“学萃讲坛”秉承学名家风范、萃科技精华的理念，以学术为魂，以育人为本，追求技术创新，提升学术品位，营造浓郁学术氛围，共品科技饕餮盛宴！

时间：2016年10月13日上午08:30-09:30

地点：水声楼15楼报告厅

报告题目：Source identification, Defect detection and material characterization by using the dynamic local motion equation of a structure

报告人：Charles PEZERAT教授

主办单位：水声工程学院

报告人简介：Charles PEZERAT, the vice-chancellor of Le Mans University in France, Institut National des Sciences Appliquées de Lyon. His main research interests are about mechanical engineering, structural engineering and acoustic engineering

报告内容简介：The vibroacoustics activities of the LAUM contain the three main topics: Experimental characterization methods, modelling, vibration and acoustics passive controls. After a brief presentation of the team, an overview of PhD thesis subjects and industrial collaborations in each topic are described. The proposed scientific focus (second part of the presentation) is on a vibration inverse problem development that corresponds to several interests of industries in automotive, naval and aeronautics sectors.

Identification of vibration sources, defects and/or material properties consists generally in solving inverse problems. The called RIFF method (French acronym meaning Windowed and Filtered Inverse Solving) is one way to solve this kind of inverse problem. The basic principle of the RIFF approach consists in measuring vibration displacement on a meshgrid in a local area of interest, injecting measured data in the motion equation and calculating the searched unknown. Compared to other usual inverse techniques, the RIFF method has the curious particularity of needing the knowledge of the local motion equation only. Boundary conditions, sources or dynamic behaviors outside the area of interest can be completely ignored, whereas they are required for the direct problem solving. The searched unknown can then be identified locally with respect to the frequency and can be mapped by using a scanning process of the area of interest. However, as in all inverse methods, the RIFF approach is very sensitive to uncertainties in measured data, so that it is practically impossible to apply it without a regularization technique. Several approaches based on the use of low-pass wavenumber filtering are presented. Initially developed for vibration source identification, the first examples shown concern location and quantification of point forces and/or moments. A special focus is also proposed on aeroacoustic excitation identification. Since the method corresponds to a local verification of the motion equation, defects can be located. An experimental example on a notch detection is shown on a plate. Finally, the recent development on identification of elastic and damping material properties is presented. Examples showing the possibility to characterize composite materials in medium and high frequency ranges and to realize maps of material characteristics are presented. Actual works and perspectives in industrial applications are discussed at the end of the presentation.