Acoustic source reconstruction of a commercial generator set using inverse pellicular modes and inverse numerical panel contribution analysis

Many methods have been used to reconstruct a noise source from measured sound pressure information in the field. Most of these approaches are used for source visualization. However, the inverse boundary and finite element methods are reconstruction methods that are more useful for sound field predictions. The current research examines two inverse numerical simulation approaches, inverse pellicular modes analysis and inverse numerical panel contribution analysis, which are variations of the inverse finite element method. The first of these relates the sound pressure in the field to pellicular modes. Pellicular modes are generated by adding a thin layer of acoustic finite elements around the boundary and determining the acoustic modes in that thin layer. These nonphysical modes are then used as if they were structural modes and serve as the basis functions for the reconstruction. The second method divides the boundary surface of the source into sets of panels. Transfer functions can be calculated relating the panel vibration to the sound pressure at the measurement locations. These transfer functions are the basis functions for the reconstruction. In the current work, both inverse pellicular modes analysis and inverse numerical panel contribution analysis are used to reconstruct a commercial generator source and to predict the sound pressure level 7.0 m away from the source. The number of measurements required for an adequate reconstruction is investigated for both methods. The results suggest that both methods can be used to reconstruct a complicated source.