基于斯托克斯向量的MOCT故障识别模型的研究

针对传统电磁电流互感器故障时因磁饱和造成继保误动或拒动现象，提出了一种磁光电流互感器（MOCT）快速识别故障发生及类型的方法。该方法以通过石榴石型MOCT的偏振光的三个斯托克斯分量作为特征量，建立支持向量机识别模型，模型核参数及惩罚参数由粒子群寻优选取。此种建模方法仅与斯托克斯分量及故障电流等级有关，与线性双折射及环境变量无关，避免了温度等外界因素对识别的干扰。试验结果表明，识别模型受温度影响小，当训练完成的识别模型工作于不同的温度条件下时，识别误差保持在2%以内。运行中的MOCT可以依靠斯托克斯分量及该模型准确识别故障电流等级，迅速判别出故障类型，精度较其他识别方法更高。

Fault identification model of MOCT Based on stokes vector

Due to the phenomenon that traditional electromagnetic current transformer encountered operation rejection and misoperation of the relay protection when it appeared magnetic saturation under failure conditions, this paper proposed a method of the magnetic optical current transformer (MOCT) for fastly identifying fault. In order to establish support vector machine model, the method took three Stokes components of polarized light which passed through garnet optical current sensor as characteristics. It optimized the nuclear and punishment parameters by particle swarm optimization. The model was not associated with linear birefringence or environment variables but Stokes component and fault current grade, so it avoided the interference of external factors. The result shows that temperature has little effect on the identification model. Although the model works at different temperature conditions, the precision of trained recognition model keep inside of 2%. MOCT has higher accuracy than other identification methods. It can accurately identify the fault current grade depending on Stokes component and the model, and quickly identify the type of fault.