基于分形理论的刀具磨损状态识别研究

《基于分形理论的刀具磨损状态识别研究》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、西南交通大学硕士学位论文基于分形理论的刀具磨损状态识别研究姓名：朱丽琴申请学位级别：硕士专业：测试计量技术及仪器指导教师：傅攀20090101西南交通大学硕士研究生学位论文第1I页AbstractInmUlmgprocess,surfacequalityanddimensionalaccuracyoftheworkpieceisaffectedbycutting-toolweal"condition．Tosurveyanddefinecutting-toolwearexactly,itisallimportantpr

2、obleminautomationproductiontobeurgemlysolvedatpresent．Cutting-toolwe&risaverycomplexprocess．VariousfactorshaveadirectorIndirecteffect011cutting-toolweal'，resultinginuncertainty．Soit．isdifficultforexperimentaldataandresulttohavegoodstability．Furtherstudieshavesh

3、ownthatcutting-toolwearisnon-linear,randomanddissipated．TherearealargenumberOfnon-stationarysignalsinthefieldsofconditionmonitoringandfaultdiagnosisforcutting-t001．Researchinganddevelopingeffectiveengineeringmethodsforprocessingnon-stationarysignalsarenecessary

4、forpromotingsustaineddevelopmentoffaultdiagnosistechnology．Vibrationanalysisisaveryimportantmeansforconditionmonitoringandfaultdiagnosis．Thispaperaimsattheresearchonthemethodsofsignalprocessingandpatternrecognition．Therefore，firstly,anexperimentalplatformwasset

5、upforthefailurediagnosisofcutting-tool，onwhichwehavedonealotofexperiments．Thenthevibrationsignalsofcutting-toolindifferentwearconditionswerecollected．Waveletnoisereductionisusedtopretreatthevibrationsignals．Themethodoffractalrecognitioncarldiagnosticatetheweard

6、egreequantificationally．Accordingtothenon-stationarityofvibrationsignals，weintroducethede-noisingmethodbasedonwaveletdecompositionandreconfiguration．Theinfluenceondenoisingeffectiscomparedbyusingdifferentmotherwaveletandwaveletthresholdrules．Forthetestdataofthi

7、sthesis，thedenoisingeffectadoptingthemotherwavelet‘db4’andthedefaultthresholdruleisthebest．Thispaperintroducesthealgorithmofboxdimensionandinformationdimensionfromthepointofviewofengineeringapplication．Theinfluenceofnoiseonthecomputationalprecisionofboxdimensit

8、ionisdiscussed．Thealgorithmofcorrelationdimension(G-Palgorithm)isinterpretedandthemethodtochooseparametersisgiven．Time-delayandMininlunlEmbeddingDimensionhaveobtainedbyusing