幂法迭代求前几阶固有频率和振型是振动算法之一。但当相邻两阶频率十分接近时,收敛速度大为降低。本文中探索了提高收敛速度的方法:类似于松弛法,在迭代中引进了收敛因子以减少计算次数。但本法中不采取将元素逐个松弛的做法,以免造成收敛因子值难以选取的困难。文中给出了选择收敛因子值的实用方法和收敛条件。最后在计算实例中对幂法和收敛因子法作了比较。 The power law iteratively seeks the first few natural frequencies and mode shapes is one of the vibration algorithms. However, when the frequencies of the adjacent two orders are very close, the convergence speed is greatly reduced. This article explores ways to increase the speed of convergence: similar to the relaxation method, the convergence factor is introduced in the iteration to reduce the number of calculations. However, this law does not adopt the practice of relaxing elements one by one, so as to avoid the difficulty of selecting convergence factor values. The paper presents a practical method and convergence conditions for selecting the convergence factor value. Finally, the power law and the convergence factor method are compared in the calculation examples.