Evaluating the compaction quality of rock-filled subgrades rapidly and accurately poses a pressing challenge in highway engineering. To address this

this study establishes a discrete element-finite difference coupling model to simulate the response of rock-filled subgrades under impact loading. The primary parameters of the model are calibrated using indoor large-scale triaxial tests

and the model’s accuracy is verified through comparisons between calculated and field data. Furthermore

this study conducts an in-depth analysis of the dynamic response results of five commonly used gradations of rock-filled subgrades under varying degrees of compaction

discussing the influence of gradation fractal dimension and porosity on subgrade deformation response. The findings are as follows: (1) A good exponential relationship between subgrade porosity and resilient modulus is identified

and the concept of settlement ratio is introduced

with a linear relationship between settlement ratio and subgrade porosity being verified. It is suggested that both resilient modulus and settlement ratio should be used as control indicators when evaluating subgrade compaction quality. (2) A prediction function for subgrade resilient modulus considering fill gradation and porosity is obtained

revealing that particle gradation has a significant impact on resilient modulus. Specifically

as the gradation fractal dimension approaches 2.31

the resilient modulus increases more rapidly with decreasing porosity. (3) A settlement ratio of zero corresponds to the ideal compaction state of the subgrade. This study establishes a prediction model for the critical resilient modulus of the subgrade in its ideal state

considering fill gradation

and finds that the critical modulus first increases and then decreases with increasing fractal dimension D

reaching a maximum when D=2.34. These findings aim to provide new methods and theories for evaluating the compaction quality of rock-filled subgrades in engineering.