Support Vector Machines And It's Application In Electric Field Optimization Of Ultra-High Voltage Transformer

Controlling the electric field intensity much smaller than initial partial electric field intensity, is the important part of ultra–high voltage transformer insulation design.It's very important to the manufacturing of UHV transformer.The support vector machine is a new type of the machine learning algorithm after artificial neural network, which is based on statistical learning theory. The data from the sample space is mapped to a high dimensional space by using of nuclear function. The nonlinear problem is converted into linear problem. In this thesis, the support vector machine theory and algorithm are studied in depth.In this paper, it applies support vector machines to train the model. Then application of the least squares support vector machines and simulated annealing algorithm for global optimization to find the minimum size of the maximum field strength under different design sizes in a faster rate. It purposes to meet the design requirements of UHV transformer insulation and structural optimization.This provides a theoretical basis for the design and development of the product .It often repeats calculating by artificial changing the parameters. Corresponding to each result,it needs to make the necessary structural adjustments.The efficiency is low.By using this optimization method, you can get the size when the maximum electric field strength is minimum.It only needs a certain number of groups of datas.It's very efficient.