Predictions and measurements for lightweight constructions and low frequencies C. Guigou-Carter, M. Villot CSTB

Transcription

1 Predictions and measurements for lightweight constructions and low frequencies C. Guigou-Carter, M. Villot CSTB EUONOISE 2012 Prague June 2012 PAGE 1

2 Introduction For lightweight constructions, EN standardized methods for predicting building acoustic performances from the performances of building elements based on first order SEA have to be reconsidered Most direct method for acoustic performance prediction is to establish database of measured flanking sound transmission factors for different combinations of lightweight elements and junctions Individual flanking transmission path can be performed in dedicated facility by using sound intensity or by shielding all of the flanking elements other than the ones being evaluated Advantage : using measured values Major difficulty : extension to low frequencies since it is difficult to adequately shield alternative flanking transmission paths Another problem : not possible to predict the performance of constructions which have not already been tested EUONOISE 2012 Prague June 2012 PAGE 2

3 Introduction Second method of calculating the flanking sound transmission factors of lightweight building constructions involves extending the existing EN method Method has to be re-assessed for lightweight constructions mainly due to presence non-uniform vibration fields, relatively high attenuation and non-resonant fields Discussions within the COST Action FP0702 between different European countries on work and projects on lightweight construction have been quite productive and successful in order to propose a predicting method Laboratory and in-situ measurements have been and are being carried out on wood frame lightweight buildings and some of their currently used elements. Proposed prediction is being tested on more and more realistic situations EUONOISE 2012 Prague June 2012 PAGE 3

4 Model Description COST Proposal Flanking sound reduction index ij and flanking impact sound level L n,ij from element i in the source room to element j in the receiving room can be expressed as ij * i 2 * j D vs,ij 2 D vs,ji 10log S s S S i j L n,ij L n,ii * j 2 * i D vs,ij 2 D vs, ji 10log S S i j * i and * j are sound reduction indices, referring to resonant transmission only (linings i and j omitted for clarity) D vs,ij is vibration level difference between elements i and j, when element i is mechanically excited S element surfaces (S s considered) for element separating the two rooms L n,ii normalized impact sound level of element i EUONOISE 2012 Prague June 2012 PAGE 4

5 Model Description Vibration level difference Vibration level difference measured in laboratory and in situ (data base) Finite element modeling is also investigated to determined vibration level difference Classes of junction with average vibration level difference used in prediction, proposed by CSTB New invariant proposed by COST : normalized direction average velocity level difference D v,ij,n D v,ij 2 D v, ji 10log S l ij m,i S m, j * * i j j * i * D 10log( l / S ) L L D 10log( l / S ) ij v,ij,n ij S n,ij n,ii v,ij,n ij i 2 2 EUONOISE 2012 Prague June 2012 PAGE 5

6 Junction characterization CSTB measuring method 12 measurement positions on vertical wall 12 measurement positions on floor 12 measurement positions on floor Excitation on floor : tapping machine at 3 positions velocity measurements on each wall sequentially averaged on 20 s Energy average of each wall velocity level to determine junction velocity level difference Excitation on vertical wall : rain-on-the-roof excitation with hammer (2hits/s) over complete wall for 20 s Velocity level difference evaluated for each pair of sensors (one on each wall) Junction velocity level difference by averaging each velocity level difference EUONOISE 2012 Prague June 2012 PAGE 6

7 D v,n,ij (db) Te junction Separating wall - Floor Top Wall-Floor Path - B1 Top Wall-Floor Path - B2 Wall-Wall Path - B1 Wall-Wall Path - B2 Bottom Wall-Floor Path - B1 Bottom Wall-Floor Path - B2 Wall - Wall Top Wall - Floor 30 Bottom Wall - Floor Frequency (Hz) Floor joist perpendicular to wall Same junction in 2 different buildings eproducibility of measurement method as well as junction construction More complex junction less reproducibility EUONOISE 2012 Prague June 2012 PAGE 7

8 D v,n,ij (db) Te junction Façade-Floor Façade-Façade Path - Floor joist // to façade Façade-Façade Path - Floor joist to façade Façade - Façade Frequency (Hz) At low frequencies, façade-façade flanking path more important when floor joist parallel to façade since vibrational energy remains mostly in façade When floor joists perpendicular to façade, some vibrational energy from the excited wall façade can be transmitted to floor through the joist/studs-wall connections EUONOISE 2012 Prague June 2012 PAGE 8

9 Junction characterization Finite element modeling FEM approach to evaluate vibrational level difference in order to avoid quite long, cumbersome and expensive characterization on real junction Effect of the excitation type, i.e. air-borne or structural excitation, will also be investigated numerically on the normalized direction average vibrational level difference Prediction results and parametric study will be soon available EUONOISE 2012 Prague June 2012 PAGE 9

10 Model Description Correction factor Correction factor for * based on element radiation efficiencies for airborne excitation a and structural excitation s proposed by COST * 10log( a s log( Correction factor is more important at frequencies much smaller than the critical frequency of the element considered Correction factor from measured radiation efficiencies on different types of lightweight elements including floor and wall proposed by CSTB Another approach has been proposed: using measured sound reduction indices and velocity level difference D va,ij evaluated for acoustical excitation; radiation efficiencies also necessary and final expression for ij practically the same s a ) a s ) EUONOISE 2012 Prague June 2012 PAGE 10

11 Correction *- (db) Model Description Correction factor evaluation Measurement method proposed by CSTB for radiation efficiency 10log L 6 L p v 10log (A S) 30 Measurement - Partition #1 - Board side L p spatially averaged sound level radiated in receiving room according to EN ISO Measurement - Partition #2 - Board side Measurement - Floor - Board side L v spatially averages vibration level of tested element (16 measurement points) A equivalent absorption area of receiving room 10 S the element surface Excitation : airborne (EN ISO 140-3) or uniformly distributed structure-borne using tapping machine at 3 positions (EN ISO ) or rain-on-the-roof hammer impacts 0-10 Frequency (Hz) Based on measured results, correction factor (roughly) is simplified to : -below element critical frequency : 10 db -at/above element critical frequency : 0 db EUONOISE 2012 Prague June 2012 PAGE 11

12 On-going Project in FANCE ACOUBOIS Provide a prediction method for evaluating acoustic performance of lightweight buildings at the conception stage Database for acoustic performance of lightweight elements (wall, floor, façade, roof) fulfilling actual regulations (thermal, fire, humidity, etc...) for frequency range Hz In-situ measurements on 30 buildings for frequency range Hz to validate prediction method, as well as low frequency impact sound with heavy soft rubber ball Survey with questionnaire to evaluate occupants comfort also included Junction characterization by measuring vibration level difference on 8 buildings for frequency range Hz Examples of typical apartment building constructions fulfilling actual regulations Project Partners : CSTB, FCBA, QUALITEL EUONOISE 2012 Prague June 2012 PAGE 12

13 In-situ Measurements Predictions In-situ measurements between 50 and 5000 Hz in octave and third octave bands Measurement method based on ISO/WD for airborne sound Measurement method based on EN ISO for impact sound Measurement method based on EN ISO for soft rubber ball impact sound everberation measured in reception room, associated correction limited to 5 db (T r 1.6 s) Acoustic performance laboratory measurement of building components not yet fully performed Predictions based on acoustic performance of similar walls/floors for Hz, with numerical extension at low frequencies Hz EUONOISE 2012 Prague June 2012 PAGE 13

14 D nt (db) Horizontal airborne sound insulation Main transmission path related to direct path through double separating wall Low frequency Hz not well predicted D nt,w +C D nt,w +C Measurement / Prediction Bedroom-Bedroom 59 / 55 db 35 / 44 db Bedroom-Bathroom 47 / 53 db 31 / 42 db Measurement - Bedroom-Bedroom Prediction - Bedroom-Bedroom Expected problems regarding comfort due to low frequencies French regulation fulfilled Measurement - Bedroom-Bathroom 0 Prediction - Bedroom-Bathroom -10 Frequency (Hz) D nt,w +C (octave bands Hz) Measurement egulation Bedroom-Bedroom 63 db 53 db Bedroom-Bathroom 55 db 50 db EUONOISE 2012 Prague June 2012 PAGE 14

15 L' nt (db) Vertical / Horizontal impact sound insulation Measurement - Horizontal - Bedroom-Bedroom Prediction - Horizontal - Bedroom-Bedroom Measurement - Vertical - Bedroom-Livingroom Prediction - Vertical - Bedroom-Livingroom Predictions in line with measurements Differences at low frequency Hz French regulation fulfilled (L nt,w 58 db) Horizontal Bedroom-Bedroom Vertical Bedroom-Bathroom L nt,w L nt,w +C I L nt,w +C I Measurement / Prediction 42 / 44 db 42 / 44 db 44 / 47 db 56 / 53 db 56 / 54 db 60 / 63 db -10 Frequency (Hz) EUONOISE 2012 Prague June 2012 PAGE 15

16 Conclusions Prediction method based on simplified SEA and adapted to lightweight constructions has been tested with simplifications to limit measurement of input parameters Comparisons between measurement and prediction down to low frequencies reasonable at this stage Considered building fulfilling French regulation but expected comfort problems due to low frequency performances On going work concerning laboratory measurements as well as in-situ measurements More comparisons necessary to validate prediction model elevance of prediction model at low frequencies, even below 50 Hz EUONOISE 2012 Prague June 2012 PAGE 16