In reduced-order modeling of incompressible flows, the supremizer method ensures inf-sup stability for pressure recovery. However, when applied to snapshots from a penalized full-order model, the accuracy of the supremizer method deteriorates if the penalty term is ignored. But incorporating the penalty term with the supremizer method leads to a coupled system that significantly increases computational cost. We propose a novel decoupled supremizer strategy that maintains accuracy and stability while reducing costs. We derive a priori error estimate independent of the penalty coefficient and introduce new supremizer options that reduce offline costs. Numerical experiments demonstrate the accuracy of the proposed method, and reduce online computational cost to under 10% and offline cost by roughly 50% compared to the standard approach.

姚慧, 北京交通大学数学与统计学院讲师。研究方向为模型降阶和不可压流体计算。2022获得厦门大学博士学位, 师从许传炬教授。2023-2025年在法国波尔多综合理工学院进行博士后研究工作。2025年10月入职北京交通大学。在SINUM, CMAME等期刊发表多篇SCI论文。