Low frequencies sound insulation in dwellings
Low frequency noise transmission between dwellings is an increasing problem due to home entertainment systems with enhanced bass responses. The problem is exacerbated since there are not presently available methods of measurement, rating and prediction appropriate for low frequency sound in rooms. A review of the classical theory of sound insulation and room acoustics has shown that both theories are not applicable. In fact, the sound insulation of party walls at low frequencies is strongly dependent on the modal characteristics of the sound fields of the two separated rooms, and of the party wall. Therefore methods originally developed for measurement conditions where the sound field was considered diffuse, may not be appropriate for room configurations with volumes smaller than 50m3 and for frequencies where sound wavelengths are large. An alternative approach is proposed using a Finite Element Method (FEM) to study the sound transmission between rooms. Its reliability depends on the definition of the model, which requires validating measurement. FEM therefore does not replace field or laboratory measurements, but provides complementary parametric surveys not easily obtainable by measurements. The method involves modelling the acoustic field of the two rooms as an Acoustic Finite Element model and the displacement field of the party wall as a Structural Finite Element model. The number of elements for each model was selected by comparing the numerical eigenfrequencies with theoretical values within an acceptable processing time and error. The simulation of a single room and of two coupled rooms, defined by linking the acoustic model with the structural model, were validated by comparing the predicted frequency response with measured response of a 1:4 scale model. The effect of three types of party wall edge condition on sound insulation was investigated: simply supported, clamped, and a combination of clamped and simply supported. It is shown that the frequency trends still can be explained in terms of the classical mechanisms. A thin masonry wall is likely to be mass controlled above 50Hz. A thick wall is stiffness controlled, below 100Hz. A clamped thin wall provides a lower sound insulation than a simply supported, whereas a clamped masonry wall provides greater sound level difference at low frequencies than a simply supported. The sound insulation of masonry walls are shown to be strongly dependent on the acoustical modal characteristics of the connected rooms and of the structural modal characteristics of the party wall. The sound pressure level difference displays a sequence of alternating maxima and minima about a trend, dictated by the properties of the party wall. The sound insulation is lower in equal room than in unequal rooms, whatever the edge conditions and smaller wall areas provide higher sound insulation than large areas. A correction factor is proposed as a function of room configuration and wall area and edge conditions. Attempts to quantify the factor were made using statistical and deterministic analyse, but further work is required.