为精准识别台区的线损异常,保证配电网经济、稳定运行,针对台区线损的异常情况,提出一种基于二阶聚类和鲁棒性随机分割森林(robust random cut forest,RRCF)算法的台区线损异常检测方法。首先,运用二阶聚类将台区不同的运行工况进行聚类...为精准识别台区的线损异常,保证配电网经济、稳定运行,针对台区线损的异常情况,提出一种基于二阶聚类和鲁棒性随机分割森林(robust random cut forest,RRCF)算法的台区线损异常检测方法。首先,运用二阶聚类将台区不同的运行工况进行聚类,将相同工况的线损节点归并,然后将各类工况的节点线损数据导入RRCF算法中分析,通过删除和插入样本节点,并对插入节点后评判模型的复杂度进行计算,得到线损异常节点的评分值,进一步找出线损异常的节点。最终,通过有关实例验证所提方法的准确性与有效性。展开更多
A symbol level secure precoding scheme based on band-region constraint of the eavesdropper’s receiving signal is proposed to enhance the energy efficiency of cell-free multiple-input multiple-output(MIMO)networks in ...A symbol level secure precoding scheme based on band-region constraint of the eavesdropper’s receiving signal is proposed to enhance the energy efficiency of cell-free multiple-input multiple-output(MIMO)networks in the presence of an eavesdropper while guaranteeing the quality of service(QoS)of user and the security of system.Moreover,to lighten its high computational complexity,original problem is divided into several cascade sub-problems firstly,and then those sub-problems are handled by combining Lagrangian dual function and improved Hooke-Jeeves method together.Comparative ex-periment with other secure symbol-level precoding schemes demonstrate that proposed scheme can achieve the lower power consumption with almost same symbol error rate and QoS of user.展开更多
文摘为精准识别台区的线损异常,保证配电网经济、稳定运行,针对台区线损的异常情况,提出一种基于二阶聚类和鲁棒性随机分割森林(robust random cut forest,RRCF)算法的台区线损异常检测方法。首先,运用二阶聚类将台区不同的运行工况进行聚类,将相同工况的线损节点归并,然后将各类工况的节点线损数据导入RRCF算法中分析,通过删除和插入样本节点,并对插入节点后评判模型的复杂度进行计算,得到线损异常节点的评分值,进一步找出线损异常的节点。最终,通过有关实例验证所提方法的准确性与有效性。
基金the National Natural Science Foundation of China(No.61976080)the Key Research Projects in Henan Province of China(No.231111212500).
文摘A symbol level secure precoding scheme based on band-region constraint of the eavesdropper’s receiving signal is proposed to enhance the energy efficiency of cell-free multiple-input multiple-output(MIMO)networks in the presence of an eavesdropper while guaranteeing the quality of service(QoS)of user and the security of system.Moreover,to lighten its high computational complexity,original problem is divided into several cascade sub-problems firstly,and then those sub-problems are handled by combining Lagrangian dual function and improved Hooke-Jeeves method together.Comparative ex-periment with other secure symbol-level precoding schemes demonstrate that proposed scheme can achieve the lower power consumption with almost same symbol error rate and QoS of user.
