基于分数阶Hilbert变换的电路故障诊断研究  

Research on Circuit Fault Diagnosis Based on Fractional Order Hilbert Transformation

作　　者：白莹 BAI Ying(Mengcheng County Zhuangzhou Senior Vocational Middle School,Bozhou 233500,Anhui,China)

机构地区：[1]蒙城县庄周高级职业中学,安徽亳州233500

出　　处：《流体测量与控制》2023年第4期59-63,69,共6页Fluid Measurement & Control

摘　　要：为了解决非平稳信号频谐分析引起的频谱成分重叠问题,以及滚动轴承故障信号的多载波调制特性,结合轴承故障特点,采用一种基于希尔伯特包络和阶次分析的调制阶次提取方法。首先,利用希尔伯特变换求出了一种具有异响问题的车辆的加速度状态,对包络进行等角度的重新取样,得到与速度无关的包络线,并在角域频谱分析中求出调制阶,并根据各个轴承的故障特性,找出故障的根源;其次,利用分数阶希尔伯特变换对被测信号进行弹性特性和细节特性的提取,利用分数阶希尔伯特变换对二次损伤进行特征提取,并利用分数阶希尔伯特变换对待测信号进行幅度、相位和瞬时频率的提取;再次,利用Fisher的序列和遗传算法,找到了一个最优的pn分类值,在抽取了特征点后,利用两种不同的参数向量机对正常的样品进行了诊断;最后,利用相同的方法,找到具有最大识别能力的最优等级p2值,并据此对各种类型的故障进行分类。试验结果证明:所提出的方法能有效地提高系统的故障诊断精度。In order to solve the problem of overlapping spectral components caused by frequency harmonic analysis of non-stationary signal and the multi-carrier modulation characteristics of rolling bearing fault signal,Hilbert envelope and order analysis,combined with the characteristics of bearing fault,a modulation order extraction method based on Hilbert envelope and order analysis is adopted.Then the acceleration state of a vehicle with abnormal noise.Then we resample the envelope at equal angles to obtain the velocity-independent envelope,find the modulation order in the angular domain spectrum analysis,and find out the root cause of the fault according to the fault characteristics of each bearing.Then,the fractional Hilbert transform is used to extract the elastic and detail characteristics of the measured signal,the fractional Hilbert transform,and the fractional Hilbert transform to extract the amplitude,phase and instantaneous frequency.First,using the sequence and genetic algorithms of Fisher,we find an optimal p_(n) classification value,and the classifier has the best identification performance.After extracting the feature points,the normal samples were diagnosed using two different parameter vector machines.Secondly,using the same method,we can find the optimal rank p_(2) value with the maximum recognition ability,and classify the various types of faults accordingly.The test results prove that the proposed method can effectively improve the fault diagnosis accuracy of the system.

关 键 词：分数阶Hilbert变换 FISHER准则 模拟电路故障诊断 特征提取 

分 类 号：TM464[电气工程—电器]