Acoustic power transfer (APT) has become a strategic stake with the increase of wireless smart devices located in electromagnetic proof environments. This is especially the case for medical implants where the transferred power has to go through human tissues, itself made of different layers (epidermis, dermis, muscle, bones), and through the metallic case of the medical implant. Each of these layers are mechanically coupled, making the optimization of the emitter and receiver of such system a complex task. To overcome this issue, we developed a procedure based on Pspice to accurately simulate APT between an emitter and a receiver, as it propagates through an intermediate layer. The procedure is based on Leach model which is the Pspice equivalent model of Mason electro-mechanical analogy. In this article, we propose a model which fits two 3-layer APT experiments, consisting in piezoelectric emitters and receivers with an intermediate layer of alternatively titanium and tungsten. The results provided by the simulations give a qualitative fitting of the APT and would enable to design, study and optimize n-layer APT systems for applications such as wireless medical implant recharging.
Aircraft interior noise is a major design stake with respect to airline requirements for passenger and crew comfort. The focus was put here on cockpit noise. The goal of this study was to evaluate unpleasantness of such sounds and to propose metrics for its prediction. Six sounds recorded in various airplanes were used; one of them was modified in order to reduce the level of an emerging low-frequency component, leading to a number of seven stimuli. These stimuli were used in a pair-comparison experiment in a sound-proof booth. 31 listeners participated to the experiment. Results could be analysed using a BTL model, which provided an unpleasantness scaling of the stimuli. Unpleasantness was clearly related to loudness; due to the high level in the very low frequency bands, A-weighted level failed to represent loudness. Two loudness models were also used: the ISO-532 one and the ANSI S3.4 one. It appeared that the former gave better results. It is argued that this is due to the evolution of loudness values in the low frequency range, which is different between these two models.