Computational fluid dynamics has developed rapidly in the past 20 years. In addition to providing a solid material foundation for the development of computer hardware industry, it is mainly because both analytical methods and experimental methods have great limitations. For example, because of the complexity of the problem, it is impossible to solve it analytically or make experimental determination because of its high cost, while CFD method has the advantage of low cost and can simulate complex or ideal processes. The CFD software that has been tested can broaden the scope of experimental research and reduce the expensive experimental workload. Under the given parameters, the numerical simulation of this phenomenon by computer is equivalent to numerical experiment. In history, there have been examples in which new phenomena were first discovered by CFD numerical simulation and then confirmed by experiments. CFD software can generally derive a variety of optimized physical models [2], such as steady and unsteady flow, laminar flow, turbulent flow, incompressible and compressible flow, heat transfer, chemical reaction and so on. The flow characteristics of each physical problem have appropriate numerical solutions, and users can choose explicit or implicit difference schemes to achieve the best calculation speed, stability and accuracy. CFD software can easily exchange numerical values and adopt unified pre-and post-processing tools, which saves the repetitive and inefficient work of scientific researchers in computer methods, programming and pre-processing, and can put the main energy and wisdom into the exploration of physical problems themselves.