提高机器学习潜力的可靠性,用于模拟异质液体
Kamron Fazel1, Nima Karimitari2, Tanooj Shah1
1Materials Science and Engineering, Rensselaer Polytechnic Institute, Troy, New York, USA.
Journal of computational chemistry
|April 25, 2024
概括
神经网络潜力 (NNP) 在相关数据上训练时,可以准确地预测接口上的流体行为. 这种方法增强了化学和生物过程的模拟,没有经验潜力.
科学领域:
- 计算化学是一种计算化学.
- 材料科学 是一种材料科学.
- 物理化学 物理化学
背景情况:
- 在接口上的原子尺度流体行为对于化学,电化学和生物过程至关重要.
- 经典分子动力学 (MD) 模拟提供了直接的洞察力,但受到原子间电位精度的限制.
研究的目的:
- 开发和验证神经网络潜力 (NNP) 以准确预测不均质流体反应.
- 建立一个框架,以第一原则模拟流体,绕过经验潜力.
主要方法:
- 在ab initio模拟上培训NNP,将MD模拟中的不均配置与外部潜力结合起来.
- 应用训练有素的NNP来预测液态水和化NaCl在接口上的特性.
主要成果:
- 在不均数据上训练的NNP显著改善密度响应,表面张力和空洞化自由能量的预测.
- 与经验潜力和缺乏同质训练数据的NNP相比,证明了更高的准确性.
结论:
- 在NNP培训中包括不均的配置对于准确的接口模拟至关重要.
- 这项工作提供了一个强大的,没有经验潜力的框架,用于从第一原理模拟不均的流体.
相关概念视频
Typical Model Studies
356
Fluid mechanics model studies often utilize scaled-down systems to predict fluid behavior in full-scale environments, such as river flows, dam spillways, and structures interacting with open surfaces. Maintaining Froude number similarity in river models is crucial, as it replicates surface flow features like wave patterns and velocities.
356
Modeling and Similitude
263
Scaled modeling is a fundamental technique in engineering, enabling the study of large and complex systems by creating smaller, manageable replicas that recreate critical characteristics of the original. In hydrology and civil infrastructure, for example, scaled models of dams help analyze water flow, turbulence, and pressure. This method allows for accurate predictions of real-world behavior within a controlled environment, significantly reducing the cost and time involved in full-scale...
263
Velocity Potential
366
In steady, incompressible flow through a long, straight pipe with a uniform cross-section, the flow in the central region (far from the pipe walls) is irrotational. This irrotational nature means that fluid particles do not rotate around their axes, and a scalar function called the velocity potential, represented by ϕ, can be used to describe their movement. In irrotational flows, the velocity field V is defined as the gradient of the velocity potential:
366


