研究人员开发了一种使用增量奇异值分解（iSVD）的新型自适应降阶模型（ROM）框架。该方法通过允许ROM适应新数据来提高高维动力学模拟的准确性和效率，克服了当动力学超出初始训练范围时存在的局限性。iSVD方法具有历史感知能力，保留了过去动力学的信息，以提高长期预测能力。在Sod激波管和旋转爆震发动机等复杂的非线性问题上进行了测试，与现有基线相比，iSVD自适应ROM在预测精度和计算效率方面均表现出色。 AI
影响 这项研究可能带来更高效、更准确的依赖复杂动力学系统的领域模拟。
排序理由 该集群包含一篇详细介绍新计算方法的学术论文。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2605.28684v1 Announce Type: new Abstract: Reduced-order models (ROMs) can accelerate high-dimensional dynamical simulations, but their accuracy often deteriorates when online dynamics leave the regime represented by offline training data. We develop a projection-based adapt…
Reduced-order models (ROMs) can accelerate high-dimensional dynamical simulations, but their accuracy often deteriorates when online dynamics leave the regime represented by offline training data. We develop a projection-based adaptive ROM framework based on incremental singular …
Reduced-order models (ROMs) can accelerate high-dimensional dynamical simulations, but their accuracy often deteriorates when online dynamics leave the regime represented by offline training data. We develop a projection-based adaptive ROM framework based on incremental singular …