Electric field optimization of high voltage electrode based on neural network

For the insulation design of high voltage apparatus (such as gas insulated switchgear and transformers), the electric field optimization technique becomes an indispensable tool. We developed a highly efficient electric field optimization technique based on a neural network under the personal computer. Once the neural network learns the relationship between inputs and outputs, the system easily allows a target solution to be obtained. The principle of the developed optimization technique is described with the calculation example. Moreover, normalization of learning data is introduced and the applicability is expanded.