长偏移距地震资料的优化契比雪夫动校正方法

Optimized Chebyshev method for normal moveout of long-offset seismic data

传统的双曲型动校正方法仅适用于偏移距与目标层深度比值较小的情况.对于长偏移距地震资料,须采用非双曲动校正公式才能得到较高的动校正精度.但常规的非双曲方法对于大偏移距仍然有较大的误差.本文提出契比雪夫截断加模拟退火优化的思路:通过沿用高阶项补偿系数的形式以克服对于大偏移距可能存在的不稳定现象,然后对常规的泰勒展开进行契比雪夫截断以减少截断高阶项所带来的误差,最后采用模拟退火算法对方程中的常系数进行优化.数值结果表明:即便在偏移距较大的情况下(最大偏移距与目标层深度比值约为3.5时),本文提出的优化的切比雪夫动校正方法仍能将各层位的同相轴拉平至理论上的精确位置,较以往的非双曲型动校正方法具有更高的数值精度和更广的适应范围.

The conventionalhyperbolic normal-moveout equation is reasonable only when the ratio of offset to depth is small enough. For long-offset data processing, nonhyperbolic equation is required to obtain high accuracy normal moveout. However, traditional nonhyperbolic equation is still not suitable for relatively large offset. In this paper, we developed a new strategy based on Chebyshev truncation and simulate annealing algorithm. Firstly, we applied compensation factors to guarantee the stability in the presence of large offset~ secondly, we applied Chebyshevtruncation to the traditional Taylor expansion to retain the general accuracy; finally, we optimized the constant coefficients using simulated annealing algorithm to improve the whole accuracy. Numerical results show that., our strategy can obtain a flatten trace gather even in the presence of large offset （e. g. ratio of offset to depth is about 3.5）. Compared with conventional nonhyperbolic methods, the proposed method, using Chebyshevtruncation and simulated annealing, is more accurate and feasible to much severe cases.