Development of efficient Smoothed Particle Hydrodynamics (SPH) for contained flows in motor-sport applications

运动型汽车的设计和性能依赖于变速箱和油箱等关键部件。实现最佳功率传递和维持发动机燃料供应是一项非常具有挑战性的任务，在这种情况下，气流在快速变化的外力作用下以极其复杂的几何形状发生，并且涉及到以每分钟1000秒的转速旋转的齿轮产生热量。在某些条件下，加速度可以达到重力的五倍，超出了实验台的范围，这使得计算机模拟成为唯一的设计选择。传统的计算流体力学(CFD)不能很好地模拟发动机内的剧烈流体动力学，如齿轮箱，这是由于复杂且通常高度扭曲的表面运动造成的。然而，很少有模拟方法能够捕捉到完整的物理过程。这个PHD将使用平滑粒子流体动力学(SPH)。SPH没有计算网格，正在给工程模拟带来革命性的变化，这是潜在的剧烈自由表面流体动力学的理想选择，在这些流体动力学中，高度复杂的运动几何具有强烈的非线性。其目的是利用粒子分裂和合并的可变分辨率的新思想来开发一个高效的SPH模型。模拟将在图形处理单元(GPU)上加速，这是我们开源的DualSPH代码的一部分。

The design and performance of motorsport vehicles relies on key components including the gear box and fuel tank. Achieving optimum power transfer and maintaining engine fuel supply is a very challenging task where the flows occur in extremely complex geometries under rapidly-varying imposed forces and involve gear wheels rotating at many 1000s revolutions per minute generating heat. The accelerations during certain conditions, can be up to five times gravity, are beyond experimental rigs, making computer simulation the only design option. Conventional computational fluid dynamics (CFD) is not well suited to simulate the violent hydrodynamics occurring in the engine such as in the gear box due to the complex and generally highly distorted, surface motion. However, few simulation methods can capture the full physics. This PhD will use Smoothed Particle Hydrodynamics (SPH). SPH has no computational mesh and is revolutionising engineering simulation being ideal for potentially violent free-surface hydrodynamics where there is strong nonlinearity with highly complex moving geometries. The aim is to develop an efficient SPH model using new ideas for variable resolution with particle splitting & merging. The simulations will be accelerated on graphics processing units (GPUs) as part of our open-source DualSPHysics code.