Realisation of the unit Watt in airborne sound

Abstract

The central quantity in regulations concerning acoustics is the sound power emitted by a sound source. Building up from the current state of the art of sound power determination, this work will show that there is potential to build a metrological traceability chain for this measurand. Firstly, the lack thereof will be discussed and the problems that arise from this situation will be detailed (Chp. 2). This will argue the need for a proper traceability chain including a primary, secondary and transfer standards. This work will focus on the level of the primary standard, which provides a foundation to the desired traceability chain. The task requires a significant shift in measurement protocol used. All currently standardised procedures require measurements in the sound field - while also sampling the sound field quantities sound pressure or intensity. The newly introduced procedure places the measurement equipment outside of the sound field and measures the motion of the primary standard’s surface directly. The measurement protocols and environments as well as the used physical realisations of primary sources will be described (Chp. 3). Certainly, other types or designs of primary standards are possible. The structural as well as material choices for the realisations described in this work are based on analytical and numerical studies on the characteristics of the chosen circular piston design. They will be described in detail (Chp. 4). Besides radiation characteristics, the design choices for the primary standards used are obviously governed by the suitability to the measurement method. This methodology builds on the Rayleigh integral in its discretised form. By measuring surface velocities as well as phase relations at pre-defined points of the measurement surface, the emitted sound power is calculated. The boundary condition that is imposed by this choice of measurement method is that of the need of a planar radiator which has to be embedded into the floor of the measurement environments. The discretised Rayleigh integral method is a validated methods so that the main question of interest for this work is whether its convergence can be guaranteed and, if so, how many sampling points are required to approximate this convergent value with a pre-set level of uncertainty. Both of these aspects will be discussed (Chp. 5). With the theoretical framework in place, measurement data will be evaluated. As expected, these introduce a new aspects into the uncertainty calculations: noise. The evaluation of the uncertainties associated with calculated sound power levels is a central topic determining the usability of the described primary standards. Based on Monte Carlo simulations, strategies to determine uncertainties will be developed. They focus on the aspects of filtration of noisy data and usability of measurement surfaces - piston versus baffle. It will be shown that in terms of precision, the proposed discretised Rayleigh integral method has the potential to yield satisfactory results (Chp. 6). In terms of…