高比例新能源电力系统静态电压稳定裕度在线概率评估

新能源的随机性、波动性及弱调节特性给电力系统静态电压的安全及稳定性带来了挑战。针对此问题，提出一种考虑源荷双侧不确定性的高比例新能源电力系统静态电压稳定裕度在线概率评估方法。首先，基于新能源无功调节特性与传统机组的差异，分析了大量新能源替代传统机组对稳定裕度的影响。然后，分析了新能源出力不确定性对稳定裕度分布范围的影响，并建立源荷不确定性模型以生成典型场景。最后，为了应对新能源快速波动性给稳定裕度带来的影响，提出基于优化ELM-KDE的稳定裕度在线概率评估方法。利用优化极限学习机(extreme learning machine, ELM)预测典型场景稳定裕度并通过核密度估计(kernel density estimation, KDE)准确获得其概率分布函数。构建了静态电压稳定期望裕度和静态电压稳定风险度两个指标对结果进行表征。分别在New England 39和IEEE300节点系统进行了仿真测试，并将结果与传统蒙特卡洛方法计算结果对比，验证了所提方法的有效性。

Online probabilistic assessment of static voltage stability margin for power systems with a high proportion of renewable energy

The randomness, volatility and weak regulation characteristics of renewable energy have brought new challenges to the static voltage safety and stability of a power system. In view of this, an online probability evaluation method of static voltage stability margin of a power system with a high proportion of renewable energy considering the bilateral uncertainties of source and load is proposed. First, the influence of a large number of renewable energies replacing traditional units on static voltage stability margin is analyzed based on the difference of reactive power regulation characteristics between them. Then the influence of renewable energy output uncertainty on the distribution range of the stability margin is analyzed, and source and load uncertainty models are established to generate typical scenarios. Finally, in order to deal with the rapid fluctuation of stability margin brought by renewable energy, an online probability assessment method of stability margin based on optimized ELM-KDE is proposed. The stability margin of typical scenarios is predicted by an optimized extreme learning machine (ELM), and its probability distribution function is accurately obtained by kernel density estimation (KDE). The expected margin of static voltage stability and the risk of static voltage stability are constructed to characterize the results. Simulation tests are carried out on the New England 39 and IEEE 300 node systems, and the results are compared with the traditional Monte Carlo calculation results to verify the effectiveness of the proposed method.

This work is supported by the National Key Research and Development Program of China (No. 2020YFB0905900).