基于改进K-means聚类定心算法的曲轴轴颈圆度误差评定  

Crankshaft Journal Roundness Error Assessment Based on Improved K-means Clustering Centering Algorithm

作　　者：邹春龙 黄配乐 王生怀 冯乾新 王宸 Zou Chunlong;Huang Peile;Wang Shenghuai;Feng Qianxin;Wang Chen(School of Mechanical Engineering,Hubei University of Automotive Technology,Shiyan,Hubei 442002,China;不详)

机构地区：[1]湖北汽车工业学院机械工程学院 [2]东风设备制造有限公司

出　　处：《工具技术》2024年第6期141-150,共10页Tool Engineering

基　　金：湖北省自然科学基金(2020CFB755);湖北省教育厅科研计划项目(T2020018,Q20191801);湖北省优秀高校中青年科技创新团队项目(T2022027);工信部“高档数控机床与基础制造装备”项目(2018ZX04027001);湖北省重点研发计划(2021BAA056)。

摘　　要：曲轴轴颈的圆度误差作为曲轴必检的核心尺寸,直接影响曲轴的寿命和性能。针对圆度误差求解数据量多和计算复杂的问题,提出一种基于改进K-means聚类定心算法的圆度误差评定方法。该算法通过对轴颈采样通道的样本点进行环形聚类获得集合UK,同时以设计的目标控制器剔除UK的噪声点,以UK的最小二乘法圆度评定误差fm来估计整个环形样本的误差。聚类值从K=5循环迭代增加,直至fm符合预设统计质量控制规划。评定结果表明,聚类定心算法的圆度误差评定方法能实现曲轴圆度误差的高效、精确评定。The roundness error of the crankshaft journal is a core dimension that must be inspected on the crankshaft,which directly affects the life and performance of the crankshaft.In order to solve the problem of large amount of roundness error data and complex calculation,a roundness error assessment method based on the improved K-means clustering centering algorithm is proposed.This algorithm obtains the set U K by performing circular clustering on the sample points of the journal sampling channel.At the same time,the designed target controller is used to eliminate the noise points of U K,and the least square roundness evaluation error f m of U K is used to estimate the error of the entire circular sample.The clustering value increases iteratively from K=5 until f m meets the preset SQC rules.The evaluation results show that the roundness error assessment method of the cluster centering algorithm can achieve efficient and accurate assessment of crankshaft roundness error.

关 键 词：曲轴 圆度评定 K-MEANS 聚类定心 

分 类 号：TG713[金属学及工艺—刀具与模具] TH162[机械工程—机械制造及自动化]