A recent feedback of wooden multi-storey buildings in France Thomas TOULEMONDE 1 ACOUSTB 4 rue Dolorès Ibarruri – TSA 30001 – 93188 Montreuil cedex – France Bertrand DE BASTIANI 2 ACOUSTB 24 rue Joseph Fourier – 38400 Saint-Martin-d’Hères – France

ABSTRACT Wooden structures are becoming quite common in the French architectural landscape, and ACOUSTB has made it a specialty regarding acoustics. Its experience ranges from office buildings to residential buildings, and from small to high-rise constructions. Airborne and impact noise insulation are of first importance. Structural aspects, construction methods and architecture are also discussed. Constructive solutions concern both CLT and traditional wood joists flooring systems, alternatives to floating screeds in office buildings and visible wooden structures in dwellings. Buildings are described, then selected constructive solutions are presented and debated with respect to targeted acoustic performance. Finally, in-situ measurement results are compared with the theory that underlays them, as well with the targeted performance.

1. INTRODUCTION

The French wood community has created an organization, ADIVBOIS, whose objective is to remove the technical obstacles to the construction of high-rise buildings with a wooden structure. ACOUSTB has been leading the acoustic workshop of this organization since its inception, with the collaboration of various professionals in the sector: engineers, researchers and manufacturers. In 2022, theoretical models are getting mature, while their applications are already experimented in the construction industry. This article presents three examples and assesses their actual performance: a small housing building, already in use, a brand new headquarters for a French institution, and a high-rise dwelling building, still under construction.

1 thomas.toulemonde@egis.fr 2 bertrand.debastiani@egis.fr

2. L’ADELINE – DWELLINGS WITH A TRADITIONAL WOODEN STRUCTURE

2.1. Architectural and structural principles

‘yaaa Teena TE . / —_ — LIA

Figure 1: L’Adeline, 14 housing units in Saint-Martin-d’Hères, R2K architects L’Adeline is a social housing program, built near Grenoble (Auvergne – Rhône-Alpes region). The 3- storey building is made of traditional wood frame, with glued laminated timber (‘glulam’) beams and posts, supporting floorings made of 22 mm OSB on joists.

Glulam post (apparent)

22 mm OSB floor

60 mm concrete screed

Timber joist 200 mm mineral wool

20 mm mineral wool

45 mm mineral wool 18 mm plaster board

Glulam beam

Glulam post

Internal partition in dwelling

Figure 2: Wood frame and horizontal partition between dwellings

2.2. Theoretical performances During the design stage, begun in 2016, acoustical standards were set to meet the French housing regulations. In particular: • Airborne insulation between housings (vertical and horizontal): D nT,w + C ≥ 53 dB • Impact noise in housings: L’ nT,w ≤ 58 dB, including hard floor finishes.

Laboratory measurements [1] on a similar flooring complex let us predict the following performances:

• Airborne sound reduction index: R w + C = 60 dB

• Impact noise on concrete screed: L n,w = 50 dB. The design team paid particular attention to lateral transmission and allowed a few wood posts to remain visible inside the dwellings.

2.3. Measured performances and perceived quality

Figure 3: Acoustical measurements after completion (2021) In living rooms and bedrooms, a PVC floor was laid on the concrete screeds at the end of the construction stage. ACOUSTB measured impact noise after completion. L’ nT,w performance turned out to be quite homogenous in the 4 vertical tested configurations, with a predicted prevalence of low-frequency noise, due to lightweight floor construction. Therefore, addition of a PVC floor finish with an efficient impact noise reduction at medium and high frequencies only makes a small difference compared with laboratory measurements (L n,w = 50 dB) tested without floor finish. The performance in the low-frequency range is considered satisfying, regarding present French standard: L’ nT,w + CI 50-2500 ≤ 55 dB.

Table 1: L’Adeline, impact noise measurements results between superimposed housings

Test reference

Emission

Reception

L’ nT,w

CI 50-2500

L’ nT,w + CI 50-2500

room

room

[dB]

[dB]

[dB]

BCV1 Bedroom, 2 nd

Bedroom, 1 st

48 5 53

floor

floor

BCV2 Bedroom, 2 nd

Bedroom, 1 st

48 5 53

floor

floor

BCV3 Living room,

Living room,

48 3 51

2 nd floor

1 st floor

BCV4 Bedroom, 1 st

Bedroom, ground floor

45 1 46

floor

All tested configurations complied with French regulation regarding airborne sound insulation (D nT,w + C ≥ 53 dB). Measurements between superimposed dwellings showed that apparent posts caused a clear transmission in the 500-1000 Hz range:

75

70

65

60

AIV5 - Apparent post - DnTw+C = 53dB

55

AIV3 - Apparent post - DnTw+C = 54dB

50

AIV2 - Masked post - DnTw+C = 60dB

45

AIV1 - Masked post - DnTw+C = 61dB

40

35

30

125 250 500 1 k 2 k 4k

Figure 4: Vertical airborne sound insulation - Apparent/masked wood posts

This phenomenon has been later confirmed by ADIVBOIS measurements on a full-scale mock-up [2] and has driven further considerations, one of them being discussed in § 4.1 hereafter. 3. ONF HEADQUARTERS – LOW CARBON OFFICES

3.1. Architectural and structural principles

Figure 5: ONF headquarters, Maisons-Alfort, VLAU & WOA architects Brand new ONF headquarters address low-carbon highest standards, with a wooden structure braced by an infrastructure and staircases made of concrete including blast furnace slag. The structure composed of lattice beams allows a great freedom of design and the creation of vast spaces and large corbels. In a strong commitment with environmental concerns, 160 mm CLT slabs have been only equipped with reused raised-floor panels on steel pedestals. No screed neither plaster ceiling were prescribed. A carpet finish is compulsory to mitigate horizontal transmission of impact noise and to limit sonority under footstep. In addition, vertical transmission in ensured by a resilient pad inserted below pedestals.

Figure 6: Raised-floor system on CLT slab

3.2. Theoretical performances ACOUSTB has been supervising laboratory measurements of a CLT floor equipped with a standard raised-floor system [3]. Impact noise standardized index is equal to L n,w = 65 dB. Airborne noise reduction index was measured equal to R w + C = 45 dB. Both tests were carried out on a 200 mm CLT slab and with a basic carpet finish. Applied to office activities, impact noise performance needs improvement to ensure L’ nT,w ≤ 58 dB, as required by NF S31-080 highest standard [4]. The gap between laboratory results and on-site expected performance is filled by using a 6.5 mm polyurethane foam pad below raised-floor pedestals.

3.3. Measured performances and perceived quality Impact noise was measured while normalized tapping machine was successively positioned on each layer of the floor complex: • L’ nT,w (CI 50-2500 ) = 76 (-4) dB on 160 mm CLT slab, vertical transmission

• L’ nT,w (CI 50-2500 ) = 60 (2) dB on raised floor with resilient pads, vertical transmission • L’ nT,w (CI 50-2500 ) = 52 (5) dB on finished floor, including glued carpet, vertical transmission

• L’ nT,w (CI 50-2500 ) = 50 (5) dB on finished floor, horizontal transmission

L' nT [dB]

80

70

60

50

CLT floor

Raised floor

40

Carpet

30

Carpet (horizontal)

20

10

0

50 80 125 200 315 500 800 1250 2000 3150 5000

f [Hz]

Figure 7: Normalized impact noise for each layer of the floor complex

4. LE BERLIER – HIGH RISE DWELLINGS WITH CLT FLOORS

4.1. Architectural and structural principles

INTERIEUR

Figure 8: Le Berlier, 77 luxury flats in Paris, Moreau Kusunoki architects Le Berlier, to be fully constructed by end of 2022, is an upper-standing housing program, with a low- carbon environmental profile and high quality architecture and finishes. Acoustic quality must meet the highest standards, equal or better than these that may be found in a French traditional concrete structure. An additional challenge is the building height, with the highest apartments based on the 15 th floor. The building structure is composed as follows:

• Concrete structure for underground, ground floor, 1 st floor and staircases

• 2 nd floor to 15 th floor made of CLT slabs on steel beams, with glued laminated timber posts and CLT load-bearing walls. Some wood posts will remain visible inside the dwellings. • Dwelling façades are made of traditional wood frames with large windows.

Concrete screed on thin resilient underlay CLT slab 18 mm plaster board on acoustic springs + mineral wool

Steel beam

70 kg/m³ mineral wool filling Fire protection

Façade complex

Figure 9: CLT floor composition

High-class housing demands both architectural quality – with relatively thin and visible wood structure – and acoustic comfort above regulation standards. We have shown on § 2.3 and [8] that acoustic radiation of apparent wooden posts occurs at mid-frequencies, typically around 500 Hz and may limit sound insulation to D nT,w + C = 53 dB. An efficient filtering of vibration energy above 100 Hz between each floor level, associated with a floor composition as shown in Figure 9, would theoretically improve this performance of a fair 5 dB. However, vibration filtering at every floor level of a high-rise building is challenging, as cumulative static deflection of resilient bearings must comply with strict structural criteria. Execution studies, carried out with Getzner engineers and Wewood construction team, turned out to define

12 mm thick polyurethane foam bearing pads. This device is inserted between the steel floor structure and wood posts, and is hidden under the concrete screed.

Figure 10: Getzner High Resilient Bearing mat for wooden post desolidarization

4.2. Theoretical performances ACOUSTB has been carrying out systematic on-site measurements on CLT floors equipped with a concrete floating screed, either on a 15 mm mineral wool (‘thick’) underlay, or on a 3 mm viscoelastic (‘thin’) underlay. Figure 6 shows impact noise measurements with a normalized tapping machine, for both configurations and a ceiling made of 18 mm plaster board on acoustic springs + mineral wool.

L' nT

90

Thin underlay Thick underlay

80

70

60

50

40

30

20

f (Hz)

Figure 11: L’ nT normalized impact noise comparison for 2 types of resilient underlays

63 125 250 500 1000 2000 4000

A concrete screed on a thick underlay turns out to have a resonant frequency in the range of 63 Hz octave band. Therefore, resulting impact noise including low-frequency correction factor are:

 L’ nT,w + CI 50-2500 = 58 dB with a ‘thick’ underlay  L’ nT,w + CI 50-2500 = 55 dB with a ‘thin’ underlay The latter value fits precisely with French recommendation for impact noise on light flooring structures. A thin underlay has thus been adopted for the present project. Sound reduction index of the complex has already proven to overpass French regulations regarding airborne noise between superimposed housings.

4.3. Vibration issues Vibrations from walking need to be considered with CLT floors as described here above. This issue is discussed in XP CEN/TS 19103 standard [5] defining EUROCODE 5 common rules for design of timber structures. Design parameters are: • Natural frequency of vibrations of the structure (slabs, floor and screed included), f e1 , with a critical criterion f e1 ≥ 8 Hz in dwellings (4 th harmonic of a typical 2 Hz walking frequency). Supporting conditions and continuous beams must be taken into account. Operating overloads may be neglected. • Instantaneous elastic bending stiffness of the composite structure w 1kN . • Vertical vibration velocity response to a unit pulse, υ 1kN . These recommendations must be considered altogether, as it turns out that some users may raise complaints in situations where only natural frequency criterion is met [6], [7], [8]. The selected floor composition, combined with the wooden structure calculated by our partner CET Ingénierie, turned out to comply with these vibration comfort recommendations.

4.3. Measurement methods Early measurements are planned in May 2022 on a prototype apartment. Results are presented during INTERNOISE 2022 congress session. Impact noise is measured both with a normalized tapping machine and with a ‘Japanese ball’ complying with JIS 1418-2 standard [9]. Lateral transmission of apparent posts are assessed by on-site measurements according to NF EN ISO 15186-2 standard [10]. The reproducibility of this intensity method is estimated to be equal to laboratory methods ISO 140-10 and ISO 140-4. Equation 1 defines the sound reduction index of a lateral element j , when a building element separates the transmitting room from the receiving room:

𝑆 𝑀𝑗

𝑆

𝑆 0 )] −[𝐿̅ 𝐼𝑛𝑗 + 10 log (

𝑆 0 )] , (1)

𝑅 𝐼𝐹𝑗 = [𝐿 𝑝1 −6 + 10 log (

where the first part of the equation is the sound power incident on the tested separating element in the emitting room and the second part is the sound power radiated by the lateral element j in the receiving room and where: • L p1 is the average sound pressure level in the emitting room, • S is the area of separating element

• 𝐿 ̅ 𝐼𝑛𝑗 is the average normal sound intensity level on the measurement surface of the lateral element j in the reception room • S M j is the total measurement surface of the lateral element j in the reception room • S 0 = 1 m². The receiving room shall be as absorbing as possible. Thick mineral wool mats are usually disposed on ceiling, walls and floor of the room.

Figure 12: Intensimetry measurement in practice

On the other hand, emitting room shall be as diffusive and reverberant as possible. This may be achieved by any means on the construction site. 4. CONCLUSIONS

In the long and complex process of construction, acoustic consultants need to make early choices and engage their responsibility upon a future result. Moreover, architecture trends and rapid development of wood construction techniques are raising new questions that need to be carefully studied, for instance, visible woodworks inside dwellings. Deadlines of the design process don’t always let time to collect feedback from the research field. ADIVBOIS, as a unique collaboration group, helps knowledge circulate freely between research institutes, engineers and the wood industry. Test results, calculation methods and on-site measurements are part of a same process of continuous improvement. Such progress benefits on acoustic comfort for users, architectural innovation and reduction of environmental footprint of buildings. 5. ACKNOWLEDGEMENTS

We gratefully acknowledge all active participants of ADIVBOIS acoustics workshops, the design and construction teams of each three projects described in this article, and ACOUSTB team members for their participation in studies and on-site measurements. 6. REFERENCES

1. Guigou Carter C., Balanant N. & Villenave M., ACOUBOIS étapes 2 & 3, rapport V2.0, Mesures

acoustiques en laboratoire (2014) 2. Guigou Carter C., Balanant N. & Kouyoumji J., Rapport Final – Maquette Acoustique ADIVBOIS

(2022) 3. CSTB, Acoustic test report n° AC18-26077995 regarding floors (2019) 4. NF S31-080, Acoustics — Offices and associated areas — Acoustic performance levels and criteria

by type of area 5. XP CEN/TS 19103, EUROCODE 5: design of Timber Structures - Structural design of timber-

concrete composite structures - Common rules and rules for building 6. Hamm P., Richter A., Winter S., Floor vibrations, World Conference on Timber Engineering (2010) 7. CODIFAB – Vibrations : approche des planchers basses fréquences – notions (2018) 8. ISO 10137, Bases for design of structures – serviceability of building and walkways against vibrations 9. JIS 1418-2, Acoustics — Measurement of floor impact sound insulation of buildings — Part 2:

Method using standard heavy impact sources 10. NF EN ISO 15186-2, Acoustics - Measurement of sound insulation in buildings and of building

elements using sound intensity - Part 2: field measurements.