柴油机缸套活塞组磨损状态的识别研究

《柴油机缸套活塞组磨损状态的识别研究》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、中南大学硕士学位论文摘要摘要缸套一活塞组是柴油机的动力源，尽管该组件在柴油机中故障率并不是最高的，然而一旦它出现异常磨损，轻则造成整机性能变差，重则整机报废。因此，对柴油机缸套一活塞组进行状态监测，对于保持柴油机良好的工作状态具有重要意义。基于以上考虑，本文进行了柴油机缸套一活塞组工作状态识别的研究。首先，本文对机身振动信号和活塞与缸套之间撞击情形进行了分析，建立了二者之间的映射关系，并明确了柴油机机型、运行工况等因素对映射关系的影响；其次，考虑到柴油机机身振动信号是非平稳随机过程，本文主要利用了小波包时频分析方法对机身振动信号进行处理，绘制了小波包三维时频能量图，提出主撞击能

2、量特征参数作为监测缸套一活塞组工作状态的重要依据；再次，本文将遗传算法与神经网络结合，改善了网络性能。在遗传算法初步优化权值的基础上，再进行BP网络的训练和识别，充分利用神经网络的非线性映射能力，实现了柴油机缸套一活塞组工作状态的识别；最后，对不同机型的缸内部件不同的磨损失效形式进行了实验，验证了本文提取特征参数的有效性。结果表明，振动分析法能在不拆机的前提下准确的识别出柴油机缸套一活塞组的工作状态。该方法简单可靠，具有良好的工程应用—1L目日lJ景。关键词缸套一活塞组，振动信号，特征参数提取，状态识别ABSTRACTCylinder-pistonassemblyisthepo

3、wersourceofadieselengine．Thoughfailurerateoftheassemblyisn’tthehighest，onceitappearsabnormalwear,theoperationqualityofthedieselengineisdecreasedthanbefore，oreventhedieselengineiSdestroyed．Thus．monitoringtheworkingstateofcylinder-pistonassemblyisveryimportantforkeepingdieselengineinagoodopera

4、tioncondition．Basedonthereasonsabove，aresearchontheidentificationofworkingstateofthecylinder-pistonassemblyiscarriedoutinthisthesis．Firstly,thethesisanalyzesthevibrationsignalofdieselenginesurfaceandthecrashingprocessbetweenpistonandcylinder,establishesthemappingrelationshipbetweenthem，andma

5、sterstheinfluencefactorsthatisthedieselenginemodels，operatingconditionsandSOon．Secondly,consideringthatthevibrationsignalisnon-stationarystochasticprocess，thethesismainlyuseswaveletpacketstodealwiththevibrationsignal．plotsthethree—dimensionaltime—frequencyenergychartofthewaveletpackets，andpr

6、oposesusingcharacteristicparameterofthemainimpactenergyasanimportantbasistomonitorcylinder-pistonassemblyworkingstate．Thirdly,thethesiscombinesgeneticalgorithmandneuralnetwork，whichwillimprovethenetworkperformance．Onthebasisofprimarilyoptimizingtheweightdistributionwithgeneticalgorithm，theth

7、esistrainsandrecognizestheBPnetwork,realizestheidentificationofthecylinder-pistonworkingstateassemblybymeansofthenonlinearmappingabilityofneuralnetwork．Finally,thethesisvalidatestheeffectivenessofextractingcharacteristicparametersbyexperimentstaken