Exploring the Necessity of Thermally broken Windows in Hong Kong

ABSTRACT from The Façade, the official publication of Hong Kong Façade Association, Issue #33:

Thermally broken windows consist of thermally broken frames and insulated glass, and have superior thermal performance compared to regular windows. Their characteristics of low U-value and low shading coefficient contribute to reduce both the heating and cooling loads of buildings in Hong Kong, thereby significantly reducing carbon emissions.

Thermally broken windows are globally recognized as one of the most important technologies for achieving a safer and greener society, and have been widely used around the world for three decades. Shenzhen, China has been using thermally broken windows since the new national code changed the maximum allowable window U-value from 3.5 W/(m²·K) to 2.4 W/(m²·K). However, Hong Kong hardly uses any thermally broken windows, which is worth discussing.

This paper discusses the barriers to the adoption of thermally broken windows in Hong Kong. The first issue is the shortcoming of the current OTTV method, which was obtained based on a set of conditions that are no longer valid. The new conditions lead to a different result, which leads to the overestimation of the building energy performance. The second issue is about the current practice of LEED in Hong Kong. Engineers use the glass center U-values as the window U-values, which doesn’t follow the actual requirement of LEED. This discrepancy means that the actual energy performance of the buildings is unlikely to meet the design requirements. The third issue is that the Buildings Department (BD) does not approve thermally broken windows due to structural safety concerns, and a simple and mature solution to address BD concerns is presented. The paper also discusses its limitations and the future plan for the comparable energy simulation and experimental testing.

Keywords: thermally broken window, U-value, OTTV, building envelope

Technical review by: Y.Z. ZHOU of Warmframe Technology Corporation, Beijing, China, and Richard LI of AECOM Building Engineering (Façade), Hong Kong

1. BACKGROUND

In the 19th century, monolithic glass was used in windows. To improve the thermal performance of windows, American inventor Thomas D. Stetson patented an insulated glass unit (IGU) in 1865. However, it wasn’t until the oil shortages of the 1970s that IGUs became more widely used. At that time, glass manufacturers began to develop low-emissivity (low-E) coatings to improve energy efficiency, nearly doubling the insulating performance and reducing the U-value by 50% of a conventional IGU with uncoated clear glass. 40 years ago, a German company developed the first thermal break for windows, the plastic bar installed between the outer and inner parts of the metal window frame. Today, although modern thermally broken windows consist of different glass and frame profiles, the original concept of using a low-conductivity material as the thermal break has remained unchanged (Figure 1).

• 

Figure 1: Modern Thermally Broken Windows

One of the key parameters for differentiating windows is the U-value, which may be described as the speed of heat transfer through the outer and inner surfaces of the window. The lower the U-value, the better. For the whole window, the best U-value is 0.8 W/(m²·K) or less, which consists of well-insulated frames and high-performance IGUs; while the worst U-value is 6.8 W/(m²·K) or more, which consists of regular frames and single glass. It should be noted that the window U-value is the area-weighted average of the frame and glass, not just the glass itself, as specified in the relevant standards, i.e. ISO 15099, ASHRAE 90.1, etc.

Germany was one of the early adopters of thermally broken windows, switching from regular windows in the 1990s when the window U-value limit changed from 3.1 W/(m²·K) to 1.8 W/(m²·K). The United Kingdom completed a similar upgrade in the 2000s. The United States and China began using thermally broken windows in the 2000s, but only for buildings in cold climates back then; buildings in hot climates didn’t use thermally broken windows until 2020. In Shenzhen, almost all the buildings used regular windows in the past, but since the new code, GB55015-2021: General Code for Energy Efficiency and Renewable Energy Application in Buildings, was introduced in 2021, the maximum allowable window U-value was changed from 3.5 W/(m²·K) to 2.4 W/(m²·K), requiring thermally broken windows. Hong Kong doesn’t have a window U-value requirement. Wong (2017) showed that the most popular configuration of windows for commercial buildings in Hong Kong were non-thermally broken (regular) frames with low-e IGUs [1]. The typical window U-value in Hong Kong is estimated to be an average of 3.6 W/(m²·K) (Figure 2).

• 

Figure 2: Window U-value Requirements Since 1990

• INTRODUCTION

The appearance of thermally broken windows can be exactly the same as regular windows, seen from the inside or outside. It is the thermal breaks that make the thermal performance of thermally broken windows exceed that of regular windows. THERM software was used to quantify and visualize this performance. Developed by the Lawrence Berkeley National Laboratory (LBNL), THERM is one of the most widely used thermal simulation software programs in the world. Simulation results show that the benefits of thermally broken windows include:

Higher energy efficiency and less carbon footprint

U-Value and Shading Coefficient (SC) are the main characteristics related to the energy efficiency of windows. Table 1 shows the U-value and SHGC of a state-of-the-art thermally broken window and a regular window with the same low-e IGU.

Table 1: U-value and SHGC of Windows

Frame to window area	State-of-the-art thermally broken window	Regular window
U-value	SC	U-value	SC
15%	2.0 W/(m2·K)	0.25	3.6 W/(m2·K)	0.28
25%	2.3 W/(m2·K)	0.23	5.2 W/(m2·K)	0.27
Note: Same glass is used for both the thermally broken and regular windows. Glass U-value = 1.6 W/(m2·K), Glass SC = 0.27

Although the ratio of the frame to the window area is only 15% to 25% in regular windows, the influence of the frame on the energy efficiency can be as high as 70%. This is because the frame U-value is about 16 W/(m²·K), which is about 10 times the glass U-value. As a comparison, the frame U-value of the state-of-the-art thermally broken window with Warmframe thermal break blanket and warm-edge glass spacer is only 4 W/(m²·K), which is 75% less than the regular frame (Figure 3). Thanks to the high performance of the thermally broken window, the energy efficiency of the building can be greatly improved. The lower U-value and SC contribute to reducing both the heating and cooling loads of buildings in Hong Kong, thereby significantly reducing carbon emissions.

• 

Figure 3: Window Frame U-values

Warmer in winter and cooler in summer

THERM simulations were performed to evaluate the interior surface temperatures of the windows. The outdoor and indoor environmental conditions refer to Hong Kong, and the details are as follows:

• Winter outdoor nighttime temperature = 7℃
• Summer outdoor daytime temperature = 32℃
• Indoor temperature = 25℃
• Solar irradiance = 783 W/m²

• 

Figure 4: Window Frame Temperatures

Figure 4 shows the interior surface temperatures of the state-of-the-art thermally broken window and a regular window. In winter, the temperatures are 21℃ and 16℃; in summer, the temperatures are 27℃ and 40℃, respectively. The results show that the regular window is either too hot in summer or too cold in winter, while the state-of-the-art thermally broken window shows its excellent resilience to the environment.

By maintaining a stable indoor temperature, thermally broken windows increase the comfort level for the occupants. They eliminate cold spots near windows, contributing to create a more comfortable space throughout the year. Seppänen et al (2006) found that when the indoor temperature is at 20-25℃, the productivity of the occupants will reach over 98% of the best; while the temperature is lower than 15℃ or higher than 30℃, the productivity of the occupant will drop to below 90% [2].

Reduced condensation

According to the temperatures in Figure 4 and the indoor temperature of 25℃, we can predict when there will be condensation. Figure 5 shows the curve of dew point temperature at 25℃. For the regular window, condensation will occur when the indoor relative humidity reaches 58%, which is quite common in a humid winter in Hong Kong. However, for the thermally broken window, the indoor relative humidity must be 79% for the window to condense, which is unlikely to happen.

• 

Figure 5: Dew Point Temperature & Relative Humidity Chart

Lower noise transmission

Because sound vibrations are more easily transmitted within the integral aluminum frame and form the sound bridging, the thermal breaks in thermally broken windows can be treated as sound insulators. Although there is no concrete data to show the difference, it is believed that thermally broken windows perform no worse than regular windows. In fact, there are many commercially available products that have been tested to an STC (Sound Transmission Coefficient) of 45, which means that the noise control of the windows is even better than a 4″ hollow CMU (Concrete Masonry Unit).

• RESEARCH QUESTIONS

It is globally recognized that the window is one of the most important factors in achieving sustainability and carbon neutrality for buildings. For example, a window should have a U-value of 1.4 W/(m²·K) to meet the Low Energy Building (LEB) standard, a U-value of 0.8 W/(m²·K) to meet the Passivhaus standard, and an even lower U-value of 0.7 W/(m²·K) to meet the Nearly Zero Energy Buildings (NZEB) standard. None of these are possible without thermally broken windows. Hong Kong, only 50 km away from Shenzhen, has similar climate and challenges, but hardly uses any thermally broken windows other than the regular ones, which seems odd.

After extensive discussions with building professionals including architects, engineers, consultants and contractors in Hong Kong, the most common comments can be summarized as follows:

• Hong Kong uses the Overall Thermal Transfer Value (OTTV) as the sole indicator for the energy efficiency design of buildings, and there is no window U-value requirement in the code.

• For the project seeking LEED certification from the U.S. Green Building Council (USGBC), when window U-values are required by the ASHRAE standard, the engineers simply treat the glass center U-values as equivalent to the window U-values, neglecting the frames.

• The Hong Kong Buildings Department (BD) hesitates to approve thermally broken windows due to structural safety concerns. If you do want to use the thermally broken window in the project, you have to initiate a review process which will cost extra time and money, with no guarantee of approval.

This paper conducts an in-depth analysis of the above issues based on relevant theories, standards.

• DISCUSSIONS

• Overall Thermal Transfer Value (OTTV)

“Overall Thermal Transfer Value” was first introduced in 1975 by the ASHRAE Standard ASHRAE 90-75. The original intent of the OTTV index was to limit the energy consumption of the air conditioning systems by specifying the maximum allowable heat transfer value of the building facades and roofs. Based on the ASHRAE 90-75, Singapore formulated the first OTTV (renamed as ETTV) code in Asia in 1979. Both the ASHRAE and Singapore OTTV formulas consist of three factors:

• ① Heat conduction through opaque walls (Q_wc)
• ② Solar radiation through windows (Q_sol)
• ③ Heat conduction through windows (Q_gc)

In 1991, the Hong Kong Government investigated the feasibility of implementing OTTV control in Hong Kong through a consulting study by J. Roger Preston Co. (JRP). Following the JRP’s report and feedback from the building industry, the Hong Kong Government promulgated a legislative control on building envelope design through a Code of Practice for Overall Thermal Transfer Value in Buildings (OTTV Code) in 1995. Unlike ASHRAE and Singapore, the Hong Kong OTTV Code ignored the heat conduction through windows (Q_gc) (Figure 6). Hui (1997) discussed that the cause was the relatively small temperature difference between the outdoor and indoor, but didn’t disclose the exact data [3].

• 

Figure 6: Hong Kong, ASHRAE and Singapore OTTV Formulas

• Heat Conduction through Windows (Q_gc)

Thirty years have passed since JRP came to the conclusion regarding the heat conduction through windows (Q_gc). Design criteria, standards, technologies, and even ambient temperatures have changed since 1991, which can significantly affect the engineering results and the accuracy of the conclusion. Since the JRP’s report is not available, it is necessary to reproduce the original circumstance for further evaluation.

The ignored part, the heat conduction through windows (Q_gc), can be calculated using the Equation (1) from ASHRAE 90-75 as follows:

Q_gc = U_f×A_f×(T_out–T_in)                                                                                                                                                                      (1)

Where,

U_f = the U-value of the window, say 3.6 W/(m²·K)

A_f = the area of the window, say 40%, 50%, 60% of the whole external wall

T_in = the indoor temperature, say 25.5℃

T_out = the outdoor temperature

Regarding the T_out value, three conditions are discussed:

i) Lam et al. (1993) collected the outdoor temperature from May to October of 1980 – 1989, and an average of 27.3℃ was obtained [4].

ii) The similar approach as that of Lam was used to obtain an average of 28.2℃ for the period of 2014 – 2023 (Figure 7).

iii)           32.0℃ according to the outdoor temperature for the air conditioning specified in 2021 ASHRAE Handbook – Fundamentals.

• 

Figure 7: Average Outdoor Temperature in Hong Kong (1980-1989, 2014-2023)

The heat conduction through windows (Q_gc) can then be obtained (Table 2A).

Table 2A: Heat Conduction Through Windows (Q_gc)

Parameters	Af	Tout
i) 27.3℃	ii) 28.2℃	iii) 32.0℃
Heat conduction through windows (Qgc) (unit: W/m2)	40%	2.6	3.9	9.4
50%	3.2	4.9	11.7
60%	3.9	5.8	14.0

We know that the initial 1995-version OTTV Code set the OTTV limit at 35 W/m². The current 2019-version and BEAM Plus 2.0 both set the OTTV limit at 21 W/m². Again, as explained earlier, the Hong Kong OTTVs (35 or 21 W/m²) only consider the heat conduction through opaque walls (Q_wc) and the solar radiation through windows (Q_sol), however, if we consider the heat conduction through windows (Q_gc) as required by ASHRAE 90-75, the ratio of Q_gc to the total OTTV can be obtained (Table 2B, 2C).

Table 2B: Heat Conduction Through Windows (Q_gc) to Total OTTV (1995 version)

Af	Qwc + Qsol	Tout = 27.3℃
Qgc	Total OTTV	Qgc to Total OTTV
40%	35.0	2.6	37.6	6.9%
50%	3.2	38.2	8.5%
60%	3.9	38.9	10.0%

Table 2C: Heat Conduction Through Windows (Q_gc) to Total OTTV (2019 version)

Af	Qwc + Qsol	Tout = 28.2℃	Tout = 32.0℃
Qgc	Total OTTV	Qgc to Total OTTV	Qgc	Total OTTV	Qgc to Total OTTV
40%	21.0	3.9	24.9	15.6%	9.4	30.4	30.8%
50%	4.9	25.9	18.8%	11.7	32.7	35.8%
60%	5.8	26.8	21.7%	14.0	35.0	40.1%

Table 2B shows that the average ratio of Q_gc to the total OTTV is only 8.5%, which may be the reason why it was ignored in the initial OTTV Code. While Table 2C shows that the average ratio of Q_gc to the total OTTV increases to the non-negligible 18.8% and 35.8% with a stricter OTTV limit and higher outdoor temperature.

• Heat Conduction through Opaque Walls (Q_wc)

According to the Hong Kong OTTV Code, the heat conduction through opaque walls (Q_wc) can be calculated using the Equation (2):

Q_wc = A_w × U × α × TD_EQw                                                                                                                                                                    (2)

Where,

A_w = the area of the opaque wall, say 60%, 50%, 40% of the overall external wall

U = thermal transmittance of the opaque wall, say 3.2 W/(m²·℃) for a poor-insulated wall, 1.0 W/(m²·℃) for a well-insulated wall

α = absorption of the opaque wall, say 0.6

TD_EQw = equivalent temperature difference for wall, say 5.26℃

The heat conduction through opaque walls (Q_wc) can then be obtained (Table 3A).

Table 3A: Heat Conduction Through Opaque Walls (Q_wc)

Parameters	Aw	U
Poor-insulated 3.2 W/(m2·℃)	Well-insulated 1.0 W/(m2·℃)
Heat conduction through opaque walls (Qwc) (unit: W/m2)	60%	6.1	1.9
50%	5.1	1.6
40%	4.0	1.3

In the 1990s, there were not many buildings with insulated opaque walls, but since 2020 more buildings have been built with insulated opaque walls due to the stricter OTTV, so it is reasonable to compare the heat conduction through opaque walls (Q_wc) with windows (Q_gc)at the same time, see Table 3B.

Table 3B: Heat Conduction Through Opaque Walls (Q_wc) and Windows (Q_gc), unit: W/m²

Aw/Af	Heat conduction through opaque walls (Qwc)	Heat conduction through windows (Qgc)
1990s Poor-insulated	2020s Well-insulated	1990s Tout = 27.3℃	2020s Tout = 28.2℃
60/40	6.1	1.9	2.6	3.9
50/50	5.1	1.6	3.2	4.9
40/60	4.0	1.3	3.9	5.8

Table 3B shows that Q_wc is much larger than Q_gc in the 1990s, which proves that Q_gc is relatively insignificant, confirming the reason why it was ignored in the initial OTTV Code. However, in the 2020s, due to the stricter OTTV requirement, more buildings are built with insulated opaque walls to reduce Q_wc. Meanwhile, the U-values of the windows are not improved, resulting in Q_gc being greater than Q_wc.

• Summary

This study confirms that the heat conduction through windows (Q_gc) was relatively insignificant among the three factors (Q_sol > Q_wc > Q_gc) based on the reality of the 1990s. However, the conditions are different now, including the increasing outdoor temperature and insulated opaque walls, which makes Q_gc become the second important factor (Q_sol > Q_gc > Q_wc), accounting for a non-negligible 40% of the total OTTV (Figure 8). There are a few more things can make Q_gc even critical,including a larger window-to-wall ratio (WWR), a lower OTTV than the current 21.0 W/m², higher outdoor temperature, etc.

• 

Figure 8: Ratio of the Heat Conduction through Opaque Walls and Windows

It shall also be noted that the OTTV method was abandoned by ASHRAE in 1989 because of its limitations in meeting the ever-increasing demand for building energy conservation. Oraee et al. (2015) discussed the detailed limitations [5].

• Leadership in Energy and Environmental Design (LEED)

Leadership in Energy and Environmental Design (LEED) is the most widely used green building rating system in the world. If the project is seeking LEED certification from the U.S. Green Building Council (USGBC), the design team shall follow ASHRAE standards in terms of energy efficiency. ASHRAE 90.1 (1989 and later versions) refers to the NFRC Standard NFRC 100 Procedure for Determining Fenestration Product U-factors, an area-weighted method of calculating the window U-value (U_f) using the Equation (3):

U_f = (∑(U_fr×A_fr)+∑(U_eg× A_eg)+∑(U_g× A_g))/(A_fr+A_eg+A_eg)                                                             (3)

Where,

U_fr = heat transfer coefficient of the window frame, W/(m²·K)

U_eg = heat transfer coefficient at the edge of the glass, W/(m²·K)

U_g = heat transfer coefficient at the center of the glass, W/(m²·K)

A_f = area of the window frame, m²

A_eg​ = area of the glass edge, m²

A_g = area of the center of the glass, m²

LEED Building Design & Construction V4 – Minimum Energy Performance made it very clear that “Window U-values shall be input as the assembly U-value, which takes into account the U-value of the framing system. The center-of-glass value is not acceptable.” (Figure 9).

• 

Figure 9: LEED Building Design and Construction V4 – Minimum Energy Performance

However, after interviewing with some of the building professionals including mechanical engineers, façade engineers and LEED consultants, we notice that most of the LEED projects in Hong Kong use the glass center value instead of the assembly U-value. This may be because the OTTV Code actually treats the center-of-glass Shading Coefficient (SC) as the window SC (Figure 10).

• 

Figure 10: Section-7.5 of Hong Kong OTTV Code

The U-value discrepancy between center-of-glass and the assembled window (including glass and frame) is quite large. This is because the frame U-value can be 3-10 times the glass center U-value. The typical glass center U-value is 1.5 W/(m²·K) for an argon-infilled Low-e IGU. Due to the frame effect, the window U-value varies from 1.8 W/(m²·K) to 5.6 W/(m²·K). See Figure 11.

• 

Figure 11: Window U-value and Frame Area

Accuracy in calculating window U-values is essential, as building energy simulations are widely used today to evaluate building energy performance during the design phase. LEED requires to do the Whole-Building Energy Simulation to demonstrate the actual improvement of the proposed design so that projects can earn specific credits and ultimately achieve different levels of certification. Entering a relatively lower (glass-centered) U-value and receiving a better score risks losing it in the commissioning phase because the actual energy performance of the window is much worse than that of the glass center. This situation can be improved by using well-insulated windows that have U-values relatively closer to the glass center, but it is still a problem for now.

• Structural Safety

The Hong Kong Buildings Department (BD) is concerned that the outer and inner parts of the metal window frame are separated by the low-conductive thermal break material, i.e. PVC, polyamide, polyurethane, instead of being an integral frame, if the thermal break or its connection fails, there is nothing to hold the glass, which may cause the glass to fall off the building (see Figure 12).

• 

Figure 12: Safety Concern of the Thermal-break

The Hong Kong BD is not the first party having such concerns. Since the invention of thermally broken windows, there have been extensive discussions about the structural safety. The solution is simple and mature: bolt the outer metal frame along with the thermal break to the inner metal frame. In 2012, the Shanghai Government issued DGJ08-56-2012 Technical code for building curtain wall that regulates the use of bolts (Figure 13).

• 

Figure 13: Section 13 of DGJ08-56-2012 Technical code for building curtain wall

Figure 14 explains the mechanism of the bolts as a fail-safe measure. The window details shall be designed with this in mind. Despite the rarity of the thermal break failure, if it does fail, there is little chance to replace it, which means that the initial design should consider that the other performances (air and water tightness) won’t be affected due to the failure. Experienced facade designers should be able to deal with such scenarios.

• 

Figure 14: Thermal-break Safety Solution

Last but not the least, the penetrated bolts through thermal breaks introduce additional thermal bridges to the windows. The thermal bridging effect of the bolt shall be properly evaluated according to the relevant standards. For example, the NFRC simulation manual specifies an isothermal-plane method area-weighting the window with and without thermal bridging. Figure 15 shows the U-values of thermally broken windows with and without bolts. As can be seen from the figure, the spacing and size of the bolts will increase the window U-value for about 10% compared with the no-bolt thermally broken window [6]. However, the worst case with bolts for the thermally broken window still outperforms the regular non-thermally broken window. Therefore, the bolt solution should be considered reliable in terms of structural safety and thermal performance.

• 

Figure 15: Window U-values with/without Bolts

• CONCLUSIONS

Thermally broken windows have been recognized worldwide for three decades as one of the most important technologies to achieve sustainability and carbon neutrality for buildings, but have not yet been adopted in Hong Kong. Through the thorough analysis, this paper concludes that the barriers to the adoption of thermally broken windows are as follows:

• The original OTTV code ignores the heat conduction through windows (Q_gc), but now Q_gc has become the second important factor (40%) of the total heat transfer through the building envelope due to the latest changes in the OTTV limits, outdoor temperature and opaque walls.

• For LEED projects, the window U-value should take into account the U-value of the frame and glass using the area-weighted method, but the current practice is to treat the glass center U-value as the window U-value, which results in a large discrepancy between design and reality.

• The Hong Kong BD is concerned about thermal break failure resulting in glass falling from the building and doesn’t approve thermally broken windows. Bolting the outer metal frame along with the thermal break to the inner metal frame is a simple and mature solution, but has not been promoted by stakeholders due to the OTTV and LEED issues mentioned above.

These issues have not received sufficient attention, which delays the adoption of thermally broken windows, but more importantly, will probably lead to overestimation of building energy performance, or in other words, the actual energy performance of buildings is unlikely to meet the design requirements. To achieve Hong Kong’s 2050 carbon neutrality target, the application of thermally broken windows as one of the best building envelope technologies contributes to improving the energy efficiency of buildings and reducing the carbon footprint. Therefore, it is necessary to take timely actions to address the above three issues and remove the barriers, so that thermally broken windows can be adopted according to their technical value.

• LIMITATIONS AND FUTURE WORK

This paper did not include simulations comparing the actual energy performance of buildings in the Hong Kong climate with and without thermally broken windows. Comparable simulations and experimental tests are crucial to validate the actual benefits of thermally broken windows, and will be done in the next step.

REFERENCES

[1]       Wong, S., 2017, Energy Saving Potential of Thermal Broken Fenestration System in Hot Climate Counties, World Sustainable Built Environment Conference 2017 Hong Kong, Track 3: Advancing SBE Assessments, p. 868-873.

[2]       Seppänen, O., Fisk, W.J., Lei, Q.H, 2006, Effect of Temperature on Task Performance in Office Environment, Lawrence Berkeley National Laboratory.

[3]       Hui, S.C., 1997, Overall thermal transfer value (OTTV): how to improve its control in Hong Kong, Proc. of the One-day Symposium on Building, Energy and Environment, vol. 16, p. 12-21.

[4]       Lam, J.C., Hui, S.C.M. and Chan, A.L.S., 1993, Overall thermal transfer value control of building envelope design part 2-OTTV parameters, Hong Kong Engineer, 21(9), p. 40-44.

[5]       Oraee, M. and Luther, M.B., 2015, The next step in energy rating: the international ETTV method vs. BCA Section-J Glazing Calculator, 49th International Conference of the Architectural Science Association, p. 423-433.

[6]       Zhou, Y.Z., 2023, Impact of bolt penetrating the insulation layer on the curtain-wall thermal performance based on THERM simulation, Proc. Of the 15th Building Physics Conference of China, Building Facades, p. 959-962.