论文部分内容阅读
使用Zelt和Barton的方法,通过一个计算效率高的有限差分求解eikonal方程,正演计算走时和射线路径.使用最小二乘QR分解法,求解稀疏线性系统方程组.使用正则化层析反演,结合用户给定的最小的、最平坦和最平滑的扰动限制,每一个加权因子随深度变化.结合数据残差和模型粗糙度的最小化,为数据残差提供一个最平滑的近似模型.该反演方法为非线性反演,需要一个初始模型,在每一次迭代时,需要计算新的射线路径.使用二维初至走时数据,对某油田二维井间地震实际资料进行无限频率初至走时层析反演.将反演所得到的速度与井的测井速度曲线相比较,二者吻合程度较高,表明该反演方法所得速度的分辨率比较高.证实了二维无限频率初至走时层析反演可以为全波形反演提供一个分辨率较高的长波长速度模型,从而为全波形反演井间地震实际资料提供了一个比较可靠的初始速度模型.
Zelt and Barton’s method are used to solve the eikonal equation by finite difference with high computational efficiency, and the forward and backward ray paths are forward calculated. The least square QR decomposition method is used to solve the sparse linear system equations. Using regularized tomography inversion, Combined with the minimum, flattest and smoothest disturbance limits given by the user, each of the weighting factors varies with depth, providing a smoothest approximation model for data residuals in combination with the minimization of data residuals and model roughness. The inversion method is a nonlinear inversion and needs an initial model, which needs to calculate a new ray path at each iteration.Using two-dimensional first-arrival time-travel data, Time-lapse tomographic inversion. Comparing the velocity obtained from the inversion with the well log velocity curve, the agreement between the two is quite high, indicating that the resolution of the inversion method is relatively high, confirming that the two-dimensional infinite frequency initial Tomographic inversion can provide a high-resolution long-wavelength velocity model for full waveform inversion, providing a more reliable method for full waveform inversion of crosswell seismic data The initial velocity model.