The Mechanical Structure Damage Detection And Localization Based On Modal Models

All kinds of equipments and facilities are developed to large scale, efficient and automatic direction for modern industrial sector e.g. machinery,metallurgy, energy source, aerospace, transportation, building etc. Single equipment or facility generate increasing benefits,thus reliability and safety of the machine running or facility operating are growing concerns. damage and failure will compromise their security and reliability for components and parts of machines and Industrial facilities, structures and structure components,even lead to a major accident casualties. Therefore, health monitoring and quantificational detection of early damage under running and working condition for machines and building facilities are very important problem, there has been a great deal of interesting in health monitoring and damage detection techniques at home and abroad. This paper studies online identification of the modal parameters, and damage detection and its precise locating problems based on vibration response data for machinery equipments under running condition.Modal models describe structure inherent dynamic characteristics, thus structural damage will result in changes in modal parameters. Monitoring changes of modal parameters during operation of the mechanical equipment can detecte occurrence and development of structural damage. Here, modal parameters are obtained by operation modal analysis only using vibration response data, and iuput signals (operating exciting) of white noises are assumed. However, operation excition is often not gaussian white noise, e.g. It is possible that the colored nois which contains of some of the harmonic components, even specific noise which contains of complex spectrum components. It can lead to uncertainty of modal analysis. When the natural frequency is close to or far from harmonic frequencies, study the effect of the harmonic components in the identified modal parameters, and how to deal with harmonic modal.In this paper the analysis and comparison of main characteristics and problems for existing modal analysis methods are performanced firstly in the case of no harmonic component. Using time-domain response data, the identification method of structure modal parameters is studied. In order to ensure the accuracy of parameter identification, the choice of the reference point is studied.A method of the damage detection and localization is proposed by means of modal models obtained from operational modal analysis in this paper. The Mechanical structures may generate local damsge by reason of increasing of the operation time or sudden events. Modal parameters obtained from operational modal analysis may be changed by reason of structural damages. These damags may be found, and damage positions are determined preliminarily in accordance with the change situation of modal parameters (natural frequency and damping ratio) for the different time stages.Structural damage is rapidly detected by monitoring the frequency migration of the corresponding point on the intact and damage structure. This process is automatically performed,to achieve automatical identification of modal parameyers that is required in structural health monitoring.The frequency response curvature and damage detection theory are studied. An improved method of calculating frequency response curvature is proposed. By combinating this method and principal component analysis, damage region is determined.The high density transducers are configured renewedly at components which may occurs damages to increase its spatial resolution. The second modal test is performed to obtain high spatial resolution data of the mode shape. Then normalizing and cubic spline interpolation are performed to make further improvement on the spatial resolution. Finally, the accurate positioning of the damages is achieved by 2D continuous wavelet transform of the mode shape data.