基于分层DVS的电路故障诊断研究分析  

Research and Analysis of Circuit Fault Diagnosis Based on Layered DVS

作　　者：陈英 CHEN Ying(Shenzhen INVT Electric Technology Co.,Ltd.,Shenzhen 518071,China)

机构地区：[1]深圳市英威腾电气股份有限公司,广州深圳518071

出　　处：《通信电源技术》2024年第13期181-184,共4页Telecom Power Technology

摘　　要：为进一步提高电路故障诊断效率,提出利用分层离散沃尔泰拉级数(Discrete Volterra Series,DVS)算法诊断电路故障,使用分层设计简化DVS结构,并引入贝叶斯信息准则(Bayesian Information Criterion,BIC)选择阶数,从而解决特征提取步骤中复杂的DVS参数计算问题。实验结果表明,采用分层DVS方法提取电路故障的特征,可以在0.01 s之内辨识和处理故障,精确检测和识别电路故障。同时,利用分层DVS方法,能够有效地消除电路信号噪声。DVS算法可在55.32 ms时间内发现电路故障,且故障检测精度与准确率均高于蚁群算法和随机森林算法。To further improve the efficiency of circuit fault diagnosis,a layered Discrete Volterra Series(DVS)algorithm is proposed to diagnose circuit faults.The DVS structure is simplified using a layered design,and the Bayesian Information Criterion(BIC)is introduced to select the order,thereby solving the complex DVS parameter calculation problem in the feature extraction step.The experimental results show that using the layered DVS method to extract the characteristics of circuit faults can identify and process faults within 0.01 s,and accurately detect and identify circuit faults.Meanwhile,utilizing the layered DVS method can effectively eliminate circuit signal noise.The DVS algorithm can detect circuit faults within 55.32 ms,and the fault detection accuracy and precision are higher than those of the ant colony algorithm and random forest algorithm.

关 键 词：分层离散沃尔泰拉级数(DVS) 电路故障 频率 信号噪声 

分 类 号：TN707[电子电信—电路与系统]