The vibration transmission loss at junctions including a column

Transcription

1 The vibration transmission loss at junctions including a column C. Crispin, B. Ingelaere, M. Van Damme, D. Wuyts and M. Blasco Belgian Building Research Institute, Lozenberg, 7, B-19 Sint-Stevens-Woluwe, Belgium, charlotte.crispin@bbri.be Some numerical predictions of vibration transmission loss at junctions including a column were carried out by different laboratories. The computed results show an increase of the transmission loss at high frequencies for the bending waves. This increase is smaller at low frequencies and the transmission loss can even decrease in the case of a collinear path through the junction. A measurement campaign on the vibration reduction index K ij was led in the BBRI s laboratory on such junctions in order to verify these observations and to establish a draft for empirical formulae allowing the K ij prediction. Different configurations were tested as the thickness of the column and the types of walls. This article presents the results and a discussion of these measurements. 1 Introduction In Belgium, the standard NBN S1-4 which gives the minimal threshold values of acoustic insulation between rooms will be revised. The recommended insulation level will be increased in order to satisfy 7% of the occupants for a basic comfort and 9% of them for a luxury comfort. These new requirements demand a modification of the current construction methods. The BBRI s mission is to establish building guidelines which will help the professionals of the construction to reach these new requirements. Currently, a good insulation of the partition wall is often sufficient to fulfill the old acoustic requirements between apartments. Nevertheless, according to the future standard, where higher insulation levels will have to be reached, the noise transmission through the lateral partitions will be of major importance. A research is in progress to the laboratory of the BBRI in order to study these transmission paths and the practical means to reduce them. This research concentrates more particularly on the application of resilient joints or columns in building construction. The effect of columns on the transmission loss through junctions was studied theoretically and discussed by different researchers [1], [], [], but few measurements have already been carried out on full scale building elements in laboratory. A measurement campaign of the vibration reduction index K ij was led in the BBRI s laboratory on several different junctions containing a column in order to verify the theoretical observations and to establish a draft for empirical formulae allowing the K ij prediction. Measurement set-up and results.1 Set up Measurements of the vibration reduction index K ij have been performed according to the method described in the draft pren (1) and 4 (4) [4], [5]. This normalized quantity is related to the vibrational power transmission over a junction. It is expressed in decibels and is given by the following equation (1): ij ij = Dv,ij lg [db] (1) aia j K + In this equation, D v, ij is the direction averaged velocity level difference in db, a i (resp. a j ) is the equivalent absorption length of the wall i in metres (resp. wall j) and l ij is the junction length in metres. The stationary excitation was obtained by multiple hammer hits. The velocity levels were measured simultaneously on the emission wall (i) and on the reception wall (j) with accelerometers. The hammer was moved to three different positions. For each position, four couples of transducer positions were used. The structural reverberation time was calculated as the average value out of twelve decay curves. For each test, the validity of the results was checked according to the terms of the This verification revealed that there were less than 5 modes per onethird octave frequency band between and Hz for the tested concrete walls and a modal overlap factor lower than.5 below 5 Hz for these walls. However, the walls had at least one mode per one-third octave frequency band above 1 Hz. So, the analysis was acceptable above this frequency. For information only, the measurements were compared to the results obtained by a SEA simulation []. l 197

2 Forum Acusticum 5 Budapest. Measurement results The first tested junction (junction (a)) is a T-junction. The coplanar walls are composed of concrete blocks of 19 cm (S 1 = 9.5 m², S = 9. m², E = 1.E+ N/m², ρ = 51.8 kg/m²). The perpendicular wall is made of cellular concrete of cm (S = 8.4 m², E = 1.1E+9 N/m², ρ =.4 kg/m²). The column has a square crosssection of.4 m² and a height of.4 m. The figures 1a and 1b show the measurements of the vibration reduction index K ij with and without column for an inline and a corner transmission path respectively. The measurements without column are taken up from a previous measurements campaign [7]. These figures present also the SEA simulations Figure 1a : Junction (a) In-line transmission path In the axis, the measurement for the junction with column is close to the measurement for the junction without column until 15 Hz ( Hz according to the SEA simulation). In the angle, the measurement of the K ij through the column shows a fall compared with the junction without column. This fall increases with increasing frequency up to 15 Hz. At this frequency, which depends on the moment of inertia of the column, the K ij increases. The SEA simulation gives the same trend but the minimal value is situated around 5 Hz. The same junction, but with a larger column crosssection, is shown in figure a and b (junction (b)). The column has a square cross-section of.1 m². In the axis, there is a marked improvement of the K ij from 5 Hz compared to the measurements without column. We can see an appreciable increase of the K ij from 4 Hz for the corner transmission path by doubling the thickness of the column. This trend is also noticed on the SEA simulation. Nevertheless, the measurement of the K ij with the square column of 4 cm *4 cm stays lower than the same junction without column! Figure a : Junction (b) In-line transmission path 1 Figure 1b : Junction (a) Corner transmission path Measurement (With Column ) SEA (With Column) Measurement (Without Column) SEA (Without Column) Figure b : Junction (b) Corner transmission path 198

3 Forum Acusticum 5 Budapest The third junction (junction (c)) is the same as the second junction apart from the fact that the perpendicular wall is replaced by a wall of concrete blocks (S = 8.4 m², h =.19 m, E = 1.1E+ N/m², ρ = 51.8 kg/m²). The cross-section of the concrete column is.1 m². The results are shown in figure a and b. 1 frequency. This dependency, which is confirmed by the SEA simulations, shows an improvement by log(f) for the K ij up to 1 Hz. Above 1 Hz, the K ij increases more rapidly which can be explained by the effect of the distance Figure 4a : Junction (d) In-line transmission path Figure a : Junction (c) In-line transmission path Figure 4b : Junction (d) Corner transmission path Figure b : Junction (c) Corner transmission path The measurements are in good agreement with the SEA simulations. The K ij for the corner transmission path shows a minimum value near 4 Hz (5 Hz for the simulation SEA). Above this frequency, the K ij increases with increasing frequency. The effect of the column is noticed from Hz according to the SEA prediction. The same trend can be observed for the inline transmission path. In the fourth T-junction (junction (d)), the three walls are composed of bricks. (S 1 = 9.5 m², S = 9. m², S = 8.4 m², h =. m, E = 8.9E+9 N/m², ρ = 11.4 kg/m²). The cross-section of the concrete column is still.1 m². The measurements in-line and in the angle show that in this case the K ij depends on the Measurement (With Column ) SEA (With Column) SEA (Without Column) The three walls of the following junction (junction (e)) have a surface mass lighter than the previous one. The walls are composed of cellular concrete blocks (S 1 = 9.5 m², S = 9. m², S = 8.4 m², h =. m, E = 1.E+9 N/m², ρ =.4 kg/m²). On the figure 5, we can see an important improvement of the K ij compared to the junction without column for both transmission paths. We can also observe a dependence of the K ij by *log(f) when a column is included in the junction. 199

4 Forum Acusticum 5 Budapest Figure 5a : Junction (e) In-line transmission path Figure 5b : Junction (e) Corner transmission path Measurement (With Column ) SEA (With Column) Measurement (Without Column) SEA (Without Column) In figure are shown the measured and simulated K ij for a transmission at a corner junction (Junction (f)). This junction is composed of a cellular concrete wall (S 1 = 8.4 m², h =.1 m, E = 1.1E+9 N/m², ρ =.4 kg/m²) and of a concrete blocs wall (S = 9.5 m², h =.19 m, E = 1.1E+ N/m², ρ = 51.8 kg/m²). The concrete column connecting the two walls has a square cross-section of.4 m² and a height of.4 m. The measurement shows a slight decrease of the K ij with increasing frequency till 15 Hz where the curve reaches a minimum value of 4.7 db. Above 15 Hz, the vibration reduction index increases with increasing frequency. This transition frequency depends on moments of inertia of the column and so by its thickness. The SEA simulation which gives a trend similar to the measurement shows that the inclusion of a square column decreases the K ij Figure : Junction (f) Corner transmission path Discussion and conclusion The figure 7 presents the measured K ij (The difference between the junction with column and without column) for the junctions (a) and (b) which are different by the thickness of their column. This figure shows that the K ij increases from 4 Hz for the corner transmission path by doubling the thickness of the column. Nevertheless, the measurement of the K ij stays a negative value! Column thickness effect Figure 7 : Effect of the insertion of a column with a cross-section of cm² and a cross-section of 4 4 cm² for the corner transmission path For the in-line transmission path, there is a marked improvement of the K ij for the junction with a column of cm² from 5 Hz compared to the measurements without column (figure 8). The

5 Forum Acusticum 5 Budapest improvement obtained with the square column of 4 4 cm² is close to 5 db from 5 Hz Column thickness effect Figure 8 : Effect of the insertion of a column with a cross-section of cm² and a cross-section of 4 4 cm² for the in-line transmission path So, The measurement results of the K ij for the junctions composed with heavy walls (junctions (a), (b) and (c)) show the same trend as the numerical simulations reached by J.A. Steel [1]: the effect produced by a column for a corner transmission path leads to a light decrease of the transmission loss for the bending waves in low frequency. In the axis, the measurements show that the column has no effect in the low frequencies but, in high frequencies, the K ij increases strongly. However, when the walls have densities lower than the column (Junction (d) and (e)), the measurements show an increase of the K ij for all frequencies. The figures 9 and present the effect of the densities contrast between the walls and the column for the transmission path in the angle and in the axis respectively. On these figures, the K ij is the one obtained by the SEA simulation because for some cases we don t have the measurement without column. In both cases (the in-line and the corner transmissions), the two main observations are: The more the walls have light surface mass the more K ij is high and when a column is included in a junction composed with light walls compared to this one, a frequency dependency appears. This dependency is more or less by log (f). An empirical formula has been developed from these results but there are not yet enough measurements to validate it. Some measurements are already planned with columns of different shapes Figure 9 : Wall density effect on the K ij of a T-junction composed with a concrete column of 4 4 cm² for the in-line transmission path Figure : Wall density effect on the K ij of a T- junction composed with a concrete column of 4 4 cm² for the corner transmission path ρ 1,, = 5 kg/m² ρ 1, = 5 kg/m², ρ = kg/m² ρ 1,, = 11 kg/m² ρ 1,, = kg/m² 4 Acknowledgements The authors wish to acknowledge the financial assistance of the Belgian Ministry of Economic Affairs. 1941

6 Forum Acusticum 5 Budapest 5 REFERENCES [1] J.A. Steel, Sound transmission between plates in framed structures. Journal of sound and vibration, 178(), (1994). [] Cremer, L., Heckl, M., Structure-Borne Sound. Structural Vibrations and Sound Radiation at Audio Frequencies. (Springer-Verlag, Berlin, 1988). [] R. J. M. Craik, Sound transmission through buildings using statistical energy analysis. (Ashgate, 1) [4] pren ISO/DIS (1). Acoustics Laboratory measurement of the flanking transmission of airborne and impact noise between adjoining rooms Part 1 : Frame document. [5] pren ISO (4). Acoustics Laboratory measurement of the flanking transmission of airborne and impact noise between adjoining rooms Part 4 : All other cases. [] E. Sarradj, Software Free SEA.91 (). [7] C. Crispin, C. Mertens, M. Blasco, B. Ingelaere, M. Van Damme, D. Wuyts, The vibration reduction index Kij: laboratory measurements versus predictions EN (), Proceedings of INTERNOISE 4, Prague, Czech Republic. 194