论文部分内容阅读
利用餐前地震数据正确地估计地下介质的背景速度对于复杂构造成像具有重要意义。这一问题的复杂性需要一种良好的优化方法来求解。本文给出了一种新的用于求解连续全局优化问题的计算方法，并将其应用于非线性速度反演计算。与随机法（如模拟退火、遗传算法、均场退火等）不同的是，新的优化方法利用一动态方程来控制寻优计算过程。该动态方程的优点是利用梯度信息使解快速逼近极小点；当解逼近当前极小点后，动态方程的第二项可以使解逃逸局部极值，使其具有全局寻优能力；梯度信息的利用，使求解效率大大提高。这种基于优化算法的速度估计方法在计算目标函数时勿需大量的反射波旅行时拾取，所需要的只是从叠加剖面中得到的零偏移距旅行时，利用相似度准则建立优化目标函数。采用上述寻优方法，自动完成背景速度估算。文中给出算例，证明了方法的计算效率较随机法有明显提高，速度估计算例也给出了正确的结果。
Using pre-meal seismic data to correctly estimate the background velocity of the subterranean media is of great importance to complex structural imaging. The complexity of this problem requires a good optimization method to solve. In this paper, a new calculation method for solving continuous global optimization problems is given and applied to nonlinear velocity inversion calculation. Unlike stochastic methods (eg, simulated annealing, genetic algorithms, homogenization annealing, etc.), the new optimization method uses a dynamic equation to control the optimization process. The advantage of this dynamic equation is that the gradient information is used to approximate the solution to the minimum point quickly. When the solution approaches the current minimum point, the second term of the dynamic equation can make the solution escape the local maximum and make it have global optimization ability. The gradient information The use of the solution to greatly improve the efficiency. The speed estimation method based on the optimization algorithm does not need a large number of reflected wave travels when calculating the objective function. All it takes is to calculate the objective function by using the similarity criterion when traveling at zero offset from the superimposed profile. Using the above optimization method, the background speed is estimated automatically. An example is given in the paper, which proves that the computational efficiency of the method is obviously higher than that of the random method, and the speed estimation example also gives the correct result.