中文摘要：

针对高速数字系统的混合参数宏建模，采用模态矢量拟合算法具有模型拟合精度高，易于电路仿真集成的优点，但是初始生成的宏模型可能会在局部频带内不满足系统无源性。该文提出一种基于特征值模态的混合参数宏模型无源性补偿方法。首先构建了包含宏模型特征对的无源性目标函数，利用特征对加权的形式来控制参数扰动带来的模型精度变化，然后结合相应的无源性约束条件，将无源性补偿过程等效为可以解析求解的最小二乘优化问题。该方法在补偿混合参数宏模型无源性的同时，兼顾模型在任意端接激励源条件下的应用要求。结合实例并与留数扰动、模态扰动等无源性补偿方法进行比较，结果表明该方法在精度或效率方面均具有一定的优势。

英文摘要：

The hybrid macromodels of high-speed digital systems fitted by the Modal Vector Fitting （MVF） approach have the advantages of accurate fitness and easy to be simulated with equivalent circuits, but the initial-built models may violate the system passivity in partial-band. To solve this problem, a method of passivity enforcement of hybrid macromodels based on modal components is presented. Firstly, it proposes a new formulation for the objective function, wherein the eigenpairs of the macromodel are included. The relative accuracy of the model is controlled with weighting by the eigenpairs in the process of the parameter perturbation. Then the relevant passivity constrain function is also discussed and united with the above objective function. According to the analysis of the equations set, it is essentially that the passivity compensation is equivalent to the solution of a least-square-based optimization problem, which is able to be solved using the analysis method. This proposed method can effectively enforce the passivity of hybrid macromodel and be satisfactory with arbitrary terminal conditions. Compared with the reported methods such as Residue Perturbation （RP） and Modal Perbation （MP）, experimental results demonstrate that the presented method has a certain advantage on the model accuracy or computational efficiency.