A high proportion of wind power systems with correct as set weak network connections may suffer from cascading trip faults due to the lack of conventional power plants and voltage/var support, which brings about great challenges to secure wind farm operations. To accurately and efficiently evaluate the dynamic var margin for an online field application within large-scale integrated wind areas, the transient dynamic voltage processes of hundreds or thousands of wind turbines must all be considered due to the dramatic interior voltage fluctuation effect of wind units, which brings a heavy computational burden for solving the TSCOPF problem. Therefore, to address these large-scale challenges, this paper proposes an objective-oriented approximation dimensionality reduction equivalent (OOA-DRE) approach, which accurately transforms the high-dimensional differential system to a low-dimensional system with few differential equations and low-dimensional algebraic equations for the wind unit with the maximum electrical distance value. Meanwhile, this paper uses the Levenberg-Marquardt method to obtain the global pre-evaluation error, so that the evaluation error reaches the minimum value. Finally, an actual wind power system verifies that the calculation time is reduced by approximately 2/3 compared with the original model on the premise of ensuring accuracy for large-scale systems.

中文翻译：

---

高透风电力系统的快速无功裕度评估：一种面向目标的降维近似方法


大部分已正确设置弱网连接的风电系统，由于缺乏常规电厂和电压/无功支持，可能会出现级联跳闸故障，这给风电场的安全运行带来了巨大挑战。为了准确、高效地评估大规模综合风区在线现场应用的动态无功裕度，由于风电机组内部电压波动影响剧烈，必须考虑成百上千台风机的暂态动态电压过程，这给求解 TSCOPF 问题带来了沉重的计算负担。因此，为了解决这些大规模挑战，本文提出了一种面向目标的近似降维等效（OOA-DRE）方法，该方法可以准确地将高维微分系统转换为具有很少微分方程和低维数的低维系统。具有最大电气距离值的风单元的维代数方程。同时，本文采用Levenberg-Marquardt方法求得全局预评价误差，使评价误差达到最小值。最后通过实际风电系统验证，在保证大规模系统精度的前提下，计算时间比原模型减少约2/3。