A macroscopically nominal flat surface is rough at the nanoscale level and consists of nanoasperities. Therefore, the frictional properties of the macroscale-level rough surface are determined by the mechanical behaviors of nanoasperity contact pairs under shear. In this work, we first used molecular dynamics simulations to study the non-adhesive shear between single contact pairs. Subsequently, to estimate the friction coefficient of rough surfaces, we implemented the frictional behavior of a single contact pair into a Greenwood-Williamson-type statistical model. By employing the present multiscale approach, we used the size, rate, and orientation effects, which originated from nanoscale dislocation plasticity, to determine the dependence of the macroscale friction coefficient on system parameters, such as the surface roughness, separation, loading velocity, and direction. Our model predicts an unconventional dependence of the friction coefficient on the normal contact load, which has been observed in nanoscale frictional tests. Therefore, this model represents one step toward understanding some of the relevant macroscopic phenomena of surface friction at the nanoscale level.

宏观上标称的平坦表面在纳米级水平上是粗糙的，并且由纳米微孔组成。因此，宏观水平的粗糙表面的摩擦性能是由纳米粗糙接触对在剪切作用下的力学行为所决定的。在这项工作中，我们首先使用分子动力学模拟来研究单接触对之间的非粘性剪切。随后，为了估计粗糙表面的摩擦系数，我们将单个接触对的摩擦行为实现为Greenwood-Williamson型统计模型。通过采用当前的多尺度方法，我们使用了源自纳米级位错可塑性的尺寸，速率和取向效应，来确定宏观摩擦系数对系统参数（如表面粗糙度，分离度，加载速度和方向。我们的模型预测了摩擦系数对正常接触载荷的非常规依赖性，这在纳米级摩擦试验中已经观察到。因此，该模型代表了迈向理解纳米级表面摩擦的一些相关宏观现象的第一步。