Coal seam is thin compared to the wavelength of seismic wave and usually shows strong anisotropy. It may form special geological bodies such as goaf and collapse in coal mining. The existence of these phenomena may lead to instability in numerical simulation of goaf area in coal seam. And the calculation speed of simulation is always a factor that restricting the development of simulation technique. To improve the accuracy and efficiency of seismic numerical simulation of goaf area, an improved vacuum method has been imported in a rotated staggered grid scheme and the calculation is implemented by combining parallel computing and task parallelism. It makes sure that the numerical simulation method can be utilized in a geological model with large differences in elastic parameters among layers and improve the performance of a parallel application by enabling full use of processor resources to expedite calculation speed. We set up anisotropic coal seam models, and then analyze the characteristics of synthetic seismograms and snapshots of different goaf areas with or without collapse by modeling. Results show that the proposed method can accurately simulate the goaf area and the calculation method can run with high speedup ration and parallel efficiency. The research will further advance the technology of anisotropic seismic exploration in coal fields, provide data for seismic inversion and give a theoretical support for coal mine disaster prediction.

与地震波的波长相比，煤层较薄，通常表现出强烈的各向异性。它可能形成特殊的地质体，例如采空区和煤矿开采中的崩塌。这些现象的存在可能导致煤层采空区数值模拟的不稳定性。而仿真的计算速度始终是制约仿真技术发展的因素。为了提高采空区地震数值模拟的准确性和效率，在旋转交错网格方案中引入了一种改进的真空方法，并结合了并行计算和任务并行性进行了计算。它确保了数值模拟方法可用于层间弹性参数差异很大的地质模型中，并通过充分利用处理器资源来加快计算速度来提高并行应用程序的性能。我们建立了各向异性煤层模型，然后通过建模分析了不同采空区的合成地震图和快照特征。结果表明，所提方法能够准确模拟采空区，并且计算速度快，并行效率高。该研究将进一步促进煤田各向异性地震勘探技术的发展，为地震反演提供数据，为煤矿灾害的预测提供理论依据。