摘要
为改善变压器调压区电场环境,提出在变压器调压区安装均压环以改善其电场分布。利用有限元法计算了均压环各结构参数时的调压区电场。提出了在均压环参数设计的2种不同的优化方式(常规的分步优化方式和神经网络最优化方式),并加以比较后得出其对最终优化结果的不同影响。在神经网络最优化方式中,通过神经网络构建均压环参数—调压区电场的输入—输出模型,利用最优化方法构建优化算法,并利用程序语言实现完成了变压器调压区均压环结构参数的神经网络最优化设计,真正实现了均压环参数的全局优化设计。
The Voltage Adjusting Zone (VAZ) is critical to transformer voltage regulation. When the voltage level rises, the electric field environment of the VAZ becomes worse. Grading rings for improving the electric field distribution in VAZ are designed, and two methods to optimize the locations of grading ring are proposed. One optimization method is the normal step-by-step calculation method in which the grading parameters are optimized one by on and the other one is based on neural network model which is capable to obtain the global solution.
出处
《江苏电机工程》
2014年第4期48-51,共4页
Jiangsu Electrical Engineering
关键词
变压器调压区
均压环
有限元
分步优化
神经网络
全局优化
grading ring
voltage adjusting zone
finite element
step-by-step calculation
neural network
global optimization