颗粒物质在自然界中和工业过程中普遍存在，深入开展颗粒物质力学相关研究，在滑坡、碎屑流等颗粒流灾害研究中有着有着重要的理论意义和应用价值。目前颗粒流的主流理论都基于假想圆球颗粒建立，然而自然界中的颗粒物质几乎都是非球形的，颗粒形状对颗粒流的影响机理及理论研究任然存在很大空缺。本文在斜面密集颗粒流这一场景下使用离散单元法（DEM）开展了颗粒形状的影响研究，研究在Pouliquen流动法则框架下开展，对于每种颗粒，我们改变流动层厚度h和斜坡倾角θ，以提取h_stop(θ)曲线（低于该曲线流动停止）和Fr=β h/h_stop关系，其中β是表征流动性的关键拟合参数，Fr=u/√(gh)是弗劳德数，u是平均速度，g是重力加速度。本文系统进行了模型参数、微观力学参数和颗粒长细比等形状因素对斜面密集颗粒流的影响机理分析，最终建立了统一颗粒形状的新的流动法则。主要研究内容如下：（1）系统进行了基于斜面密集颗粒流模拟的离散元参数敏感性分析。（2）针对滑动摩擦系数（μ_s）这一重要微观力学参数，系统开展了对斜面密集颗粒流的宏微观影响分析。（3）系统进行了颗粒长细比（AR）对斜面密集颗粒流的影响机理分析。（4）构建了统一颗粒形状的流动法则。在开展了多种不同形状的颗粒（包括长条形、扁平形、多面体及真实砂颗粒）的数值模拟后，分析了这样多种不同形状颗粒在 Pouliquen 流动法则下的流动行为，为解决 Pouliquen 流动法则在描述多种形状颗粒的局限性，本文基于一个近期新提出的半理论半经验模型，通过模拟得到的大量数据，构建了能够统一不同颗粒形状的流动法则。

Granular materials are ubiquitous in nature and industrial processes. Conducting in-depth research on granular mechanics has significant theoretical and practical value in understanding granular flow hazards such as landslides and debris flows. Current theoretical models of granular flow are primarily based on the assumption of idealized spherical particles. However, most granular materials in nature are non-spherical, and there remains a substantial research gap in understanding the influence of particle shape on granular flow. In this study, the discrete element method (DEM) is employed to investigate the effects of particle shape on dense granular flows on an inclined plane. The study is conducted within the framework of the Pouliquen flow rule. For each particle type, we vary the flow layer thickness $h$ and slope angle θ to extract the h_stop(θ) curve (below which flow ceases) and analyze the relationship Fr = βh/h_stop, where β is a key fitting parameter characterizing flowability, and Fr = u/√(gh) is the Froude number, with $u$ representing the mean velocity and $g$ the gravitational acceleration. A systematic investigation is conducted on the effects of model parameters, microscopic mechanical parameters, and particle aspect ratio on dense granular flow, ultimately leading to the development of a unified flow rule accounting for particle shape. The main research contents and findings are as follows:(1) A comprehensive sensitivity analysis of discrete element parameters in dense granular flow simulations on inclined planes is performed. The influences of the linear contact model and Hertzian contact model on the results are examined, with discussions on how key parameters such as Young’s modulus, restitution coefficient, and friction coefficient affect the macroscopic and microscopic motion characteristics of particles. The results indicate that, under the same flow conditions, the rolling friction coefficient has the most significant impact. In dense flow states, the normal restitution coefficient has little effect on the results, whereas in loose states, particle velocity increases with the normal restitution coefficient. Additionally, as Young’s modulus increases, its influence on particle velocity diminishes. Within physically meaningful ranges, the stiffness ratio and tangential restitution coefficient have negligible effects on velocity. Based on these analyses, a set of recommended benchmark parameters is established for both the Hertzian and linear contact models. A parameter sensitivity analysis of the h_stop(θ) curve is also conducted, revealing that all microscopic parameters except the friction coefficient have minimal impact.(2) The effects of the sliding friction coefficient (μ_s)—a crucial microscopic mechanical parameter—on the macroscopic and microscopic behavior of dense granular flow are systematically examined. For both spherical and non-spherical particles, including spherical particles with rolling friction (μ_r=0.2) to mimic shape effects, elongated particles (AR=2), and sand particles, an increase in μ_s shifts the h_stop(θ) curve to the right and saturates at μ_s > 0.5. This trend is also observed in its influence on the dynamic repose angle θ_1 and the key fitting parameter β, though different particle types exhibit varying sensitivities. Furthermore, it is found that the rotational ability of spherical particles is unaffected by friction, whereas non-spherical particles experience restricted rotation due to shape constraints. However, as μ_s increases, their rotational ability is enhanced and eventually approaches that of spherical particles.(3) A systematic investigation is conducted on the influence of particle aspect ratio (AR) on dense granular flow. Within the Pouliquen flow rule framework, for each individual AR value, the relationship Fr = βh/h_stop successfully collapses data from different flow thicknesses $h$ and slope angles θ, but fails to achieve universal normalization across all AR values. By systematically varying the particle aspect ratio, a complex S-shaped influence trend is observed. When AR ≤ 1.3, flowability remains nearly unchanged due to unrestricted particle rotation. In the range 1.3 ≤ AR ≤ 2.0, a sharp transition in flowability occurs, reflecting fundamental shifts in microscopic interactions such as particle alignment and rotation mechanisms. For AR ≥ 2.0, flowability saturates as particle alignment ceases to change, indicating that multiple physical mechanisms jointly govern granular flow behavior. The study further integrates new simulation data with various experimental and numerical results from the literature and discusses them within different theoretical frameworks.(4) A unified flow rule incorporating particle shape is proposed. Through numerical simulations of various particle shapes—including elongated, flattened, polyhedral, and realistic sand particles—the flow behavior of these diverse particles under the Pouliquen flow rule is analyzed, which however does not collapse the data. To address the limitations of the Pouliquen flow rule in describing different particle shapes, a recently proposed semi-empirical model is employed. Using extensive simulation data, a generalized flow rule capable of unifying different particle shapes is developed. The results demonstrate that this new theoretical framework accurately predicts the flow behavior of all tested particle shapes while extending the applicability of the original model, thereby providing a more universal theoretical foundation for predicting flow velocity in dense granular flows on inclined planes. Additionally, a potentially generalizable finding is proposed: by substituting the repose angle μ_AOR for μ_1 in initial calculations, the labor-intensive measurement of the h_stop(θ) curve can be bypassed, paving the way for future applications in landslide and debris flow dynamics based on depth-integrated equations.