【摘要】 结构原型是通过任一定义的具有代表性的条件筛选出来的具有相同条件的结构类型。结构推演是一种理论上预测新化合物的方法。首先用特定元素替换大量的结构原型中的元素,得到很多化合物结构;然后,利用第一性原理对这些结构进行结构优化和能量计算,得到每个结构的能量值、形成焓等物理量;最终以形成焓为依据预测这些结构是否有可能形成稳定化合物。这一系列的处理和计算就是一次结构推演。结构推演是一种典型的高通量筛选,涉及到多通道,多任务,高并发的计算。MatCloud材料计算平台是一个国内自主研发的高通量材料计算平台,其中一个主要功能就是实现了这些多通道,多任务,高并发作业的自动计算流程。为了更为有效及高效地开展结构推演,本论文的工作是在MatCloud材料计算平台已有的工作基础上,开发出一个基于Web的结构推演自动流程计算模块。更多还原

【Abstract】 Structure prototype(1) is one arbitrarily chosen ‘representative’ entry of all entries belonging to the same structure type". Structure derivation is a method of predicting crystal structures theoretically. First, compound structures can be obtained by replacing the elements of the prototypes with certain elements. Then, ab initio methods can be used to calculate energy, formation enthalpy and other physical quantities of these structures. Finally, whether these compound structures are compound-forming systems can be predicted according to formation enthalpy. This series of operations and calculations is called structure derivation. Structure derivation is a typical high throughput screening operation, which involves calculations of multi-task, multi-channel, and high-concurrency. MatC loud is a high throughput computational material platform, which can facilitate the automation of these multi-task, multi-channel, and high-concurrency simulation tasks. Based on MatC loud, the paper discusses the development of this Web-based structure derivation tool to help with new material discovery.更多还原