提出了一种适用于非结构网格下格子中心型时域有限体积方法的基于最小平方法的二次函数重构技术以取代以往采用的线性重构技术,包括插值模板的选择和重构函数的确定。通过对典型的电磁场本征值问题及散射问题的计算验证,表明了文中方法较以往方法大大改善了算法的耗散性,提高精度的同时可以显著地减少CPU时间和存储量。 A quadratic function reconstruction technique based on the least squares method is proposed to replace the lattice-centric finite-volume time-domain method in unstructured grids to replace the linear reconstruction techniques used in the past, including interpolation template selection and reconstruction function determine. Through the calculation and verification of typical electromagnetic field eigenvalue problems and scattering problems, it is shown that the proposed method greatly improves the dissipation of the algorithm compared with the previous methods, and can significantly reduce CPU time and memory while improving accuracy.