开关柜合格性检验的力矩奇异性特征分析方法
A Characteristic Analysis Method of Torque Singularity for Switchgear Conformity Inspection
张铁 1杨明达 1邹焱飚1
作者信息
- 1. 华南理工大学机械与汽车工程学院,广东 广州 510640
- 折叠
摘要
面向开关柜的合格性检验,提出一种基于力矩信号奇异性分析的检验算法.根据开关柜的合格性与力矩信号的奇异性特征的相关性,提出将力矩信号的12个统计学特征通过Fisher判别法压缩至5个,并将由该特征组成的样本集输入至SVM分类器中训练出能判别开关柜合格性的SVM模型.最后对所提算法进行了实验验证,结果表明所提算法对开关柜的合格性检验的识别率达100%,延迟在30ms以内,所识别点均落在设定区域内.阻尼干扰下,与小波变换理论算法相比,在兼顾快速响应的同时鲁棒性更好,且不需要大量的训练样本和训练时间.
Abstract
For switchgear conformity inspection,an acceptance test algorithm based on singularity analysis of torque signals was proposed.According to the correlation between the eligibility of the switchgear and the singularity characteristics of torque sig-nal,12 statistical features of torque signal were compressed to 5 by Fisher discriminant method,and the sample set composed of the features was set as input to train the SVM classifier which could distinguish the eligibility of the switchgear after training.Fi-nally,the proposed method was verified by experiments.The experimental results show that the recognition rate of the proposed al-gorithm is100%,and delay within30ms,all the identified points fall within the set area.In the case of damped interference,com-pared with the wavelet transform theory method,it has better robustness while taking into account fast response,and does not need a large number of training samples and training time.
关键词
机器人/合格性检验/力矩/奇异性特征分析Key words
Robot/Acceptance Test/Torque/Characteristic Analysis of Singularity引用本文复制引用
基金项目
广东省科技计划(2020A0103010)
出版年
2024
NETL
NSTL
维普
万方数据