A model for acoustic cavitation flows able to depict large geometries and time scales is proposed. It is based on the Euler–Lagrange approach incorporating a novel Helmholtz solver with a non-linear acoustic attenuation model. The method is able to depict a polydisperse bubble population, which may vary locally. The model is verified and analyzed in a setup with a large sonotrode. Influences of the initial void fraction and the population type are studied. The results show that the velocity is strongly influenced by these parameters. Furthermore, the largest bubbles determine the highest pressure amplitude reached in the domain, which corresponds to the Blake threshold of these bubbles. Additionally, a validation is performed with a small sonotrode. The model reproduces most of the experimentally observed phenomena. In the experiments, neighboring bubbles are found which move in different directions depending on their size. The numerical results show that the responsible mechanism here is the reversal of the primary Bjerknes force at a certain pressure amplitude.

提出了一种能够描绘大几何形状和时间尺度的声空化流模型。它基于欧拉-拉格朗日方法，结合了新颖的亥姆霍兹求解器和非线性声衰减模型。该方法能够描绘可能局部变化的多分散气泡群。该模型在带有大型超声波发生器的装置中进行了验证和分析。研究了初始空隙率和群体类型的影响。结果表明，速度受这些参数的影响很大。此外，最大的气泡决定了域中达到的最高压力幅度，这对应于这些气泡的布莱克阈值。此外，还使用小型超声波发生器进行验证。该模型再现了大部分实验观察到的现象。在实验中，发现相邻的气泡根据其大小向不同的方向移动。数值结果表明，这里的主要机制是在一定压力幅度下初级 Bjerknes 力的反转。