The Mode Analysis Based On ICA And Its Application In Fault Diagnosis

Modal parameters(Natural frequency?damping ratio and modal vibration mode)reflect the dynamic characteristics of structure.Accurate identification of structural modal parameters on the mechanical equipment condition monitoring and fault diagnosis is of great significance.The modal parameter identification method is mainly divided into the traditional experimental modal analysis(EMA)and operational modal analysis(OMA).Experimental modal analysis are mostly based on frequency response function or pulse response function of structure vibration.Operational modal analysis only according to the output of the system response to identify the modal parameters without knowing the excitation of system,such as SSI method and ARMA method,etc.Above analysis methods are mostly affected by noise big,easy with modal noise,require incentive is smooth and white noise,model order difficulties and other issues.ICA is developed in the late 1990 s in the field of data processing and analysis technology,a new purpose from multichannel observations by the number of statistically independent characteristics of source signals of the hidden in the mixed signal independent component(source)signal separation or extracted.Due to the consistent with the theory model of the free vibration response of structure,the ICA for modal parameter identification provides a new train of thought.In view of this,this paper mainly researches the following contents:(1)This paper expounds the research purpose and significance of this topic,briefly introduces the equipment condition monitoring and fault diagnosis technology,and the structure of existing modal parameter identification method of basic theory and research status are reviewed.(2)Introduces in detail the basic principle of ICA,criterion of independence and non Gaussian of random variables have been presented and discussed.And This paper selective mainly introduced several kinds of typical algorithm of ICA.(3)This paper introduced the basic principle of ITD method,and discuss the time delay of ITD method.It is deduced by theoretical analysis that the multiple of the time delay relative to the sampling interval shall be less by half than the multiple of the sampling frequency relative to the system maximum intrinsic frequency.Through numerical simulation and cantilever beam test to confirm the conclusion and analyse the influence of time delay on modal parameter identification.(4)The consistency between structural free vibration model and ICA model is discussed in this paper.The free vibration response signal of structure is decomposed into a number of single degree of freedom decoupling signal by the ICA technology,and then through the Hilbert transform to identify modal parameters.The feasibilityof different ICA algorithm to identify the structural modal parameters under different condition was analyzed through numerical simulation.(5)the SOBI of high identification precison was applied to cantilever beam crack fault experiment.Through the SOBI method to identify modal parameters of cantilever beam before and after the fault,to observe the change of the modal parameters when structure cracks and looseness fault occurs,it provides a powerful reference for fault diagnosis of mechanical equipment.