基于神经网络的分接开关均压环结构优化研究

基于110kV无励磁分接开关的结构特点,利用Ansoft软件建立了分接开关均压环在变压器油中的三维数值计算模型,研究了均压环的环径、管径和安装位置对分接开关电场分布影响的规律,并建立了基于BP神经网络的均压环结构参数的数学模型,获得了均压环的结构优化参数。计算结果表明,高低端均压环环径、管径和安装位置对其表面电场分布均有显著的影响,优化后的均压环表面最高场强下降,能在一定程度上降低电晕放电概率,为厂家设计人员提供了参考。

Optimization Structure of Shielding Rings of Tap Changers Based on Neural Network

According to the features of structure in 110 kV tap-changer, three-dimensional (3D) model of shielding rings of tap changer in transformer oil is established by using Ansoft software. This paper studies the effect of ring radius, tube radius and installation position of the shielding rings on electric field distribution. And then based on the BP neural network, the mathematical model of structure parameters about shielding rings is established to obtain the optimized structural parameters of the shielding rings. The results show that external electric field distribution is significantly affected by the ring radius, tube radius and installation position of the shielding rings; the maximal electric field around the shielding rings is obviously decreased; this structure can reduce the probability of corona discharges in a certain extent, which provides references for designers of manufacturers.