电磁场、低速流体场和温度场计算方法相关数据

本文围绕单场计算方法的高效高精度仿真展开，集中呈现了四项关键技术研究。首先，提出了考虑磁滞与各向异性的三维定点谐波平衡有限元法，通过棱边-节点组合单元与规范变换降低自由度并保持矩阵对称，结合自适应松弛因子技术，显著提升非线性收敛速度，在保证误差小于10⁻⁵条件下，迭代次数与计算时间分别降低约31%和29%，且铁心损耗仿真与实验误差小于5%。其次，针对油浸式变压器绕组温升计算，引入多孔介质理论构建简化模型，实现了绕组热点温度与位置的准确预测，仿真与光纤测量结果误差在4℃以内，为三维流-热耦合分析提供了有效工程简化途径。再次，针对电工钢叠片铁心建模，提出基于等效复磁导率的改进均质化方法，在降低边缘网格密度的同时提升计算效率，对TEAM 21d-M模型的损耗仿真最大误差不超过6%，内存与耗时显著优于传统方法。最后，发展了时空自适应的气体动理学格式（STAMR-GKS），结合叉树网格与二阶GKS求解器，通过局部时间步长与高效并行策略，在包括激波与流体界面等复杂流动中展现出良好的鲁棒性与高计算效率。研究成果为电磁、热流体及多物理场仿真提供了兼具精度与效率的先进数值工具。

This paper focuses on efficient and high-precision simulation of single-field computational methods, and systematically presents four key technical studies. First, a 3D fixed-point harmonic balance finite element method (FP-HB-FEM) considering magnetic hysteresis and anisotropy is proposed. By adopting edge-node combined elements and gauge transformation, the number of degrees of freedom (DOFs) is reduced while the matrix symmetry is maintained. Combined with the adaptive relaxation factor technique, the nonlinear convergence rate is significantly improved. Under the condition that the error is less than 10^{-5}, the number of iterations and computational time are reduced by approximately 31% and 29% respectively, and the error between core loss simulation and experimental results is less than 5%. Second, for the temperature rise calculation of oil-immersed transformer windings, the porous media theory is introduced to construct a simplified model, which realizes accurate prediction of the hot-spot temperature and its position of the windings. The error between simulation results and fiber-optic measurement data is within 4℃, providing an effective engineering simplification approach for 3D fluid-thermal coupling analysis. Third, for the modeling of electrical steel laminated cores, an improved homogenization method based on equivalent complex permeability is proposed. It reduces the edge mesh density while improving computational efficiency. For the TEAM 21d-M model, the maximum error of loss simulation is no more than 6%, and the memory usage and computational time are significantly better than those of traditional methods. Finally, a spatio-temporal adaptive gas kinetic scheme (STAMR-GKS) is developed. It combines octree meshes and a second-order GKS solver, and adopts local time steps and an efficient parallel strategy, showing good robustness and high computational efficiency in complex flows including shock waves and fluid interfaces. The research results provide advanced numerical tools with both accuracy and efficiency for electromagnetism, thermal fluid and multiphysics field simulations.