大电网机电暂态仿真数值收敛性的自动识别和提高

大电网暂态稳定计算在电压和频率极端变化条件下,时常出现数值仿真振荡现象,迫切需要自动识别并采取必要的措施调整以提高仿真收敛性。为此,文中提出了暂态仿真数值收敛性指标,基于交替求解中各积分步迭代中发电机及其控制系统和动态负荷的输出变量的方差与均值之比,自动识别影响仿真收敛的关键动态元件,并从迭代初值改进、积分步长自动调整、合理的迭代逼近和迭代解选取方面,给出了提高仿真收敛性的方法。采用所述方法对IEEE 14节点系统和实际系统数据进行了暂态仿真数值收敛性指标和关键动态元件确定以及积分步长自动调整,结果验证了该方法的正确性和有效性。

Automatic Identification and Improvement of Numerical Convergence of Electromechanical Transient Simulation in Large Power Grid

The phenomenon of numerical simulation oscillation often occurs in the transient stability calculation for large power grid under the condition of extreme voltage and frequency change. It is urgent for automatic identification and adjustment to improve simulation convergence. For this reason, this paper proposes several numerical convergence indices of transient simulation. Based on the ratio of variance and mean of output variables of generators, generator control systems and dynamic load during the integral iteration in the alternating solution process, the key dynamic components affecting the simulation convergence are automatically identified. From the aspects of the iterative initial value improvement, automatic integral step adjustment, rational approximation, and reasonable iterative solution selection, some methods are put forward to improve the dynamic simulation convergence. Finally, the proposed methods are applied in the transient simulation of IEEE 14-bus system and an actual system data, to determine the numerical convergence indices and key dynamic components and automatically adjust the integral step. The results show the correctness and effectiveness of the proposed methods.