摘要
在役潜艇的建造和修理需要检测其耐压壳体的截面圆度,因现场条件限制或多设备融合测量,测量采样点经常呈现非均匀性,导致最小二乘圆拟合的结果与实际情况不符,研究旨在减小非均匀采样带来的影响,提高拟合圆的精度。基于非均匀采样定权提出一种非均匀采样加权总体最小二乘圆拟合方法,同时考虑观测向量和系数矩阵都有扰动,用于潜艇耐压壳体测量数据的处理。通过对4种不同预设形状的采样分析数值实验表明,该方法能提高非均匀采样情况下圆拟合的精度,但当随机误差加大至1.5 mm,该方法不能完全优于最小二乘法。非均匀采样加权总体最小二乘圆拟合法计算得到的圆心偏差及半径偏差都比最小二乘法的结果更小,具有较好的工程实用价值,可用于在役潜艇耐压壳体圆度测量非均匀采样时的圆拟合计算。
The maintenance and repair of in-service submarines need to detect the roundness of the section of the pres-sure hull.Due to site conditions or multi-equipment fusion measurement,the measurement sampling points often show non-uniformity,resulting in the result of least squares circle fitting and the actual situation.The research aims to reduce the influ-ence of non-uniform sampling and improve the accuracy of fitting circles.Based on non-uniform sampling weighting,a non-uniform sampling weighted overall least squares circle fitting method is proposed,which also considers the disturbance of observation vector and coefficient matrix,which is used for the processing of submarine pressure hull measurement data.Nu-merical experiments of sampling analysis of four different preset shapes show that this method can improve the accuracy of circle fitting in the case of non-uniform sampling,but when the random error increases to 1.5 mm,this method cannot be completely better than the least squares method.The center deviation and radius deviation calculated by the non-uniform sampling weighted overall least squares circle fitting method are smaller than the results of the least squares method,which has good engineering practical value,and can be used for the measurement of the roundness of the in-service submarine pres-sure hull.Circle fit calculation when uniformly sampled.
作者
邓为耀
彭飞
王中
孟庆旭
韩玉超
王雪妮
DENG Wei-yao;PENG Fei;WANG Zhong;MENG Qing-xu;HAN Yu-chao;WANG Xue-ni(Ship and Ocean College,Naval University of Engineering,Wuhan 430033,China)
出处
《舰船科学技术》
北大核心
2023年第18期51-57,共7页
Ship Science and Technology
基金
国家自然科学基金资助项目(51609253)。
关键词
潜艇
耐压壳体
圆拟合
非均匀采样
加权总体最小二乘法
submarine
pressure hull
circle fitting
non-uniform sampling
weighted total least squares
