基于神經(jīng)網(wǎng)絡(luò)的車輛橫向半主動(dòng)懸掛控制

1、西南交通大學(xué)碩士學(xué)位論文基于神經(jīng)網(wǎng)絡(luò)的車輛橫向半主動(dòng)懸掛控制姓名：薛飛申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：檢測(cè)技術(shù)與自動(dòng)化裝置指導(dǎo)教師：金煒東20100501西南交通大學(xué)碩士研究生學(xué)位論文第1I頁(yè)AbstractInthehighspeedtrain，thetrainlateralridequalitygreatlyaffectsthesafetyofthetrainoperationandthepassengersridecomfort，andh

2、owtoimprovethetrainlateralridequalityisaseriousproblemThetrainvehiclesuspensionsystemisdividedintoactivesuspensionsystemandpassivesuspensionsystem，andSemiactivesuspensionsystemisaspecialformofactivesuspensionsystemWhenth

3、etrainisrunning，theparametersofthepassivesuspensionsystemcannotbeadjusted，thusthetrainlateralridequalityturnsworse；theparametersoftheactivesuspensionsystemcanbeadjustedtimelybutthecostoftheactivesuspensionisSOexpensiveth

4、atitcouldnotbeusedinthereality；thecostofthesemi—activesuspensionsystemischeap，SOitisaneffectivewaytoimprovethetrainlateralridequalityThispaperappliesneuralnetworksemiactivesuspensioncontrolsystemtoimprovethetrainlateralr

5、idequalityThisistheimportantcontentofthispaperResearchontheadvantagesanddisadvantagesoftheLateralsemiactivesuspensiondynamicsmodelindifferentdegreesoffreedomofthetrain，andselectthe17degreesoffreedomwhichcarlbettershowthe

6、modeoftherunningtrainasthecontrolledobjectmodelThen，usetheMATLABtobuilditsSIMULINKmodelResearchonthenumericalsimulationoftrackirregularity，andsimulatetheGermanhighspeed10winterferencetrackspectrumwithPrecisionDuringthesi

7、mulationofthesecondarylateralsuspensionsforthehighspeedpassengervehiclewithnobolsterbogie，theeffectiveusingneuralnetworkpredictivecontroltoimprovetheridecomfortarestudiedTheresultsofsimulationshowthat，comparedtothepassiv

8、esuspensionsystem，usingskyhookdamping，theRMSvaluesoflateralaccelerationCarbodycallbereducedabout40％，maximumvaluesoflateralaccelerationCanbereduced30％to40％，theridecomfortindexCanbereducedabout14％SOtheneuralnetworkpredicti

9、vecontrolusinginthesemiactivesuspensionsystemiseffectiveDuringthesimulationofthesecondarylateralsuspensionsforthehighspeedpassengervehiclewithnobolsterbogieondifferentspeed，therobustofneuralnetworkpredictivecontrolarestu

10、diedTheresultsofsimulationshowthat，withtheincreasedspeed，theridecomfortindexofbothpassivesuspensionsystemandsemi—activesuspensionsystemareincreased，buttheridecomfortindexofsemi—activesuspensionsystemstillreduced14％to15％S