基于模糊神经网络计算VFTO作用下的过电压分布

Overvoltage Distribution Calculation under VFTO Based on Fuzzy Neural Network

为了研究特快速暂态过电压(VFTO)对大型变压器绕组绝缘的影响,笔者提出了基于模糊神经网络的暂态建模算法。首先假设传输线中不存在频变效应,利用模糊神经网络计算大型变压器绕组的分布参数;然后对容易损坏的前几匝线圈每匝为一单元,其余以两匝为一单元建立多导体传输线模型,提出了基于时域有限差分法的暂态建模算法;对于频变效应明显时,运用矢量匹配法实现频变阻抗的有理逼近,提出了含频变参数的暂态建模算法;最后通过仿真实例验证算法的有效性和可行性。

To investigate the effect of transformer windings under VFTO,a new transient modeling algorithm is proposed in this paper.Firstly it is assumed that no frequency-dependent impedance in transmission line,fuzzy neural network is used to compute distributed parameters of large transformer windings.Secondly in the case of easily destroyed former several windings,each is modeled as one unit,others are modeled with two windings as one unit,a transient modeling algorithm based on Finite Difference Time Domain（FDTD） method is put forward.When frequency-changing effect is obvious,Vector Fitting method is adopted to deal with frequency-dependent parameters by expanding time-domain form of internal impedance into a sum of finite exponential functions,which is easy for time-domain convolution,so a transient modeling algorithm considering frequency-dependent parameters is presented.Lastly through simulation illustrations validate effectiveness and feasibility of the presented algorithms.