基于改进深度森林的暂态电压稳定快速评估

Fast Assessment of Transient Voltage Stability Based on Improved Deep Forest

随着高压直流输电工程的密集投运,暂态电压稳定问题日益凸显,对系统安全稳定运行产生了严重的威胁。基于改进的深度森林,提出了一种暂态电压稳定智能化评估方法。通过最大相关最小冗余(maximum correlation minimum redundancy,mRMR)代替多粒度扫描提取强表征特征子集,避免了计算量大、内存占用多问题。然后,对极限梯度提升树(extreme gradient boosting,XGBoost)进行集成以构造新的级联森林,进一步提升模型预测精度。在改进的新英格兰10机39节点测试系统中开展算例分析,结果表明,所提方法具有评估精度高,计算耗时短和鲁棒性强的优点,可辅助电网运行人员在故障后及时预判暂态电压失稳风险,从而提升系统安全稳定运行能力。

With the intensive commissioning of high-voltage direct current(HVDC)projects,the transient voltage problem has become increasingly prominent,which seriously threatens the safe and stable operation of the power system.Based on the improved deep forest,an intelligent assessment method for transient voltage stability is proposed in this paper.By replacing multi-granularity scanning with maximum correlation minimum redundancy(m RMR),a subset of strongly represented features is extracted,which avoids the problems of large computational load and high memory usage.Next,the extreme gradient boosting(XGBoost)is integrated to construct a new cascade forest,which can further improve the prediction accuracy of the model.The case studies have been carried on the improved New England 10 machine and 39 bus system.The results show that the proposed method has high accuracy,short calculation time and strong robustness.It can help grid operators predict the risk of transient voltage instability in time and improve the safe and stable operation of the power system.