【摘要】 场站级的故障等值建模与分析方法是新能源集中并网地区的保护原理设计与定值计算研究的基础。传统的场站等值建模多采用单机倍乘和分群聚类算法，选取特征参数等值最终将场站模型简化，或者基于机组控制策略采用迭代算法进行求解，存在等值参数误差和计算不收敛的问题。因此，该文首先分析逆变型电源故障后电流的控制策略；然后通过两点迭代的方法求取线路电压与机组输出电流的对应曲线，进而通过遍历末端机组电压，运用前推的方式得到场站并网点电压与电流相量的对应关系，求得逆变型新能源发电场站的等值曲线；最后通过仿真计算及实际短路试验数据验证显示，对于大规模逆变型新能源场站故障稳态的等值计算，所提方法无需复杂迭代计算，计算精度高，与试验录波数据平均等值计算误差为3.28%。更多还原

【Abstract】 Fault analysis in electrical power systems is a crucial prerequisite for designing protection principles and configuring settings. With the increasing integration of renewable energy sources into centralized grids, the distinctive fault characteristics of inverter-type new energy power conversion equipment are becoming more pronounced, directly influencing the performance of protection actions. However, the lack of effective fault equivalent analysis models tailored for new energy stations poses challenges in quantitatively analyzing protection operation boundaries and proposing improvement methods, necessitating the development of novel equivalent analysis methodologies.Currently, fault equivalent analysis for new energy stations can be categorized into single-machine multiplication and multi-machine equivalent analysis methods. The single-machine multiplication method simplifies the calculation process by omitting clustering steps, resulting in reduced computational complexity.However, it often sacrifices accuracy. Conversely, multi-machine equivalent analysis considers various operational variables of units and attempts to group them accordingly. Nevertheless, due to differences in unit output parameters and overhead line parameters among units, errors unavoidably arise in the equivalent calculations.Iterative methods, although applicable for fault calculations in networks with new energy sources, struggle due to the inherent topological characteristics of such stations. During the iterative process, it is necessary to individually correct each inverter-type unit, leading to poor convergence and extended computation times.To address the challenges of multi-machine equivalent analysis and the convergence issues encountered in iterative calculations for multi-node new energy stations, a novel methodology based on forward propagation of terminal unit voltages for deriving station equivalent curves is proposed. By iteratively establishing the mapping relationship between single units and line voltages, and subsequently propagating terminal unit voltages towards the grid connection point to calculate current and voltage phasors, the mapping curve of station grid voltage and current phasors, i.e., the station equivalent curve, is obtained. Representing an entire new energy station in the system topology with a curve enhances equivalent accuracy while simplifying the computation process, thereby reducing the likelihood of convergence issues arising from a high number of nodes during iteration.To validate the accuracy and feasibility of the proposed method, this paper analyzes the impact of errors in single-machine equivalent models on the final station equivalent curve. Assuming the maximum amplitude error occurs in a single unit, the influence of the single-unit equivalent model on the station equivalent curve error is analyzed. Combined with the basic principle of phasor addition, it is shown that under the condition where all units have the maximum error in the same direction, the final error will not exceed the error of the single-machine equivalent model. The error of the equivalent curve calculated with model errors aligns with theoretical analysis results, demonstrating that the proposed method does not amplify errors in single-machine equivalent models during the calculation process.Simulation results indicate that compared to traditional multi-machine equivalent and improved fault equivalent modeling, the average error of short-circuit currents is reduced by 1.13. Moreover, equivalent errors under various operating conditions remain within 0.9. Furthermore, utilizing actual artificial short-circuit test data from Xinjiang to fit single-machine curves, the average equivalent calculation error between curves obtained via the proposed method and experimental waveform data is 3.28.更多还原