THERMAL PERFORMANCE OF COMPLEX

MULTIMATERIAL FRAMES FOR EVACUATED GLAZING

Yueping Fang, Philip C. Eames, Trevor J. Hyde and Brian Norton

Centre for Sustainable Technologies, School of the Built Environment, University of Ulster, Newtownabbey, BT37 0QB, N. Ireland

Tel: +44 (0) 28 9036 8567, Fax: +44 (0) 28 9036 8239, email:

Abstract - The thermal performance of complex multimaterial frames framing evacuated glazings were simulated using a 2-dimensional finite element model and validated experimentally using a guarded hot box calorimeter. The simulated 0.5m by 0.5m evacuated glazing consisted of two low-e coated 4 or 6mm thick glass panes supported by 0.32mm diameter pillars separated on a 25mm grid and contiguously sealed by a 10mm metal edge seal. The predicted heat loss from a window with an evacuated glazing and a complex multimaterial frame is about 80% of that for a window comprised of an evacuated glazing set in a single material solid frame.

1. INTRODUCTION

Evacuated glazing (Griffiths et al., 1998; Hyde et al., 2000) is made of two glass panes with a 0.15mm vacuum space, supported by a pillar array arranged on a 25mm square grid, contiguously sealed by a 10mm metal edge seal. For window designs using evacuated glazing with a very low heat loss, the frame must exhibit low heat loss commensurate with that of the glazed element. Window frames are made commonly from aluminium, wood or plastic. Aluminium is strong and lightweight, can be fabricated easily by extruding and requires little maintenance. However aluminium has a high thermal conductivity of 237Wm-1K-1. Although heat transfer between separate panes set within an aluminium frame may be reduced by including a thermal break in the frame, the risk of internal condensation means it may not be the best choice for cold climates (Aschehoug et al., 1994). Hardwood window frames generally have good strength and low thermal conductivity (typically 0.12 Wm-1K-1, Arpaci et al., 2000). Interior condensation on a wooden frame is less common. However a wooden frame can require considerable maintenance, including periodic painting to protect from moisture ingress, cracking and warping. The most common plastic framing material is polyvinylchloride (PVC). A PVC window frame requires little maintenance. However PVC window frames are not inherently strong, which limits both the size of the frame and the weight of the glass to be framed. Thus for large glazed areas, an internal metal extrusion is often employed to stiffen a PVC frame. In this work a novel complex multimaterial frame was designed and its thermal performance was investigated using an experimentally validated 2-dimensional finite element model developed in-house.

2. FRAME DESIGN

The essential criteria for the frame design was to provide a high degree of thermal insulation yet maintain the structural strength and rigidity necessary to support the evacuated glazing during its serviceable life. Since the thermal conductivity of some insulants are much lower than those of common frame materials, the use of insulant to form the frame was considered. To provide frame rigidity and combat the inherent low mechanical strength and rigidity of insulants, an exoskeleton structure made of common frame materials was employed. The geometrical configuration of the exoskeleton frame, as shown in fig. 1, was selected on the basis of strength and thermal performance. Aluminium, wood, polypropylene, polyethylene, polyvinylchloride or teflon may be used to form the exoskeleton structure, with a range of solid insulants filling the internal cavities.

Fig. 1 Acomplex multimaterial window frame housing an evacuated glazing sample with a metal edge seal.

3 EFFECT OF FRAME WIDTH ON THE THERMAL PERFORMANCE OF AN EVACUATED GLAZED WINDOW

A finite element model (Eames and Norton, 1993) was modified to analyse the heat transfer through a complex multimaterial frame framing an evacuated glazing. The temperatures of the indoor and external ambient air were set to 21.1ºC and -17.8ºC respectively. The convective heat transfer coefficients on the indoor and the external ambient surfaces were 8.3and 30Wm-2K-1, to correspond to measurement standards for winter conditions (ASTM, 1991). The infill insulant was expanding foam (Type: CP0296/1A) (Anon, 2000). For a frame with a depth H of 40mm and an evacuated space heat transfer coefficient of 0.4Wm-2K-1, the thermal performance of an evacuated window system with a range of frame widths is presented in fig. 2. The exoskeleton frame thickness was 3mm, frame rebate depth 17mm and the width of metal edge seal was 3mm.

Fig. 2 Thermal performance of an evacuated glazing window system with frame widths, w, in the range 40-60mm.

Increasing frame width leads to a decrease in the heat transfer coefficients of the frame area and the overall window system. The rate of decrease in the frame area heat transfer coefficient is greater than that of the overall system heat transfer coefficient. The heat transfer through the frame decreases with increasing frame width and that through the glass area is relatively stable. The average surface temperatures of the internal surface increases by 0.3ºC and the external surface temperature decreases by less than 0.1ºC due to the rate of heat transfer through the glass decreasing.

4. EVACUATED GLAZING THERMAL PERFORMANCE AS A FUNCTION OF FRAME REBATE DEPTH

The thermal performance of the evacuated glazing systems with a complex frame width, w, of 50mm and depth, H, of40mm and a range of rebate depths, z, of between 10 and 25mm and a constant evacuated space heat transfer coefficient of 0.4Wm-2K-1 were evaluated. The simulated results presented in fig. 3 show that increasing the rebate depth leads to a decrease in the heat transfer coefficients of the overall window system, the frame and the exposed glass areas with a concomitant increase in the internal side average surface temperatures and a decrease in the external side average surface temperature. Increasing the frame rebate depth from 10 to 25mm increases the path length for heat transfer through the metal edge seal by 30mm leading to a reduction in the heat transfer coefficient of the window system by 0.08Wm-2K-1. Increased frame area reduces the exposed glass area. When designing an optimal evacuated glazing system, the trade-off between solar gains and thermal performance of the total window system would need to be considered.

Fig. 3 Thermal performance of an evacuated glazing with recess depths of 10-25mm in a complex expanding foam filled frame.

5. THE EFFECT OF THE INSULANT MATERIALS ON THE THERMAL PERFORMANCE OF AN EVACUATED GLAZED WINDOW

The thermal performance of an evacuated glazing rebated into a complex multimaterial frame with several insulant materials was simulated and the results are presented in fig 4.

Fig. 4 Predictedthermal performance of an evacuated glazing with a range of insulant materials within the cavities of a polypropylene exoskeleton frame

The insulants were expanding foam, urethane, glass wool and vermiculite flakes. For these simulations, the exoskeleton material was polypropylene with a thickness of 3mm. Fig. 4 shows that increasing insulant thermal conductivity leads to a significant increase in the heat transfer coefficient of the frame area and a small decrease in the heat transfer coefficient of the glass area, this results in an increase in the total heat transfer through the window system.

6. THERMAL PERFORMANCE AS A FUNCTION OF EXOSKELETON THICKNESS

The predicted thermal performances for varying thicknesses of exoskeleton fabricated from aluminium and polypropylene with expanding foam infill are illustrated in figs. 5 and 6 respectively.

Fig. 5 Thermal performance of an evacuated glazing with an aluminium exoskeleton thickness in the range 0.5-3mm for a complex multimaterial frame.

Fig. 6 Predicted thermal performance of an evacuated glazing with a polypropylene exoskeleton thickness in the range 0.5-3mm.

It can be seen from figs. 5 and 6 that increasingthe exoskeleton framework thickness significantly increases the heat transfer coefficient through the frame area while that through the glass areas remains almost constant. For an aluminium frame, fig. 5 shows that the rate of increase in heat transfer coefficient of the glazing system decreases with increasing exoskeleton framework thickness, i.e. the effect of the exoskeleton thickness on the thermal performance of the evacuated glazing system decreases. For the evacuated glazing system modelled, the increase in heat transfer coefficient of the overall window is less than 0.195Wm-2K-1 when the aluminium exoskeleton thickness increases from 0.5mm to 3mm. The corresponding increase in heat transfer coefficients of frame and glass areas were 366.53 and 0.05Wm-2K-1 respectively. Although the increase in heat transfer coefficient of the frame area is great, the increase in heat transfer coefficient of the overall window system is less than 0.195Wm-2K-1. From the thermal view point the effect of using a thicker aluminium exoskeleton frame should not be sufficient to prevent its use if required to maintain the correct strength and rigidity for the window system.

When the polypropylene exoskeleton thickness increases from 0.5 to 3mm, fig. 6 shows that the changes in heat transfer coefficient of the overall window system, frame and glass areas are 0.09, 0.58Wm-2K-1 and less than 0.01Wm-2K-1 respectively. The strength consideration is more important than that of the thermal performance for practical evacuated glazing window systems.

7. THERMAL PERFORMANCE AS A FUNCTION OF THE FRAME EXOSKELETON MATERIALS

Different exoskeleton frame materials give distinct thermal performances for evacuated glazing window systems. A range of exoskeleton frame materials were modelled as shown in fig 7. The frame depth was 40mm, width 50mm, rebate depth 17mm and exoskeleton thickness 3mm. The cavity infill material was expanding foam with a thermal conductivity of 0.024Wm-1K-1. The heat transfer coefficient of the vacuum space was set at 0.4Wm-2K-1. The predicted thermal performance for these systems is shown in fig. 7. It can be seen that increasing the exoskeleton frame thermal conductivity leads to an increase in the heat transfer coefficient of the frame area and a reduction in the heat transfer coefficient of the glass area. The rate of increase in the heat transfer through the frame is greater than the rate of decrease in the heat transfer through the glass area, leading to an increase in the total heat transfer through the glazing system.

Fig. 7 Thermal performance of an evacuated glazing system with different exoskeleton materials.

When the simulated exoskeleton frame was aluminium with similar dimensions to the wood and plastic frames, the heat transfer coefficient of the overall evacuated glazing window system was 2.59Wm-2K-1. The heat transfer coefficient of the overall evacuated glazing window systems with the nonmetal exoskeleton frame materials simulated in fig. 7 was less than 1Wm-2K-1.

Depending on the glazing size, the magnitude of the effect of using metal or nonmetal exoskeleton frame materials on the thermal performance of an evacuated glazing can be significant. Using the nonmetal materials in fig. 7, as the exoskeleton frame material, although the heat transfer coefficient variation of the frame area exceeds 1Wm-2K-1, the difference in thermal performance of the overall window systems is less than 0.2Wm-2K-1. This resulted from two factors: (i) since the frame area was 14% of the total glazing window area, the effect of the frame area thermal performance on that of overall window area was relatively small; (ii) the thermal conductivities of the non-metal materials used were similar.

8. COMPARISON OF THE PERFORMANCE OF A SOLID AND A COMPLEX MULTI-MATERIAL FRAME FOR AN EVACUATED GLAZING

The thermal performance of a frame as shown in fig. 8 made of a single solid material with no cavities was simulated. The predicted thermal performance results for an evacuated glazing with a complex multimaterial frame and a solid wood frame are compared in fig. 9.

Fig. 8 An evacuated glazing with a metal edge seal rebated into a single material solid frame.

The solid single material frames considered were made of either polypropylene or wood. The complex multi-material frame was made of either a polypropylene or wood exoskeleton framework filled with expanding foam. The dimensions of both the single and complex multimaterial frames were the same, with a 50mm width, 40mm height and a 17mm rebate depth.

Fig. 9 Thermal performance of an evacuated glazing with a solid polypropylene or wood frame and a complex multimaterial frame with the same exoskeleton materials.

Fig. 9 shows that the thermal performance of the evacuated glazing window system with the complex multimaterial frame is significantly better than that with the solid single material frame. The heat transfer coefficient of the former system is about 80% of the latter system. The heat transfer coefficients of the evacuated glazing with the complex multimaterial frames made with polypropylene or wood exoskeleton are similar due to the thermal conductivities of polypropylene and wood being similar and the significant effect of the infill material within the cavities of both frames in determining the overall heat transfer.

9. EXPERIMENTAL INVESTIGATIONS

Fig. 10 Schematic diagram of the constructed guarded hot box.

A guarded hot box calorimeter, as shown in fig. 10, was designed and constructed in accordance with ISO, 1994 to measure the thermal performance of evacuated glazing with the complex expanding foam filled frames.

The heat transfer through the test sample can be determined (ISO, 1994) by:

(1)

Heat conductance of specimen is given by

(2)

The environmental temperature Tn represents the weighting of air and radiant temperatures and is used for the purpose of determining the heat flow to the surface.

(3)

The overall heat transfer coefficient of the specimen was obtained by using

(4)

To ease fabrication, the structure of complex multimaterial frame A shown in fig. 1 was modified to frame B with the geometrically simplified structure as shown in fig. 11. The isotherms within the mask wall were calculated using the finite element model and are shown in fig. 11. The 4mm thick exoskeleton structure was made from wood with expanding foam used to fill the cavities. The width of the metal edge seal was 10mm.

Fig. 11 Enlarged view of the isotherms in the evacuated glazing rebated 16mm into a complex frame B.

Fig. 12 Predicted thermal performance of the two 50mm wide and 40mm high complex expanding foam filled frames A and B with varying vacuum space heat transfer coefficient.

Fig. 11 shows that the metal edge seal significantly affected the temperature distribution. Predicted and measured flanking losses were analysed by Fang et al., (2002). The experimental uncertainties in the measurements of the flanking loss and glazing heat transfer coefficients were determined to be 1.4% and 5% respectively. After the complex frame A was simplified to the frame B, its thermal performance was simulated and is presented in fig. 12.

Fig. 12 shows that simplifying complex frame A to frame B leads to the heat transfer coefficient of the overall window system, frame and glass areas increasing. For frame B, increasing the heat transfer coefficient of the evacuated space leads to larger increases in the heat transfer coefficient of the overall window area and the glass area than those of frame A. When the heat transfer coefficient of the vacuum space is 0.4Wm-2K-1, the heat transfer coefficient of the overall window system increases by about 0.04Wm-2K-1for frame B. When the heat transfer coefficient is 1.2Wm-2K-1, the heat transfer coefficient of the overall window system increases by approximately 0.12Wm-2K-1. The greater the evacuated space heat transfer coefficient, the greater the increase in the overall window system heat transfer coefficient. The thermal performance degradation caused by simplifying the frame structure is small.

A comparison was made between the solid and complex frames. The predicted and experimentally determined heat transfer through the specimen and the mask wall are presented in table 1. In the measurements, the air temperatures in the hot and cold chambers, for the solid wood framed system, were 39.0ºC and 13.4ºC; and for the complex multimaterial frame, were 38.6ºC and 11.2ºC. The dimensions of the solid and complex multimaterial frames were the same, 50mm wide, 40mm high with a 16mm rebate depth.

Frame
type / ∆T
ºC / W / /
/ /
/

Solid / 25.6 / 21.8 / 3.28 / 3.24 / 3.39 / 3.40 / 1.35
Complex / 27.5 / 20.8 / 3.22 / 3.23 / 2.80 / 2.83 / 1.24

Table 1 Comparison between the measured and predicted heat transfer coefficients of the exposed glass and total glazing areas with solid and complex frames.

It can be seen from table 1 that the complex expanding foam filled frame reduces the heat transfer coefficient by approximately 20%.

Fig. 13 Comparison of the measured and predicted mean surface temperatures of the evacuated glazing with a complex frame with a 16mm rebate.

The measured heat transfer coefficients were very close to the predictions calculated using the finite element model. The measured and predicted mean surface temperatures of the exposed glass and mask wall are presented in figs. 13 and 14.

Figs. 13 and 14 show that the measured mean surface temperatures for the exposed glass and total glazing areas are in very good agreement with the predicted values calculated using the finite element model. In fig 13, the measured frame surface temperatures on the hot and cold side surfaces deviated from those predicted by 4%, the measured glass surface temperature deviated from those predicted by less than 3%.

Fig. 14 Comparison of measured and predicted mean surface temperatures of the mask wall and evacuated glazing with a solid wood frame with a 16mm rebate.

9.1EFFECT OF COMPLEX MULTIMATERIAL FRAME REBATE DEPTH

An evacuated glazing sample was rebated into two complex multimaterial frames with rebate depths of 16mm and 25mm. The width and height of both frames was 50 and 40mm respectively. The heat transfer through the vacuum glazing and the mask wall measured using the guarded hot box calorimeter is shown in table 2. The comparison between the measured and predicted heat transfer coefficients of the exposed glass and total window system are presented in fig 15.