特高压传输距离较远且工作环境复杂,容易发生故障,线路保护动作正确率不高.基于±800 kV特高压直流输电工程,将目标探测算法与特高压直流保护动作评价相结合,提出了一种基于YOLOv3算法的特高压直流保护动作评价方法.首先分析了影响线路继电保护动作行为的因素;然后构建了基于YOLOv3算法的特高压直流线路保护动作评价网络结构,开展基于PSCAD(power systems computer aided design)的特高压直流线路保护动作行为仿真,生成网络训练和测试所需的数据样本集,再利用训练样本集对评价网络进行训练;最后对该网络进行了测试.测试结果表明,该方法可以正确评价特高压直流线路保护的动作行为,具有学习能力强、速度快、精度高等优点,解决了传统人工评价人力不足的问题.
Evaluation of UHVDC line protection action based on YOLOv3 algorithm
The transmission distance of UHV(ultra-high voltage)transmission is long,and the working environment is complex,so it is easy to fail,and the correct rate of line protection action is not high.In this paper,an evaluation method of UHVDC(ultra-high voltage direct current)protection action based on YOLOv3 algorithm is proposed,considering the combination of target detection algorithm and evaluation of UHVDC protection action for±800 kV UHVDC transmission projects.Firstly,the factors influencing the action behaviour of line relay protection are analysed.Secondly,an evaluation network structure of UHVDC line protective action based on YOLOv3 algorithm is constructed,the protection action simulation of UHVDC line based on power systems computer aided design(PSCAD)is carried out.The data sample set required for network training and testing is generated,and then the evaluation network is trained by using the training sample set.Finally,the network is tested.The test results show that this method can correctly evaluate the action behaviour of UHVDC line protection,and has the advantages of strong learning ability,fast speed and high accuracy,which solves the problem of insufficient traditional manual evaluation manpower.
relay protectionprotection action evaluationultra-high voltage direct current(UHVDC)YOLOv3 algorithmpower systems computer aided design(PSCAD)simulation
NETL
NSTL
万方数据
