Continuum thermo-mechanical surrogate modeling of frictional contact at atomic length scales

我们建议开发一种基于潜在原子接触行为的连续热-机械摩擦模型的系统确定的计算多尺度方法。这项工作的动机是对有效的连续摩擦模型的需求，这些模型计算效率高，易于扩展，但仍然直接和完全基于潜在的原子接触行为。我们特别关注的是两个固体之间的滑动接触，一个是硬的，一个是软的，其中也可以存在像水这样的界面介质。我们的方法是我们的研究小组共同开发的现有多尺度方法的扩展，该方法使用分子动力学（MD）模拟来开发界面系统的有限元接触模型所需的本构模型，包括多组分系统，如自组装单层和薄膜。这种现有的方法允许系统地、自下而上地确定连续体模型，可以准确地预测原始分子模型的正常接触行为。这里提出的工作将通过纳入热效应来扩展方法，以适当地表征摩擦滑动期间的切向接触行为。该方法包括建立和运行一系列具有MD模拟的“虚拟实验”，以及提取信息传递到连续体水平所需的适当机械和热变量。在这些实验的基础上，建立了合适的连续体模型，并在一组基准案例中针对分子行为进行了实施和验证。我们在这里将自己限制在聚合物和结晶固体之间的接触，只考虑接触行为（而不是体积行为）是耗散的。然而，方法学本身将以一般术语表述，以便容易地承认其他材料配对。

We propose to develop a computational multiscale methodology for the systematic determination of continuum thermo-mechanical friction models based on the underlying atomistic contact behavior. The work is motivated by a demand for effective continuum friction models that are computationally efficient and simple to scale up, yet still based directly and exclusively on the underlying atomistic contact behavior. Our particular focus is on sliding contact between two solid bodies, onehard and one soft, where an interfacial medium such as water can also bepresent. Our approach is an extension of an existing multiscale methodology developed jointly by our research groups, which uses molecular dynamics (MD) simulations to develop the constitutive models needed for finite-element contact models of interfacial systems, including multicomponent systems such as self-assembled monolayers and thin films. This existing methodology allows for a systematic, bottom-up determination of continuum models that can accurately predict the normal contact behavior of the original molecular model. The work proposed here will extend the methodology by incorporating thermal effects to properly characterize the tangential contact behavior during frictional sliding. The methodology involves setting up and running a series of "virtual experiments" with MD simulations as well as extracting appropriate mechanical and thermal variables needed for the information transfer to the continuum level. Based on those experiments, suitable continuum models are formulated, implemented and validated against the molecular behavior in a set of benchmark cases. We restrict ourselves here to contact between polymeric and crystalline solids, considering only the contact behavior (but not the bulk behavior) to be dissipative. The methodology itself, however, will be formulated in general terms to readily admit other material pairings.