechanismof4135nonsuperchargedfourcylinderfourstrokedieselenginesurfacevibration，s

4、howedthemaindieselenginevibrationexcitationSOUrceandtheirtimingandexplaintheintrinsicrelationshipbetweentheexcitationSOurcesandtheenginesurfacevibrationsignals；introducedcyclostationarytheorypointedoutcyclostationaryprop

5、ertiesofcylinderheadsvibrationsignals，establishmentedastatisticalmodelofthevibrationsignals，giventhereasonwhythesamplingmothodbased011themodelmustbeequaltimeintervals(2)ThepaperhasstudiedaRBFneuralnetworkidentificationme

6、thodsbasedonnonhypotheticalAfterprocessingthevibrationsignalsandthecylinderpressuresignalsrespectivelybyasynchronousaveragemethod，treatedthemasthenetworkinputsandoutputs，trainedneuralnetworks，andestablishmentedthenonline

7、armappingrelationsbetweencylinderpressuresignalsandvibrationsignals，andusedthisrelationshiptoidentifythecylinderpressureTheresultsshowthattheidentificationcurvesofthepressureandthemeasuredpressurecurvesarealmostacoincide

8、nce，thattheneuralnetworkmethodhasnotonlyhighrecognitionaccuracybutalsotheadvantagesofsmallelTOIS，andgoodadaptablitytotheworkingconditionAtthesametime，thepaperpointedoutthatsynchronizationaverageofthesignalseffectivelyenh