Identification approaches for source separation in vibration and acoustic problems Seminar

Source identification is a recurrent issue in vibration and acoustic problems. The correct localization, quantification and ranking of sources contributing to an observed acoustic or vibration field is indeed a major challenge in the post-processing and analysis of measurement datasets. Relating observations to sources is an inverse problem. It requires the modeling of the system under consideration, but this is often not sufficient and some a priori information has to be additionally considered so as to make the inverse problem well-posed. Different strategies can be implemented to do so, based on heuristic approaches, optimization algorithms or Bayesian formulations. These strategies will be briefly reviewed, and their implementations will be illustrated in the frame of few experimental applications.