发表机构
Direction Scientifique Générale, ONERA; Institut Polytechnique de Paris, Palaiseau, France（法国国家航天研究中心科学方向部; 巴黎理工 institute）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对可压缩流学习求解器的问题，提出具熵稳定内部通量的学习有限体积格式。通过多种评估协议发现，仅保证机制在相等网格周期性情况最强，相等成本下学习收益不稳定，框架等成本增益不变，还对格式弱点进行修正并展示了空间门优势。
AI 中文摘要
可压缩流的学习求解器通常在相等网格分辨率而非相等计算成本下与经典方法比较，且通常无法保证其解保持物理上的可接受性。我们提出一种用于非结构化网格上二维欧拉方程的学习有限体积格式，通过构造保证其可接受性且具有熵稳定的内部通量。我们在任何计算前固定的协议下评估它：冻结阈值、证伪条款、负控制、学习组件的因子分解以及与精细化经典基线的等成本比较。分解产生了核心结果：仅保证机制，在两个学习头关闭时（未学习的框架），在每个周期性情况下在相等网格上是最强的格式。在相等挂钟时间成本下情况相反。学习仅在其从未见过的边界条件类型的壁面情况上稳健地有收益（10.8%）。其周期性收益随评估抽取而改变符号（在一个留出的情况上为 +10%，在最难的情况上为 -12%）。框架是唯一一种等成本增益从不改变符号的方法，每步有 1.74 倍的可衡量开销。有保证的变体完成了 36 次模拟中的 36 次，包括马赫数外推和未见过的壁面情况，且无负性事件。我们在推理时修复了有保证格式剩下的一个分布外弱点，即马赫数外推：通过尺度不变的网络输入、特定熵下限且无需重新训练，修正后的方案在一个马赫数情况下超过无约束方案，在另一个情况下将其差距缩小三分之一，在未见过的壁面上超过框架，并保持保证。一个空间门闭合了循环：仅在壁面附近激活头优于框架和修正后的方案，并能不变地转移到第二种壁面几何形状上。
英文摘要
A second order finite volume scheme rests on two local quantities: a gradient reconstructed in each cell, and a limiter which scales it down where the reconstruction would overshoot. Both are set by fixed formulas, and on coarse unstructured meshes a small network can supply better values. But a network is free to output anything, and the usual safeguard is a penalty in the training loss, which discourages inadmissible states without preventing them. We replace the penalty by a hard constraint. The network still sets both quantities, and every value it can produce lies inside safe bounds: its stencil weights cannot cancel a neighbour, and its limiter is capped by the local flow. The flux, the wall treatment and the time step are not learned and carry their own guarantees. Admissibility therefore holds for every value of the weights rather than as an outcome of training, and no negative density or pressure occurred in any computation reported here. Because the scheme is safe whatever the network does, we could ask what the network contributes. We test it on supersonic channel flow over an obstacle, including the forward facing step of Woodward and Colella. Learning lowers the error by 38% on an unseen geometry and 29% on an unseen obstacle topology, measured against the same scheme with the network switched off. The method aims at the accuracy of a fine mesh for the cost of a coarse one, and refining once improves the error fourfold while multiplying the run time by eight. Learning secures half of this improvement for a sixth of this time. All of this comes from one of the two quantities the network sets. The gradient reconstruction reproduces the full effect on its own, and the limiter accounts for about a tenth as much. This also explains why the gain fades beyond the Mach numbers the weights were trained on.
Comments16 pages, 1 figure, 4 tables. Replaces the previous version: the study is redone on steady supersonic channel flow, and the main result is new. Learning is shown to act through the gradient reconstruction rather than the limiter, which explains where the method helps and where it stops
