Conference_programme: 18: Room acoustics
Lecture: Comparing Measurement and Simulation Results using Model Calibration for Room Acoustical Evaluation of Industry Halls.
Author(s): Ziegler Thomas
Summary:
Powerful simulation tools have become standard for room acoustical prediction but results are often questionable due to inaccurate or incomplete input data and user-unconsciousness on the limitations of geometrical acoustics. In case of renovation projects we can alleviate the problem of inaccurate input data by calibrating the simulation model to approximate acoustical measurements of the untreated room before predicting the effect of the acoustical treatment. \nExtending preceding comparative studies of industry halls the present paper focuses on the limitations of the model calibration approach by varying input parameters like scattering coefficients and level of detail of internal fittings. Sound pressure level, Dl2 and reverberation time of a mid-sized, flat industry hall with non mixing geometry and dense internal fittings are measured and simulated with CATT-Acoustic. Results using model calibration show that mean sound pressure level over measurement points, LAeq per measurement point, reverberation time and Dl2 after the acoustical treatment can be predicted with reasonable accuracy. Octave-band sound pressure levels at single measurement points, however, show significant prediction errors. Reasons for deviations in predictions are discussed in detail. \nA simple, least-effort empty room scenario is motivated and compared to a simulation scenario with internal fittings. Though accurate predictions with the empty room scenario are feasible, it shows significantly higher sensitivity with respect to variation of scattering coefficients. Somewhat attenuated, this conclusion holds true for prediction of sound pressure levels. Furthermore, we show that rough model calibration according to reverberation time does not cause higher prediction errors as compared to detailed model calibration according to sound pressure levels taking into account room properties in the vicinity of measurement points. Calibration and prediction errors only weakly correlate in case of detailed model calibration.
Corresponding author
Name: Dr Thomas Ziegler
e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
Country: Austria
