中文摘要：

由于深层钻井少,资料少,当前主要利用浅层数据拟合公式进行深层时深转换.在浅层拟合数据段内,受拟合数据约束,各种拟合方法拟合度都很高,但在深层,由于缺乏拟合数据的约束,拟合外推的曲线存在不确定性.本文应用南海北部深水区实际钻井数据,通过分析与拟合过程相关的拟合方法、拟合数据以及拟合公式中常数项的变化,指出了做好深层时深转换注意事项：（1）拟合方法的选择以层速度变化为标准,幂函数拟合式得到的层速度增速随深度增加而减小,符合速度变化规律,幂函数转换的深度最为合理;二次多项式拟合式得到的层速度增速过快,其转换深度大于实际深度;三次多项式拟合式得到的层速度在深层会出现速度反转而降低,其转换深度小于实际深度,但在速度反转之前的拟合结果是正确的.（2）拟合数据的变化对不同的拟合方法影响不同,三次多项式对拟合数据的变化最为敏感,时深拟合结果极不稳定,时间-速度关系变化大.与三次多项式相比,二次多项式和幂函数拟合结果对拟合数据的变化相对稳定.（3）常数项是否为零对二次多项式影响最小,幂函数次之,对三次多项式影响最大.

英文摘要：

Due to being short of wells in deep strata,fitting formula from shallow strata is used for time-depth conversion in deep strata.There is a high goodness of fit for different fitting methods in shallow strata because of well data constraint. However,it is uncertain for fitting extrapolation in deep strata for lacking data constraint. Using real well data in deep water of northern South China sea,this paper point out three important items by analyzing fitting method,fitting data and constant term of fitting formula which are relevant to fitting process：（ 1） For choice of fitting method is based on tendency of interval velocity,the power function is the best one because the growth of interval velocity deriving from it abides the law of velocity that the growth of interval velocity decreases with depth increasing,so the time-depth conversion of the power function is more reliable.The conversion depth of 2nd-order polynomial function is deeper much more than actual depth because the growth of interval velocity increases heavily; the conversion depth of 3rd-order polynomial function is shallower than actual depth for inversion of interval velocity at deep depth,however,it is still applicable before the inversion.（ 2） The little change of fitting data can cause different fitting curves on different fitting methods. The fitting of the 3 rdorder polynomial function is most sensitive to the change of fitting data,and the curve of fitting is unstable significantly. Compare with the 3 rd-order polynomial function,the fitting curve of the 2 ndorder polynomial function and the power function are stable to the change of fitting data.（ 3） Constant term is zero or not can lead to obvious change to the fitting of the 3 nd-order polynomial function,little change to the fitting of the power function,subtle change to the fitting of the 2 nd-order polynomial function.