基于改进ANFIS算法的故障分类

Fault Classification Based on the Improved ANFIS Algorithm

高压输电线路是电网的骨架,发生故障如不能及时排除会造成很大的影响,因此在故障定位之前首先要确定故障类型。利用比例共轭梯度法(Scaled Conjugated Gradient,SCG)对ANFIS(自适应神经模糊推理系统)算法进行改进,对线路发生的故障类型进行分类。利用MATLAB软件对改进ANFIS算法与标准ANFIS算法进行对比分析,得出改进ANFIS算法收敛速度更快,所需的训练时间和训练步数也更少,并且其故障分类结果也更加准确。

The high voltage transmission line is the skeleton of the power grid, faults of which, if not eliminated in time, would have a severe impact on the system. Therefore, it is necessary to determine the fault type before fault can be located. The improved ANFIS （Adaptive Neural Fuzzy Inference System） , based on the Scaled Conjugate Gradient （SCG） is used to classify the fault type of the line. MATLAB is used to compare the improved ANFIS algorithm with traditional ANFIS algorithm. Thus, it is concluded that the improved ANFIS algorithm has faster convergence rate, less training time and training steps, and more accurate fault classification result.