支持向量机在高压绝缘子污秽程度评定中的应用

高压绝缘子污秽程度的评定可以表述为多类模式识别问题,由于影响绝缘子表面污秽状况的因素复杂,并且实际获得的样本数据有限,因此传统的智能方法往往达不到工程应用的精度要求。研究利用支持向量机在解决小样本、非线性及高维识别中的优势,将可用于多类模式识别的DAG SVM s模型用于高压绝缘子污秽程度的评定。该模型易于实现,且能够找到模式间的最优分类超平面,泛化能力较高。用SVM评定高压绝缘子污秽程度时,可以综合考虑不同污秽程度绝缘子泄漏电流的电气特性、环境参数与绝缘子污秽程度之间的非线性关系,从而实现具有极大模糊性的绝缘子表面污秽程度的评定。结果表明,此方法对解决绝缘子污秽程度的评定问题具有良好的适应性和实用性。

Application of support vector machine in assessing contamination condition of high voltage insulators

Assessing the insulator surface contamination condition can be described as multi-pattern recognition.The traditional methods for assessing this can not achieve the required accuracy for some engineering application due to the limited sample data sets and the complex factors that affect the surface contamination condition of insulator.In the research,based on the full advantage of SVM's ability to solve the problem with relatively few samples and nonlinear and high dimensions,the DAG(Directed Acyclic Graph) SVMs model for multi-pattern recognition is used to assess the contamination condition of the high voltage insulator.The model is easily realized and can find out the super-plane between patterns without local minima,and has strong universal ability.The nonlinear relationship between the electrical characteristics of different contamination condition insulators,the environment factors and the contamination condition is considered synthetically by the SVM method,and the surface contamination condition of insulator is assessed.The results show that the model is suitable to the contamination assessment.