基于MRMR-DRSN的电力系统暂态稳定评估

Power System Transient Stability Assessment Method Based on MRMR-DRSN

随着电力系统的广泛互联互通和相量测量单元(phasor measurement unit,PMU)的广泛应用,电力系统的安全运行面临着巨大挑战。为实现对电力系统运行状态快速、准确、有效的评估,提出了一种基于最大相关-最小冗余(max-relevance and min-redundancy,MRMR)准则和深度残差收缩网络(deep residual shrinkage network,DRSN)的暂态稳定评估方法。首先,利用MRMR准则进行特征选择,并将筛选后的关键特征和相应的类标签作为DRSN模型的输入和输出进行离线训练。然后,制定模型更新机制以应对电力系统运行工况变化。最后,基于PMU实时数据和训练好的DRSN,可立即提供暂态稳定评估结果。在IEEE 10机39节点系统上进行测试,结果表明,所提方法相较于其他数据驱动方法的综合评估性能更优异,同时还具有较好的抗噪性能和鲁棒性。

With the wide interconnection of power system and the wide application of phasor measurement unit(PMU),the safe operation of power system faces great challenges.To assess the operation state of power sys⁃tem quickly,accurately and effectively,an assessment method for transient stability based on max-relevance and min-redundancy(MRMR)criteria and deep residual shrinkage network(DRSN)is put forward.Firstly,feature selection is carried out using MRMR criteria,and key features and corresponding class tags are selected as input and output of DRSN model for off-line training.Then,the model updating mechanism is developed to cope with the change of power system operation status.Finally,transient stability assessment results can be provided immediately based on PMU real-time data and trained DRSN.The proposed method is tested on the IEEE 10-machine 39-bus system,and the results show that compared with other data-driven methods,the pro⁃posed method has better performance in comprehensive assessment,and has better anti-noise performance and robustness.