基于位姿在线反馈调整的大型法兰对装

《基于位姿在线反馈调整的大型法兰对装》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。

1、摘要：为解决工业结构中大型法兰对接装配存在的低精度、低效率等问题，提出一种基于姿态矢量及LD-PSD在线反馈的对装方法.首先，以关节臂测量机获取标定块测量点的实际坐标值，与理论坐标值对比形成拟合误差矩阵，以该矩阵2范数最小为优化目标，利用奇异值分解法（SVD）建立全局坐标系与测量坐标系映射关系，并获取全局坐标系下待装配件若干点空间坐标，结合最小二乘法及随机霍夫变换（RHT）建立待装配体当前位姿数学模型.设计六自由度位姿调整装置，以静法兰为目标位姿，依据关键姿态矢量信息解算动法兰位姿调整参量，实现位姿预调整.LD-PSD在线反馈系统检测并反

2、馈装配质量评价指标，规划微转动及微平动路径，对静法兰位姿多次纠偏以满足对接装配精度.试验结果表明：装配过程高效，装配精度高，能够实现法兰连接结构的精确对装.设计的LD-PSD在线反馈系统可实现位姿的调整、检测、反馈再调整的闭环装配操作.关键词：关节臂测量机；关键姿态矢量；LD-PSD；奇异值分解法；最小二乘法；随机霍夫变换中图分类号：TP23文献标志码：AAssemblageofLargeFlangeBasedonAttituedVectorandPositionOnlineFeedbackHOUZengxuan，丫ANGXiuwei，H

3、UANGLei,XUJun(SchoolofMechanicalEngineering,DalianUniversityofTechnology，Dalian116024，China)Abstract：Inordertosolvetheproblemsoflowprecisionandlowefficiencyintheassemblageoflargeflange,anewmethodwasconsideredonthebasisofthekeyattitudevectorandLD-PSDonlinefeedbackfortheass

4、emblageoflargeflange.Firstly,jointarmmeasuringmachinewasusedtomeasuretheactualcoordinatesofcalibrationblock，whichwascomparedwiththetheoreticalvaluestobuildfittingerrormatrix.Takingtheleast2-normoftheerrormatrixasoptimizationobjective，amappingrelationbetweentheglobalcoordi

5、natesystemandmeasuringcoordinatesystemwithSVDwasthenestablished.Secondly，thespacecoordinatesofseveralpointsinthepartstobeassembledunderglobalcoordinatesystem，andthemathematicalmodelofthecurrentpositionofthepartsbythecombinationoftheleastsquareandRHTthekeyattitudevectorinf

6、ormationforthepositionandattitudeadjustmentparametersoftailnozzlewerecalculated.Lastly，theLD-PSDonlinecorrectionsystemdetectedandfeedbackedthequalityevaluationindexoftheassemblytoplanthemicromotionandrotationpathandadjustnozzlepositioninordertosatisfytheassemblyaccuracy.A

7、ccordingtotheresultsofexperiment，theassembleprocessthroughtheproposedmethodisstableandefficient,theassembleprecisionishigh,andthepreciseassembleofnozzleandcombustionchamberisrealized.LD-PSDsystemcanrealizeaclosed-loopoperationoftailnozzle,suchasadjustment，detection,feedba

8、ck，andreadjustment.Keywords：jointarmmeasuringmachine；keyattitudevector；LD-PSD；singularvaluedecom