Goods and products transported by road are subjected to vehicle vibration which, without proper protective packaging, can suffer damage. To reduce shipment costs, protection has to be optimised to limit product damage occurrence while keeping packaging weight and size to a minimum. Optimisation is realized by simulating the vibration of transport vehicles. To achieve an accurate simulation, each vehicle vibration mode has to be modelled. These include: the nonstationary random vibration induced by road roughness and speed variations, the transient vibration created by road surface aberrations and the harmonic vibration created by the vehicle engine and drive train. Identifying and indexing these mixed-modes within complex road vehicle vibration signals is essential to define the severity and occurrence of the different modes in order to develop an accurate model. This paper shows that indexing can be performed using the orthogonal wavelet transform such as Daubechies 10. Assuming that each mode is preponderant in different analysis scales, the wavelet coefficients can be used to perform the indexing. This allows more sensitive mode detection and a more precise time indexing thanks to the multi-resolution nature of the wavelet transform compared to other time-frequency analysis methods.

This paper presents beamforming techniques for source localization on aicraft in flight with a focus on the development at DLR in Germany. Fly-over tests with phased arrays are the only way to localize and analyze the different aerodynamic and engine sources of aircraft in flight. Many of these sources cannot be simulated numerically or in wind-tunnel tests because they they are either unknown or they cannot be resolved properly in model scale. The localization of sound sources on aircraft in flight is performed using large microphone arrays. For the data analysis, the source signals at emission time are reconstructed from the Doppler-shifted microphone data using the measured flight trajectory. Standard beamforming techniques in the frequency domain cannot be applied due transitory nature of the signals, so the data is usually analyzed using a classical beamforming algorithm in the time domain. The spatial resolution and the dynamic range of the source maps can be improved by calculating a deconvolution of the sound source maps with the point spread function of the microphone array. This compensates the imaging properties of the microphone array by eliminating side lobes and aliases. While classical beamfoming yields results that are more qualitative by nature, the deconvolution results can be used to integrate the acoustic power over the different source regions in order to obtain the powers of each source. ranking of the sources. These results can be used to rank the sources, for acoustic trouble shooting, and to assess the potential of noise abatement methods.