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. Due to less deep drilling and less data, the deep-time deep conversion is mainly based on the shallow data fitting formula in the past.In the shallow fitting data segment, due to the fitting data constraint, the fitting degree of various fitting methods is very high, However, at a deeper level, because of the lack of fitting data, the curve of fitting extrapolation has uncertainties.In this paper, the actual drilling data of the deepwater area in the northern South China Sea are used to analyze the fitting method and fitting data, The change of the constant term in the co-ordination formula points out the following points to make the deep time-to-depth conversion: (1) The choice of fitting method takes the variation of the layer velocity as the standard, and the growth of the layer velocity obtained by the power function fitting decreases with the increase of the depth Small, consistent with the law of speed variation, the power function conversion depth is the most reasonable; quadratic polynomial fitting layer rate growth rate is too fast, the conversion depth is greater than the actual depth; cubic polynomial fitting layer velocity will appear in the deep The speed of inversion is lower and the conversion depth is less than the actual depth, but the fitting result before speed reversal is correct. (2) The variation of fitting data has different effects on different fitting methods. The cubic polynomial The change of time is the most sensitive, the time-depth fitting result is very unstable, and the time-speed relationship changes greatly. Compared with the cubic polynomial, the quadratic polynomial and power function fitting result are relatively stable to the fitting data. (3) Whether the term is zero has the least effect on the second-order polynomial, the second is the power function, and the third-order polynomial has the greatest influence.