基于改进蝙蝠算法优化支持向量机的变压器故障诊断研究

Research on transformer fault diagnosis based on improved bat algorithm optimizing support vector machine

针对传统蝙蝠算法收敛速度慢,求解精度低,易陷入局部最优等缺点,提出一种在速度公式中加入自学习因子并在位置公式中加入比例权重系数的改进方法。利用改进的蝙蝠算法对影响支持向量机分类精度的两个主要参数C和g进行寻优,获得最佳的参数组合并建立故障诊断模型,再结合溶解气体分析(DGA)数据获得故障诊断结果。实验表明,改进后的蝙蝠算法可以加快算法的收敛速度并提高求解精度。通过与传统的蝙蝠算法、粒子群算法、遗传算法寻优SVM获得故障诊断结果相比较,所提改进蝙蝠算法具有更高的故障正判率。

In order to overcome the shortcomings of the traditional bat algorithm,such as slow convergence rate,low solution accuracy,and easy to fall into local optimum,an improved method is proposed in which the self-learning factor is added to velocity formula and proportional weight coefficient is added to position formula. The improved bat algorithm is used to optimize the two main parameters C and g which affect the classification accuracy of support vector machine. The optimal combination of parameters is obtained and the fault diagnosis model is established.The fault diagnosis results are obtained by combining the dissolved gas analysis( DGA) data. Experiments show that the improved bat algorithm can accelerate the convergence speed and improve the accuracy of the algorithm.Compared with the fault diagnosis results obtained by optimizing SVM with traditional bat algorithm,particle swarm optimization algorithm and genetic algorithm,the improved bat algorithm has higher fault diagnosis accuracy.