基于LSTM的电力暂态稳定在线评估及预测研究

Research on Online Assessment and Prediction of Power System Transient Stability Based on LSTM

摘要: 电网规模的扩大使得电力系统运行状态变得更加复杂,对电网安全稳定运行提出了更高要求。提出了基于深度学习中长短时记忆(long-and-short term memory,LSTM)的电力暂态稳定在线评估模型。该模型通过获取全网各节点电压、电流、功率等电气量,实时计算得到电网失稳可能性评分,并在新英格兰10机39线系统上对该模型进行测试与优化。实验结果表明,该模型能通过实时运算得到电网稳定性的评估及预警,具有准确性高、预警能力强、支持在线监测的特点。

Abstract: The expansion of the scale of the power grid has made the operating state of the power system more complicated,and put forward higher requirements for the safe and stable operation of the power grid. An online assessment model of power transient stability based on long and short term memory（LSTM）in deep learning is proposed in this paper. The model obtains the voltage, current, power and other electrical quantities of each node in the whole network,and calculates the instability coefficient（result of grid stability evaluation）in real time. The model is tested and optimized on the New England 10-machine 39-wire system. The experimental results show that the model can obtain power grid stability evaluation and early warning through real-time calculations, and has the characteristics of high accuracy,early warning capability,and supporting the online detection.