Duration dependence of night-time noise effect for passively cooled residential bedrooms Anthony Chilton 1 Max Fordham LLP 42-43 Gloucester Crescent, London NW17PE Peter Leonard 2 Max Fordham LLP 42-43 Gloucester Crescent, London NW17PE

ABSTRACT The Acoustics, Ventilation, Overheating: Residential Design Guide[1] (“The AVO Guide”) was launched by the Association of Noise Consultants and Institute of Acoustics in January 2020. It provides a methodology for assessing the adverse effect due to break-in of external transport noise through façade openings provided for the purpose of mitigating overheating. Figure 3-2 of the AVO Guide indicates qualitatively how the SOAEL (Significant Observed Adverse Effect Level) is expected to vary with the duration for which the ‘overheating condition’ occurs. However, no quantitative guidance regarding the dependence on duration is given. This study describes how the methodology used in the WHO Night Noise Guidance for Europe 2009[2] to evaluate outside-to-inside level difference can be extended to provide a quantitative definition of the SOAEL for transport noise affecting passively cooled residential bedrooms at night. A quantitative version of the AVO Guide Figure 3-2 is derived and presented. Comparison is made between the SOAEL values presented here and the duration/frequency independent values suggested in the recently published Approved Document Part O[3]. Commentary is provided regarding the urgent need for better availability of relevant information on the frequency of occurrence of window openings for residential buildings in use, as opposed to data generated by the dynamic thermal models produced to show compliance with CIBSE TM59[15] and/or Approved Document Part O.

1. AVO GUIDE - QUALITATIVE APPROACH

1.1. Level 2 Assessment Based on indoor ambient noise level as shown in Figure 3-3 of the AVO Guide[1] (reproduced here as Figure 1). Paragraph 3-16 of the AVO Guide states that the upper category in Table 3-3 is defined with reference to the WHO Night Noise guidelines[2] (‘NNG’). Although not explicitly stated, the lower category in Table 3-3 is defined with reference to the guideline values in Table 4 of BS8233:2014 (‘BS8233’)[4].

1 a.chilton@maxfordham.com 2 p.leonard@maxfordham.com

It is important to note that both the NNG and BS8233 values are yearly average values. As such, it is implicit that the Level 2 assessment is based on yearly average noise levels. Table 3-3 makes qualitative statements (as highlighted in Figure 1) regarding how the duration for which the noise level is experienced will influence the adverse effect of the noise.

1.1. AVO Diagram Figure 3-2 (“the AVO diagram”) illustrates qualitatively how adverse effect is expected to vary with duration. Figure 3-2 is reproduced here as Figure 2. Figure 3-2 of the AVO Guide does not quantify the scale of either the x or y axes. Figures B-2 and B-3 of the AVO Guide show versions of the AVO Diagram that quantify the Internal Ambient Noise Level axis (y-axis) but the Duration axis (x-axis) remains qualitative. The reason that quantitative guidance was not provided regarding duration was that the AVO Guide authors felt that there was not a sufficiently robust basis for a qualitative approach.

Figure 1: Table 3-3 from AVO Guide “Guidance for Level 2 assessment of noise from transport noise sources relating to the overheating condition”. See AVO Guide[1] for the referenced notes.

Figure 2: ‘AVO Diagram’ – Figure 3-2 from AVO Guide[1] “Qualitative guidance on combined

effect of internal ambient noise level and duration for the overheating situation”.

2. WHO NIGHT NOISE GUIDELINES FOR EUROPE 2009[2]

2.1. Conversion between Inside and Outside Noise Levels Section 1.3.4.4 of the NNG describes the correction that is used to convert between inside and outside noise levels. The relevant relationship (Equation 3 from the NNG) is reproduced as Equation 1 here.

L night = L night,inside + Y dB (1)

where:

• L night is the one-year L Aeq over the 8-hour night-time period, outside at the most exposed

façade • L night,inside is the one-year L Aeq over the 8-hour night-time period, inside • Y is the year average insulation of the (bedroom) façade. The NNG uses a value of 21dB for Y.

2.2. Evaluation of the Year-Average Insulation The value of 21dB for Y is taken from a field study[5] of sleep disturbance and aircraft noise undertaken between November 1999 and April 2001 in the areas around Schiphol Airport in the Netherlands. The study had 418 adult participants at 15 locations within 20km of Schiphol airport exposed to varying degrees of aircraft noise exposure. Each participant was involved for an 11-day period from a Monday morning until Friday morning the following week. Reference should be made to the original reports for a full description of the study. For the purposes of this discussion, the relevant pieces of information that were recorded as part of the study are as follows:

• Indoor noise levels in the participants bedroom • Outdoor noise levels at the location

Intemat ambient level frm anspor. Rely Duration for which the “overheating condition occurs Most ofthe tine

• For both the indoor and outdoor noise measurements, aircraft noise events have been

identified by reference to the flight track monitoring system of the Civil Aviation Enforcement Agency • The values of L max_o (outside) and L max_i (inside) have been identified for each aircraft noise

event. L max_o and L max_i being the maximum value of the L Aeq,1s value during the event. • A total of 63,242 aircraft noise events were monitored during the study • The outside-to-inside level difference (L max_o - L max_i ) has been evaluated for each aircraft

noise event. • Both the cumulative distribution and mean values of the outside-to-inside level difference

have been reported. Values are reported separately for single-glazed and double-glazed windows. • Participants were required to complete a questionnaire each morning. This includes

questions about the position of the bedroom window during the previous night. Five options are given for the window position ranging between fully closed and fully open. • The distribution of the reported window positions is reported. Again, values are reported

separately for single-glazed and double-glazed windows. Table 1 shows the cumulative distribution of L max_o – L max_i as reported in Table B6 in reference 5. Table 2 shows the distribution of reported night-time window positions from the questionnaire responses. It is made clear in Section 1.3.5 of NNG that the origin of the 21dB value for Y in Equation 1 is the average value for single glazed bedrooms as reported in Table 1, with an acknowledgement that there being only a “slight difference between single and double glazed windows”. Table 1: Cumulative distrib uti on of measured outside-to-inside level difference. From referen ce 5.

Frequency Single glazing (42% of bedrooms)

Double Glazing (58% of bedrooms) 10% 13 12 20% 16 16 30% 18 18 40% 20 20 50% 22 23 60% 23 25 70% 25 27 80% 27 28 90% 30 31 Average 21.3 22.2 Table 2: Frequency of night-time window positions as reported in the questionnaire responses. From reference 5.

Window Position Frequency Single glazing (42% of bedrooms)

Double Glazing (58% of bedrooms) Fully closed 47.4% 40.6% Small opening 21.7% 27.9% Open at hand’s breadt h 23.2% 22.6% Half opened 5.6% 4.3% Fully opened 2.1% 4.6%

3. USE OF WHO NNG CONVERSION TO DERIVE AVO SOAEL

The LOAEL line shown in Figure 2 is horizontal and can therefore be unambiguously defined by the internal ambient noise level alone. With reference to Figure 1, it is suggested that the lower category noise levels (i.e. those from Table 4 of BS8233) can be taken to define the LOAEL for the overheating situation. Indeed, this is the approach taken in Figures B-2 and B-3 of the AVO Guide. The SOAEL line shown in Figure 2 is, however, not horizontal indicating that there is a dependence on both internal ambient noise level and duration. The argument presented in this paper is that the cumulative distribution shown in Table 1 can be used as a basis for quantitatively defining the SOAEL line in Figure 2.

3.1. Applicability of the values shown in Figure 3 As discussed in paragraph 3.16 of the AVO Guide, the upper category in Table 3-3 (Figure 1 here) for the night-time period is defined with reference to the NNG interim target of 55dB L night . L night is the one-year L Aeq over the 8-hour night-time period, outside at the most exposed façade. Table 3-3 of the AVO Guide (Figure 1 here) is expressed in terms of internal ambient noise level. It is therefore necessary to convert the NNG L night value to an internal level to define the upper category. The AVO Guide makes this conversion by applying an outside-to-inside level difference of 13dB to represent the situation of a partially open window as discussed in paragraph 3.24. The reason for using the situation of a partially open window was that the AVO Guide assessment relates to the overheating situation, where windows are typically expected to be open. A significant limitation of the AVO Guide is that it does not then provide quantitative guidance as to the acceptable duration of upper category noise levels other than to imply qualitatively that they should occur “rarely” and be avoided for “most of the time”. Instead of converting the NNG L night value to an internal level by applying an outside-to-inside level difference of 13dB, it is suggested that this conversion is done using Equation 1. This is consistent with the conversion methodology put forward in the NNG itself. Rather than using a yearly average value of 21dB for Y in Equation 1, it is instead suggested that the value of Y is variable across the year and defined by the cumulative distribution shown in Table 1. The values for double glazing are used on the basis that these are more relevant to the design of new housing. The night-time internal ambient noise level, L Aeq,8hr , is calculated by subtracting the outside-to-inside level difference from the outdoor L night value of 55dB. Table 3: Weighted average level difference and resultant internal level for 55dB L night outside

Frequency Outside-to-inside level

L night (outside)

L Aeq,8hr (inside)

difference [dB]

[dB]

[dB] 10% 12 55 43 20% 16 55 39 30% 18 55 37 40% 20 55 35 50% 23 55 32 60% 25 55 30 70% 27 55 28 80% 28 55 27 90% 31 55 24 Average 22.2 55 32.8

Figure 3 shows a graphical plot of the values from Table 3 in a format that is analogous to an AVO diagram, assuming that ‘Frequency of Occurrence’ axis is equivalent to ‘Duration for which the overheating condition occurs’. Figure 3 also shows a LOAEL value of 30dB L Aeq,8hr . The x-axis of Figure 3 is the frequency of occurrence in a data set of 63,242 night-time aircraft noise vents measured over a 17-month period from November 1999 to April 2001. The presentation of the data in Figure 3 effectively assumes that the frequency of occurrence across the dataset is equivalent to frequency of occurrence across a calendar year. This same assumption has been made in the NNG by taking 21dB as the yearly average outside-to-inside level difference. There is a further assumption implicit in this approach, that the window opening behavior exhibited in the field study[5] still holds at external levels of 55dB L night (the NNG interim target). This is felt to be reasonable given that the sites in the study experienced external aircraft noise levels of up to 52dB L night , a difference of only 3dB.

50dB

Night-time internal ambient noise level

SOAEL

from transport noise sources, L Aeq,8hr

45dB

LOAEL

40dB

35dB

30dB

25dB

20dB

0%

100%

10%

20%

40%

50%

60%

70%

80%

90%

30%

Frequency of Occurence

Figure 3: The distribution of night-time indoor L Aeq,8hr from Table 3, plotted to represent the

SOAEL line in the AVO diagram format. The LOAEL is also shown at a value of 30dB.

3.2. Use of the information from questionnaire responses Table 2 shows the frequency distribution of night-time window position as reported by participants in their questionnaire responses the following morning. This is self-reported information that will involve some subjective judgement on behalf of participants and no-doubt some element of human error. However, it is still valuable to consider as it provides information that can be used to supplement the values shown in Table 3 and Figure 3. The first notable aspect is that windows are reported as being fully closed on at least 40% of the nights during the study. The AVO Guide Level 2 assessment is not concerned with the situation where windows are closed. It is therefore reasonable to curtail the x-axis in Figure 6 to the range 0% to 60%. The second notable aspect is that windows are reported “fully-opened” on only 4.6% of nights in the case of double glazing. This frequency is below the lowest data point given in Table 1 (Table B6 in reference 5) and might therefore be used to extend the definition of the SOAEL line. To do this, we need to determine an outside-to-inside level difference corresponding to the “fully-opened” window situation.

For England and Wales, Approved Document F (ADF)[6] of the Building Regulations requires provision for “purge” ventilation that can be provided by a minimum total area of openings equivalent to 1/20 th of the floor area of the room. Note that a larger area is required in the case of a “a hinged or pivot window that opens less than 30°. It can therefore be assumed that fully-opened windows imply an opening of at least 1/20 th of the floor area of the room . This information can be used to estimate the outside-to-inside level difference for the fully-opened window situation. Note that the purge ventilation “spuiventilatie” requirement[7] in the Netherlands is 6 litres/s/m² floor area. For a room height of 2.3m, this equates to 9.4 ach (air-changes per hour), which is significantly higher than the ADF value of 4ach and may imply larger window openings depending on the situation. Similarly, the ‘Simplified Method’ set out in Approved Document Part O (ADO)[3] suggests that a bedroom located in a ‘high’ overheating risk location should have openings equivalent to 13% of the floor area of the room. Again, this is a significantly larger opening than ADF “purge”. As such, the calculation presented below, which is based on ADF window openings, may slightly over-estimate the outside-to-inside level difference. Equation D.1 from BS EN ISO 12354-3:2017[8], reproduced as Equation 2 here can be used to calculate the element normalized level difference for the open window.

𝐷 𝑛,𝑒 = −10𝑙𝑜𝑔 10 ( 𝑆 𝑜𝑝𝑒𝑛

) (2)

𝐴 0

where :

• S open is the area of the opening, in m² • A 0 is the reference equivalent sound absorption area, in m². For dwellings, A 0 =10m² Equation G.1 from BS 8233:2014[4], can be simplified to the form shown as Equation 3 by assuming that there is no transmission of outdoor into the room sound except that via the window opening.

(− 𝐷 𝑛,𝑒

𝐿 𝑒𝑞,2 = 𝐿 𝑒𝑞,𝑓𝑓 + 10𝑙𝑜𝑔 10 [ 𝐴 0

10 ) ] + 10𝑙𝑜𝑔 10 ( 𝑆

𝐴 ) + 3 (3)

𝑆 10

where:

• L eq,2 is indoor sound level in the room • L eq,ff is the free-field sound level outside of the window • A 0 is the reference equivalent sound absorption area, in m². For dwellings, A 0 =10m² • S is the total area, in m², of elements through which sound enters the room. Note that the

value of this term is arbitrary here as it cancels out in subsequent steps of this derivation. • A is the equivalent absorption area of the room From the discussion given above regarding purge ventilation provisions, we can also use:

𝑆 𝑜𝑝𝑒𝑛 = 𝑓. 𝑆 𝑓𝑙𝑜𝑜𝑟 (4)

where:

• S floor is the floor area of the room, in m² • f is the area of openings expressed as a fraction of the room floor area. In the case of ADF,

f=0.05.

Sabine’s formula (e.g. Section 3.17 of BS EN ISO 16283-3:2016[9]) can be used to obtain the expression shown in Equation 5 for the equivalent absorption area of the room.

𝑇 = 0.16𝑆 𝑓𝑙𝑜𝑜𝑟 ℎ

𝐴= 0.16𝑉

𝑇 (5)

where:

• T is the reverberation time in the room in seconds. • h is the height of the room in metres The reverberation time can be approximated by the reference reverberation time i.e. assume T=T 0 =0.5s. Section 3.17 of BS EN ISO 16283-3:2016[9] states that “in dwellings with furniture the reverberation time has been found to be reasonably independent of volume and frequency and to be approximately equal to 0.5s”. Equations 2, 3, 4, and 5 can then be combined to evaluate the outside-to-inside level difference as given by Equation 6.

𝐿 𝑒𝑞,𝑓𝑓 −𝐿 𝑒𝑞,2 = 10𝑙𝑜𝑔 10 (ℎ) − 10𝑙𝑜𝑔 10 (𝑓) −8 (6)

50dB

Night-time internal ambient noise level

SOAEL

from transport noise sources, L Aeq,8hr

45dB

LOAEL

40dB

35dB

30dB

25dB

20dB

0%

10%

30%

40%

50%

60%

20%

Frequency of Occurence

Figure 4: The data distribution from Figure 6 replotted to include the calculated fully-opened

window data point but to exclude the points corresponding to fully-closed windows. The “Technical housing standards – nationally prescribed space standard” 2015[10] requires that the minimum floor to ceiling height is 2.3m for at least 75% of the Gross Internal Area. Adopting h=2.3m and f=0.05 from ADF implies an outside-to-inside level difference of 8.6dB. The resultant internal level for 55dB L night outside is therefore 46.4dB L Aeq,8hr . It is worth noting that the calculated outside-to-inside level difference shows agreement with the measured value of 10.0±2.9dB reported in a recent Swiss field study[11]. As noted above, the ventilation openings

required in the Netherlands may be larger than those in the UK, which would imply a slightly higher internal level for the fully-opened window situation. Figure 4 shows Figure 3 replotted to include the fully-opened window data point and to exclude data points where windows are fully-closed.

3.3. Correlation between questionnaire responses and measured outside to inside level difference In the previous section, outside-to-inside level difference was estimated for the “fully opened” window situation, as reported in the questionnaire responses[5]. In order to further investigate the validity of this approach, it is worthwhile to look at the rest of the data from the questionnaire responses and evaluate the correlation between calculated and measured values of outside-to-inside level difference. Each of the questionnaire responses, see Table 2, has been related to a corresponding outside-to- inside level difference as shown in Table 4. With the exception of the “small opening” situation, this is done by making reference to a 2018 Swiss field study[11] of window sound insulation. For the case of a “half-opened” window, the value has been inferred from the “fully opened” window value using Equation 2 above. The “open at hand’s breadth” situation was equated to the “tilted” window in the field study[11] on the basis that “the opening angle of tilted windows was typically about 5–10 degrees”. For a typical[12] 1m high window this implies an opening distance of around 9cm-18cm, which could reasonably be described as “hand’s breadth”. The outside-to-inside level difference for the “small opening” situation has been related to the laboratory measurements undertaken as part of Napier University’s 2007 laboratory study[12]. The “small opening” situation has been taken to equate to the “ajar” position. This is described[12] as “if the mechanism was available, the window was opened a small amount to a secondary keeper set further out from the primary keeper, which allowed the window to be secured whilst still providing ventilation”. This description suggests the lower end of what might be considered a “small opening” but the authors of this paper are not aware of any more representative measurements. The BS EN ISO 140-5[13] comparable measurements are used, which have a source located 5m from the façade (position L6) and the source microphone located 2m external to the façade (position S1). From Table 4-3 of the Napier Study[12], the six windows for which the “ajar” position has been measured show an averaged level difference of D w =21.7±1.1dB. According to BS EN ISO 12354-3[8], the external level measured at 2m from the façade (position S1) is 3dB higher than the free-field external level (i.e. the level at the position of the façade without the façade being present). This suggests that the outside-to-inside level difference should be 3dB lower, i.e. 18.7±1.1dN. However, the receiver room used in the Napier Study[12] was a bare room with a reverberation time of 1.4s. Normalising to a reverberation time of T 0 =0.5s (Section 3.17 of BS EN ISO 16283- 3:2016[9]) indicates a correction of +4.5dB, resulting in a value of 23.2±1.1dB. The outside-to-inside level difference reported in Table 4 are used to calculate the night-time internal ambient noise level, L Aeq,8hr , by subtracting the outside-to-inside level difference from the outdoor L night value of 55dB. The resulting internal ambient noise levels are shown in Figure 5. Note that the bars show a vertical range accounting for the standard deviation. The horizontal width of the bars relates to the frequency of occurrence, also tabulated in Table 4. It can be seen that there is generally a good agreement shown between the levels based on the questionnaire responses and the values that levels that were measured on site in the study[5]. As expected, the “ajar” condition[12] appears to correspond to very small openings, at the lower end of the range of what is termed a “small opening” in the questionnaire responses[5]. Typically, windows latched open position is only around 1 or 2cm open, whereas the “small opening” category captures any opening that is too small to be considered “hands breadth”.

Table 4: Outside-to-inside level difference values corresponding to the five window positions reported in the questionnaire responses.

Window

Frequency

Cumulative

Outside-to- inside Level

Reference Note

Position

Frequency

Difference

[dB]

Fully closed

40.6% 100% 27.8 ± 4.4 Table 2 of Ref 11

‘closed’ window

Small opening

27.9% 59.4% 23.2 ± 1.1 Table 4-3 of Ref 12 ‘ajar’ position,

Source position L6, microphone position

(windows A-1, A-2, A-3, E, F, G)

S1

Open at

22.6% 31.5% 15.8 ± 2.7 Table 2 of Ref 11

Opening angle typically 5-10°. For

hand’s breadth

‘tilted’ window

a 1m high window,

implies opening distance of 9-18cm. Half opened

4.3% 8.9% 13.0 ± 2.9 Table 2 of Ref 11

3dB added to open window value as per

‘open’ window with correction

Equation 2 Fully opened

4.6% 4.6% 10.0 ± 2.9 Table 2 of Ref 11

‘open’ window

50dB

Measured (L max_o – L max_i )

Night-time internal ambient noise level

from transport noise sources, L Aeq,8hr

A

A = Fully opened B = Half opened C = Hand’s breadth (“tilted”) D = Small opening (“ajar”) E = Fully closed

45dB

B

C

40dB

35dB

D

E

30dB

25dB

20dB

0%

100%

10%

20%

30%

40%

50%

60%

80%

90%

70%

Frequency of Occurence

Figure 5: The distribution of night-time indoor L Aeq,8hr from Table 3 with corresponding predicted

values from Table 4 corresponding to the questionnaire responses overlaid for comparison.

4. COMMENTARY

4.1. Applicability to the AVO Guide Figure 4 of this paper provides a quantitative version of the Figure 3-2 ‘AVO Diagram’[1] for the case of night-time noise in bedrooms. Provided that information can be obtained (refer to Section 4.2 below) regarding the expected in-use frequency of occurrence of window openings for the bedroom in question, Figure 4 can be used to evaluate whether the SOAEL value has been exceeded or not. Because Figure 4 relates to internal ambient noise level, it can also be used in the case of attenuated façade openings that provide a greater outside-to-inside level difference than opening windows providing equivalent ventilation. Reference 14 provides a methodology for comparing different attenuated façade ventilators using a quantitative rating based on a combination of both their aerodynamic and acoustic performance.

4.2. Obtaining Data on Frequency of Occurrence of Window Openings It is extremely important to make the point that the frequency of occurrence of window openings relates to the actual use of the window in the occupied building and not to the values predicted from a dynamic thermal model with template occupancy, heat-loads and window opening profiles. That is to say that dynamic thermal modelling undertaken in accordance with CIBSE TM59[15] and/or Approved Document Part O[3] will not provide appropriate information about frequency of window (or other façade) openings in actual buildings. There have been a number of studies [e.g. References 16, 17, 18] conducted regarding residential window opening behaviour but additional UK studies would be beneficial and further work is urgently required to inform how this information may be used in the design of new buildings.

4.3. Approved Document Part O - Overheating Approved Document Part O (‘ADO’)[3] comes into effect in England on 15 th June 2022. Paragraph 3.3 of this document includes the following statement: “Windows are likely to be closed during sleeping hours if noise within bedrooms exceeds the following limits.

a. 40dB L Aeq,T , averaged over 8 hours (between 11pm and 7am). b. 55dB L AFmax , more than 10 times a night (between 11pm and 7am)”. This could be taken to mean imply a SOAEL value of 40dB L Aeq,8hr . Comparing this value with the distribution shown in Figure 4 indicates that, for a frequency of occurrence less than around 17% of nights (63 nights per year), the ADO SOAEL value is lower (i.e. more onerous) than that derived in this paper. Another way to make the comparison would be to say, with reference to Table 4 above, that the ADO 40dB L Aeq,8hr value implies that bedroom windows are not likely to be opened to more than “hand’s breadth” on any night. ADO does not provide any explicit reference or justification for the 40dB L Aeq,8hr value. It is also worth noting that the L AFmax value given in ADO may imply a higher outside-to-inside level difference than is implied by the L Aeq,8hr value, although no further discussion is given here regarding this issue.

5. CONCLUSIONS

A methodology is presented in which Equation 3 and the interim target of 55dB L night from NNG[2] are used together with the underlying frequency distribution[5] of measured outside-to-inside level difference to calculate the frequency of occurrence of internal ambient noise levels in bedrooms at night. This frequency distribution (presented in Figure 4) is put forward as a quantitative version of the Figure 3-2 ‘AVO Diagram’[1] for the case of night-time noise in bedrooms. Comparison is made between the SOAEL values presented here and the duration/frequency independent values suggested in the recently published Approved Document Part O[3]. Commentary is provided regarding the urgent need for better availability of relevant information on the frequency of occurrence of window openings for residential buildings in use, as opposed to data generated by the dynamic thermal models produced to show compliance with TM59[15] and/or ADO[3]. 6. REFERENCES

1. Association of Noise Consultants and Institute of Acoustics., Acoustics Ventilation and

Overheating: Residential Design Guide ., (2020). 2. Night Noise Guidelines for Europe, World Health Organisation (2009) 3. HM Government, Approved Document O – Overheating , 2021 edition, (2021) 4. British Standards Institution, British Standard 8233:2014 Guidance on sound insulation and

noise reduction for buildings , (2014) 5. Passchier-Vermeer et. al., Sleep disturbance and aircraft noise exposure – Exposure-effect

relationships TNO Report 2002.027, (2002) 6. HM Government, Approved Document F – Ventilation , 2021 edition, (2021) 7. Minister voor Wonen en Rijksdienst, Bouwbesluit 2012 - Afdeling 3.7. Spuivoorziening, (2012) 8. British Standards Institution, BS EN ISO 12354-3:2017 Building acoustics — Estimation of

acoustic performance of buildings from the performance of elements — Part 3: Airborne sound insulation against outdoor sound , (2017) 9. British Standards Institution, BS EN ISO 16283-3:2016 Acoustics — Field measurement of

sound insulation in buildings and of building elements — Part 3: Façade sound insulation , (2016) 10. Department for Communities and Local Government, Technical housing standards – nationally

prescribed space standard , (2015) 11. Locher, Wunderli et. al., Differences between Outdoor and Indoor Sound Levels for Open,

Tilted, and Closed Windows, International Journal of Environmental Research and Public Health , 15(1) , p149, (2018). 12. Department of Environment, Food & Rural Affairs, NANR116 Open-Closed Window Research

Report , (2007) 13. British Standards Institution, BS EN ISO 140-5:1998 Acoustics — Measurement of sound

insulation in buildings and of building elements — Part 5: Field measurements of airborne sound insulation of façade elements and façades , (1998) 14. Chilton, Novo, McBride, Lewis-Nunes, Johnston and Rene, Natural ventilation and acoustic

comfort, Proceedings of Acoustics 2012 , Nantes, France, (2012) 15. Chartered Institution of Building Services Engineers, TM59 Design methodology for the

assessment of overheating risk in homes , (2017) 16. Andersen, Fabi, Toftum, Corgnati, Olesen, Window opening behaviour modelled from

measurements in Danish dwellings, Building and Environment 69 p101-113, (2013) 17. Kali, Andersen, Müller, Olesen, Analysis of occupants' behavior related to the use of windows

in German households, Building and Environment 103 p54-69, (2016) 18. Morrison, Cagle, Date, A national survey of window-opening behavior in United States

Homes, Indoor Air . p1–14, (2021)