摘要
针对一类时间常数相差较大的多阶指数函数,提出一种逐级递推式拟合算法.根据不同时间常数的指数函数具有不同平衡时间的特点,首先选取最后一段,采用一阶指数函数的拟合算法进行拟合,从而确定最后一项指数函数的幅值和时间常数;然后选取倒数第二段,通过变换同样采用一阶指数函数的拟合算法进行拟合,逐级递推,分别得到多阶指数函数不同项的时间常数和相应的幅值.此方法简单实用,具有较高的拟合精度,适合于具有大进给力的纳米级热驱动部件的精确建模.
A step by step fitting algorithm was presented for multi-order exponential function with larger ime constant difference.For the exponential function with different time constant has different balance ime.The last part is firstly fitted by the algorithm for single-order exponential function,then confirm the mplitude and time constant of the last item of the multi-order exponential function,then choose the last econd section,adopt the same fitting algorithm for single-order exponential function through translating,nd get different time constants and corresponding amplitude separately step by step.The algorithm is imple and practical,with high precision of fitting.It was applied to establish the mathematical model of he nanometer drive parts with large feeding force.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2010年第12期2365-2369,2375,共6页
Journal of Zhejiang University：Engineering Science
基金
浙江省重大科技专项基金资助项目(2008C01026-1)
国家科技重大专项基金资助项目(2009ZX04014-026-04)
博士学科点专项科研基金资助项目(200803350028)
关键词
逐级递推
拟合算法
多阶指数函数
step-by-step recurrence method
fitting algorithm
multi-order exponential function
