Based on a recently developed computer code (PAIS---particle agglomeration and interaction simulation), numerical simulations are presented which model the process of acoustic agglomeration using two different agglomeration kernels: the classical orthokinetic and the new acoustic drag interaction kernel. The results of the simulations are contrasted with experimental data. Acoustic particle agglomeration is an effect that occurs when an aerosol is sonified with high-intensity sound (140--165 dB). Experimental investigations show that acoustic agglomeration may increase the efficiency of conventional particle filtering devices (such as cyclone filters or electrostatic precipitators) by modifying the aerosol's particle size distribution. However, existing models based on the classical orthokinetic kernel fail to predict certain important features of acoustic agglomeration. The new PAIS code overcomes this limitation by implementing acoustic drag interaction, a hydrodynamic effect that was only recently found to play an important role in the acoustic agglomeration process. As a result, new model predictions based on both kernels are compared with experimentally determined data. Important differences in the simulation with either kernel are pointed out. It is shown that the acoustic drag interaction kernel, in contrast to the classical model, accounts for the important agglomerations occurring between same and similarly sized particles.