腔衰荡技术应用中的快速数据拟合

Fast data fitting in the application of cavity ring-down technology

数据拟合是腔衰荡技术应用中极其关键的一个步骤,因为它一定程度上决定着腔损耗测量的精度和速度。在连续波腔衰荡技术应用中,探测器实际测得的信号并非理想的单指数衰减,而是表现为单指数衰减项和常数项之和。这种非线性模型的拟合常采用牛顿法或其改进法等非线性拟合方法进行,但这些方法计算量大、速度缓慢,而且还需要良好的初始条件。针对这种情况,在分析实际连续波腔衰荡信号的特点的基础上,本文对连续波腔衰荡数据拟合中的若干问题(模型建立、数据预处理等)进行了研究,采用加权最小二乘法来拟合线性化处理后的腔衰荡信号。实验表明,该拟合方法在减少运算量、提高拟合速度(至少6倍)的同时,能保证其拟合结果与非线性拟合结果之间误差不超过0.01ppm。

Data fitting is the most important step in the application of cavity ring-down technology, it decides the precision and speed of measurement on extent. Actual signal coming from detector dcesn＇ t ring down as ideal single exponent, it can express with single exponent adding to constant. The fitting of nonlinear signal usually uses Newton and its improved method, but these nonlinear fitting methods are so time - consuming, at the same time, they need good initialization. Therefore, basing on character analysis of cavity ring- down signal, the paper researches some questions of date fitting in cavity ring- down tech- nology, such as model establishment and data pretreatment, the data is fired with least square method after linearization. The experiments show that the method can not only reduce operation time and qnicken data fitting （ at least 6 times） but also hold outcome difference from noulinear least square method smaller than 0. 01tram.