基于插值算法和时间序列的容积表拟合方法  

Volumetric Table Fitting Method Based on Interpolation Algorithm and Time Series

作　　者：马鑫钰[1,2] 王金涛 孙斌[1] 刘翔[2] 郑鹤[2] 钟家栋 MA Xin-yu;WANG Jin-tao;SUN Bin;LIU Xiang;ZHENG He;ZHONG Jia-dong(China Jiliang University,Hangzhou,Zhejiang 310018,China;National Institute of Metrology,Beijing 100029,China)

机构地区：[1]中国计量大学,浙江杭州310018 [2]中国计量科学研究院,北京100029

出　　处：《计量学报》2019年第4期642-646,共5页Acta Metrologica Sinica

摘　　要：在实际测量罐容积表数值的基础上,使用三次样条曲线插值和指数时间序列分析的方法,样条插值生成的数值用以补全在容积表实测采样点区间内的数值,时间序列分析方法用以推算出在罐底和罐顶未测得区间的数值。在实际计量测试中,三次均匀B样条曲线插值方法模拟与实际真值最大偏差小于0.3%,时间序列分析方法测的数据提供参考价值,依次两位向前推得数据,实际偏差小于1%。On the basis of the measured tank height-volume sampling points,a cubic spline curve interpolation and exponential time series analysis method are used.The values generated by the spline interpolation are used to make up the values in the sampling points of the volumetric table.The time series analysis method is used to calculate the unmeasured values at the tank bottom and the tank top.In the actual measurement test,the maximum deviation of the simulation of the cubic uniform B-spline curve interpolation from the actual value is less than 0.3%.The data measured by the time series analysis method provide the reference value,and the two data are forwardly pushed forward with the actual deviation less than 1%.

关 键 词：计量学 容积表 样条插值 时间序列 加权系数 指数平滑法 

分 类 号：TB938.3[一般工业技术—计量学]