接触电阻计算新方法

发布时间：2018-03-15 04:34

本文选题：接触电阻　切入点：模糊系统　出处：《湖南大学学报(自然科学版)》2017年08期 　论文类型：期刊论文

【摘要】：运用模糊系统对接触电阻进行建模.首先通过全因子法进行试验,得到足量试验数据,从试验数据中选出一部分能代表整个样本空间特征的数据作为训练数据训练模糊系统,剩下的作为测试数据对模糊系统进行测试,检测系统可靠性.然后在分析基本遗传算法与递推最小二乘算法特点的基础上,将二者结合形成混合算法,混合算法是在基本遗传算法之后进行递推最小二乘算法的运算.通过训练数据分别利用递推最小二乘算法、基本遗传算法、混合算法训练模糊系统,混合算法的收敛效果优于其他2种方法.通过训练数据建立基于回归分析的接触电阻模型.利用测试数据对各模型进行测试,并比较测试结果,混合算法训练模糊系统所得模型的测试效果是各模型中最好的.测试和比较结果说明若能得到足量训练数据,用混合算法训练模糊系统的方法对接触电阻进行建模是可行的.
[Abstract]:In this paper, the contact resistance is modeled by fuzzy system. First, the full factor method is used to get enough test data, and a part of the data representing the spatial characteristics of the whole sample is selected from the test data as the training data to train the fuzzy system. The rest test the fuzzy system as test data to test the reliability of the system. Then on the basis of analyzing the characteristics of the basic genetic algorithm and recursive least squares algorithm, we combine the two algorithms to form a hybrid algorithm. The hybrid algorithm is the recursive least squares algorithm after the basic genetic algorithm. The fuzzy system is trained by the recursive least squares algorithm, the basic genetic algorithm and the hybrid algorithm. The convergence effect of the hybrid algorithm is better than that of the other two methods. The contact resistance model based on regression analysis is established by training data. The test data are used to test each model, and the test results are compared. The results of test and comparison show that if sufficient training data are obtained, it is feasible to use hybrid algorithm to train fuzzy system to model contact resistance.
【作者单位】：平高集团有限公司国家电网高压开关设备绝缘材料实验室;武汉大学电气工程学院;
【基金】：国家自然科学基金资助项目(51477120) 国家863高技术基金资助项目(2014AA051802) 国家电网科技项目‘特高压直流换流站中隔离开关、阀厅接地开关和金具关键技术研究’(SGNXJX00YJJS1400105)~~
【分类号】：TM501.3;TP18

【相似文献】