laminated glass

41
EXTENDING THE PERFORMANCE OF LAMINATED GLASS FOR TODAY'S BUILDING DESIGN CHALLENGES Stephen Bennison Research Associate E.I. Dupont de Nemours & Co Inc USA Abstract Laminated glass continues to grow as the material of choice for much architectural glazing. This growth is driven by the recognized benefits of improved safety and security, improved acoustic functionality and extended energy management performance when used in conjunction with coated glass. Along with this growth come many requests from architects and façade consultants for extending laminate performance. In this paper we discuss strategies for addressing three growing needs: 1) extending strength and temperature performance of laminates, particularly in applications such as minimally supported glazing, where the laminate is usually subjected to bending deformation; 2) extending the durability of laminated glass, specifically with respect to edge delamination and stability; and 3) extending the security function of laminates to mitigate terrorist threats from bombs. Emphasis will be placed upon: interlayer choice, laminating strategies and total system design. In many cases the full potential of the laminate can only be realized by optimizing the glazing attachment system. A new paradigm will be introduced for laminate attachment. Examples of several architectural jobs will be given that highlight the design rationale behind materials selection, processing strategies, attachment and system details. 1

Post on 28-Nov-2014

466 views

Category:

Documents


2 download

TRANSCRIPT

Page 1: Laminated Glass

EXTENDING THE PERFORMANCE OF LAMINATED GLASS FORTODAY'S BUILDING DESIGN CHALLENGES

Stephen BennisonResearch AssociateE.I. Dupont de Nemours & Co IncUSA

Abstract

Laminated glass continues to grow as the material of choice for much architectural glazing. This growth is driven by the recognized benefits of improved safety and security, improved acoustic functionality and extended energy management performance when used in conjunction with coated glass. Along with this growth come many requests from architects and façade consultants for extending laminate performance.

In this paper we discuss strategies for addressing three growing needs: 1) extending strength and temperature performance of laminates, particularly in applications such as minimally supported glazing, where the laminate is usually subjected to bending deformation; 2) extending the durability of laminated glass, specifically with respect to edge delamination and stability; and 3) extending the security function of laminates to mitigate terrorist threats from bombs. Emphasis will be placed upon: interlayer choice, laminating strategies and total system design. In many cases the full potential of the laminate can only be realized by optimizing the glazing attachment system. A new paradigm will be introduced for laminate attachment. Examples of several architectural jobs will be given that highlight the design rationale behind materials selection, processing strategies, attachment and system details.

1

Page 2: Laminated Glass

Stress Generation In Laminated Glass Plates

Dr. Joji Suetomi, Glass Research Center Central Glass Co.,Ltd.Mr. Shin Omi, Central Glass Co.,Ltd.Mr. Kazutoshi Nakaya, Central Glass Co.,Ltd.Dr. Shin’ichi Aratani, Central Glass Co.,Ltd.Glass Processing Days, 18–21 June 2001

Abstract

Laminated glasses have been used as windows for their safety properties. Recently, demands for thinner laminated glass have been increasing for the conservation of natural resource and the save of energy. From the point of view, the values of stress generation in thinner laminated glass are very important for considering the safety properties. One of authors reported the stress generation values with the sample of 3 mm thick glass + PVB film + 3 mm thick glass. However, the stress generation values with a thinner sample like 2 mm thick gkass + PVB film + 2 mm thick glass doesn’t have been reported yet. In order to investigate the stress generation in various laminated glass, measurements of stress were attempted by ring on ring method. Laminated glass plates of 2 mm thick glass + PVB film + 2 mm thick glass were used as specimens. In addition, numerical analysis was done by the solid model. As the results, the relationship between stress and loading velocity was recognized at 284 K but the stress at 294 K and 302 K was not changed by loading velocity. It was also recognized that the stress of 30 mil laminated glass was larger than that of 60 mil laminated glass at 284 K. These phenomena were caused by the visco-elastic characteristic of laminated film.

Introduction

Laminated glasses have been used as windows of buildings, automobiles and trains etc. for their safety properties. Recently, demands for thinner laminated glass have been increasing for the conservation of natural resource and the save of energy. For example, laminated glass (2 mm thick glass + PVB film + 2 mm thick glass) has been used mainly for windshield of automobile.

From the point of view, the values of stress generation in thinner laminated glass are very important for considering the safety properties. However, the stress generation phenomena on loading in laminated glass are very difficult to analyze, because laminated glass consists of two glasses and PVB film which has visco-elastic property and its transformation is different from that of single glass plate. Some papers was reported for stress in laminated glass. [1], [2] One of authors reported the stress generation values with the sample of 3 mm thick glass + PVB film + 3 mm thick glass. [3] Recently, it was reported that the allowable pressure of failure probability of 0.001 was 0.8 to 1.1 times larger than that of monolithic glass at 293 K for flexibility ratio (length of short dimension / total glass thickness) ranging from 56 to 222 for 6.3 mm thick laminated glass by ring on ring method. [4] However, the stress generation values with a thinner sample like 2 mm thick gkass + PVB film + 2 mm thick glass doesn’t have been reported yet. In order to investigate the stress generation in various laminated glass, measurements of stress were attempted by ring on ring method. In addition, numerical analysis was done by the solid model.

Experimental

The stress in laminated glass was measured by use of strain gages. Laminated glass plates of sizes 420 mm x 420 mm and two kinds of PVB thickness 30 mil (0.76mm) and 60 mil (1.52mm) were used as specimens. Experimental conditions were shown as follows;

Sample:

Laminated glass (size; 420 mm x 420 mm)A ; FL2 + 30 mil (0.76 mm) PVB + FL2B ; FL2 + 60 mil (1.52 mm) PVB + FL2

Loading method : Ring on ring methodLoading ring diameter ; 70 mm

2

Page 3: Laminated Glass

Supporting ring diameter ; 210 mmLoad : 0-981 NLoading velocity : 0.02-100 mm/min.Temperature of samples: 284 K, 294 K and 302 KControlled for more than 4 hours

The schematic view of experiment device is shown in figure 1.

Fig. 1 The schematic view of experiment device.

Numerical Analysis

Example of mesh for the specimen to analyze the stress generation is shown in figure 2. 1/4 model of specimen was splitted to 21,276 nodes and 18,120 elements.

- Constraint : Circle of radius 105 mm on bottom of the specimen- Load : Circle of radius 35 mm on top of the specimen

It was hypothesized that the Young’s modulus of PVB film didn’t change by stress, and then the static stress generation was analyzed in the case of this hypothesis. Numerical analysis was done by the solid model in each case that several kinds of the Young’s modulus were supposed. Other factors which were used for this numerical analysis is follows;

- Poisson’s ratio of glass : 0.23- Poisson’s ratio of PVB : 0.3- Young’s modulus of glass : 71.5Gpa

Fig. 2 Example of mesh for the specimen

3

Page 4: Laminated Glass

Results

The relation between stress of 1st-4th faces and loading velocity in 30 mil PVB laminated glass is shown in figure 3. The value of 4th face’s stress was larger than that of 2nd face’s stress in tension. In addition, the value of 4th face’s absolute stress was larger than that of 1st face’s absolute stress.

Fig. 3 Relation between stress of 1 st-4th faces and loading velocity. (30 mil PVB, 284 K)

The order of absolute stress is follows;

Larger ← → Smaller4th face > 1st face > 2nd face > 3rd face

The relationship between stress in 30 mil PVB laminated glass and loading velocity at 284 K was recognized. The faster loading velocity was, the less the absolute stress of each face was at 284K. The relation between stress of 4th face and loading velocity in 60 mil PVB laminated glass is shown in figure 4. In 60 mil PVB laminated glass, the relationship between stress of 4th face and loading velocity at 284 K was also recognized. Moreover, the change of stress in 60 mil PVB laminated glass was bigger than that in 30 mil PVB laminated glass. However, the stress at 294 K and 302 K didn’t depend on loading velocity. In addition, the stress at 294 K and 302 K was bigger than that at 284 K.

The relation between stress of 4th face and temperature in 60mil PVB laminated glass is shown in figure 5. The stress at 284 K was smaller than those at 294 K and 302 K.

Fig. 4 Relation between stress of 4th face and loading velocity in 60 mil PVB laminated glass.

4

Page 5: Laminated Glass

When the loading velocity was more than 1 mm/min., the change of stress by temperature was remarkable. The relation between stress of 4th face and thickness of PVB film is shown in figure 6. The stress in 30 mil PVB laminated glass was larger than those on 60 mil PVB laminated glass. At 284K, the change ratio of stress by PVB thickness didn’t depend on the loading velocity.

Fig. 5 Relation between stress of 4th face and temperature in 60 mil PVB laminated glass.

Results of stress measurements analysis were summarized as follows;

– The values of 4th face’s absolute stress were larger than those of 1st face’s absolute stress.– The relationship between stress and loading velocity was recognized at 284 K. At 284 K, the faster a loading

velocity was, the smaller a value of stress in laminated glass was.– Values of stress in laminated glass were not changed by loading velocity at 294 K and 302 K.

Fig. 6 Relation between stress of 4th face and thickness of PVB at 284K.

– Stress in laminated glass at 284 K was smaller than those at 294 K and 302 K.– The dependence on loading velocity in 60 mil PVB laminated glass was larger than that in 30 mil PVB

5

Page 6: Laminated Glass

laminated glass.

One sample of stress distribution calculated on 4th face of laminated glass is shown in figure 7. According to the results of numerical analysis, the apparent Young’s modulus of laminated film was estimated on the base of the actual stress values in laminated glass at 294 K. As the result, the apparent Young’s modulus of laminated film was about 10 MPa. This apparent Young’s modulus means the laminated glass plate doesn’t behave as one lump of object as monolithic glass plate, because the Young’s modulus of glass is about 71.5 GPa. This is to certify that the visco-elastic characteristic of laminated film affects the stress generation inlaminated glass very much.

Fig. 7 Example of simulated result (30 mil PVB).

Conclusions

The values of stress generation in laminated glass specimens were changed with loading velocity, temperature of specimens and thickness of laminated film. The relationship between stress and loading velocity was recognized at 284 K but the value of stress was not changed by loading velocity at 294 K and 302 K. The dependence on loading velocity in 60 mil PVB laminated glass was larger than that in 30 mil PVB laminated glass.

The values of stress at 284 K were smaller than those at 294 K and 302 K. In addition, the values of 4th face’s absolute stress were larger than those of 1st face’s absolute stress. These phenomena were caused by the visco-elastic characteristic of laminated film. The result of numerical analysis was also to certify that visco-elastic characteristic of laminated film affected the stress generation of laminated glass very much. Therefore, laminated glass didn’t behave as one lump of object as monolithic glass plate and stress generation in laminated glass was complex phenomenon from visco-elastic properties of laminated film. Those remarks were follows;

• The values of stress in laminated glass depend on the loading velocity at 284 K.• The laminated films at 294 K and 302 K were more flexible than that at 284 K and so the transformation of laminated films at 294 K and 302 K occurred immediately.

• The laminated film decreased the flexibility of laminated glass and so the thicker the laminated film was, the smaller the stress in laminated glass was.

Consequently, when the stress generation in laminated glass is considered, it’s very important to understand the visco-elastic characteristic of laminated film.

Acknowledgement

We are thankful to K. Miyazaki and his staffs ( Central Glass Co., Ltd ) for their help to make experiment

6

Page 7: Laminated Glass

samples.

Reference

[1] J. E. Minor and P. L. Linden : “ Failure Strength of Laminated Glass “, J. Struct. Eng., Vol.116, No.4, pp.1030 – 1039, April, 1990.

[2] R. A. Behr, J. E. Minor and H. S. Norville : “ Structural Behavior of Architectural Laminated Glass “, J. Struct. Eng., Vol.119, No.1, pp.202 – 222, January, 1993.

[3] S. Aratani, M. Kikuta, T. Murakami and M. Ida : “ Stress measurement of laminated glass ”, Proc. 12th Fall Meeting Ceram. Soc. Japan, pp.80, October, 1999.

[4] T. Murakami, T. Kanasugi, M. Miyamoto and Y. Fujitani : “ Study on wind pressure resistance of laminated glass in buildings using failure probability theory “, J. Struct. Constr. Eng., AIJ, No.537, pp.141 – 148, November, 2000.

Strength of Laminated Safety Glass

Dr. Stephen J. BennisonDr. C. Anthony Smith, Mr. Alex Van Duser & Dr. Anand Jagota,Glass Laminating Products, E.I. DuPont de Nemours & Co. Inc.Glass Processing Days, 18–21 June 2001

Abstract

The strength of laminated safety glass continues to be treated in an overly conservative fashion by façade engineers, consultants and architects in general. The origin of this conservatism is a belief that the compliant nature of a polymer interlayer, relative to glass, results in poor transfer of shear stresses between the glass components of the laminate hence leading to greater glass stress for a given loading/support condition. In this presentation we re-examine the issue of stress development in laminated glass and present an objective protocol for designing laminates for required strength performance. The approach is based upon finite element stress analysis, with a constitutive model for the polymer interlayer, coupled with a statistical glass fracture model and a description of glass stress corrosion cracking. The approach has allowed us to develop a series of design charts for laminated glass that allows selection of interlayer thickness and glass thickness for specified loading/support scenario, temperature and loading rate. An important conclusion of our work is that four-sided simply supported laminated glass plates under uniform pressure demonstrate comparable strength to equivalent monoliths. Indeed, in certain cases laminates may outperform equivalent monoliths. This finding is understandable since transfer ofshear stresses between glass plies plays only a minor role in stress development during large deformation of

7

Page 8: Laminated Glass

plates. Our design charts and the approach are being considered by various standards bodies around the world and should lead to an objective treatment of laminated glass that will allow its true structural benefits to be realized

Introduction

It has been nearly thirty years since Hooper published his seminal work on the load bearing capacity, or “strength” of laminated glass [1]. In this study Hooper looked at stress development in PVB-based laminated glass beams both from a theoretical and experimental approach. He concluded that the load bearing capability oflaminated glass should fall between two limits: 1) the monolithic limit, in which the laminate behaves essentially in an equivalent manner to monolithic glass and 2) the layered limit in which the two glass plies slide over one another and the laminate displays one half the strength of the equivalent monolith. The limit is determined by the effectiveness of shear stress transfer across the interlayer and is strongly affected by loading rate and temperature due to the viscoelastic nature of plasticized PVB. Hooper stated clearly in this paper that the results only strictly apply to beams and should only be taken as a guide in the absence of design information for the specific loading/support case of interest.

Much work has been published since looking into stress development of laminated glass, particularly for 4-side support, uniform pressure loading of plates, which is the most common loading/support scenario for glazing [2-7]. The conclusion from this body of work is that for laminate plates, glass stress development is essentially determined by membrane stretching and that the interlayer shear properties now take on a diminished role in the plate strength properties. As such, laminated glass often displays equivalent strength properties to monoliths.

Despite this extensive body of work there has been little change in the way laminated glass is treated in design and a suspicion remains regarding the strength of laminated glass. Laminated glass remains severely over-designed in the world of architecture.

This suspicion is understandable since designers are not likely to treat laminates as equivalent to monoliths if there is belief that bending stresses may contribute in some way to stress development in a plate application of interest. Also, there is a belief that laminate strength properties are strongly affected by temperature and loading duration. Designers in general would prefer to have a comprehensive design approach to laminates that give anobjective evaluation of performance for the case of interest. We have been developing such a comprehensive design approach for laminated glass over the past several years [8,9]. The approach is based upon finite element stress analysis, with a constitutive model for the polymer interlayer, coupled with a statistical glass fracture model and a description of glass stress corrosion cracking. In this contribution we present some key findings from our approach to strength design for laminated glass.

Layered Experiment- The “Oil” Laminate

As an example of how the overall stress state influences glass stress development and strength performance we present results of a simple experiment. In this experiment we have studied glass stress development in two specimens: 1) a 6 mm monolithic glass plate and 2) a laminate comprising of two plates of 3 mm glass with a thin film of mineral oil separating the plates. We refer to this as our “oil” laminate. Note that the two 3 mm plates were chosen in thickness to match closely the total 6 mm thick monolith. Fifteen rosette strain gages were attached to each specimen and each plate was loaded with uniform pressure on 4-side supports. Figure 1 plots the maximum principal stress development in each plate as a function of applied pressure. The key point to note is that glass stress development is essentially equivalent in the two specimens.

8

Page 9: Laminated Glass

Figure 1. Glass stress development in a 6 mm monolithic glass plate and a laminate comprising 2 x 3 mm glass plates with and “oil” interlayer. During uniform pressure loading and 4-side supports, glass stress development is essentially equivalent.

Looking in detail, at low applied pressures, the monolith is slightly stronger than the “oil” laminate since bending stresses dominate at small deflections, but at higher applied pressures, the “oil” laminate is actually slightly stronger than the monolith! This result is counter-intuitive. However, the “oil” laminate has transitioned to the membrane dominated stress state more quickly than the monolith and the additional stiffening from membrane stretching leads to a slight benefit. Note this effect has been predicted from stress analyses performed by Vallabhan and co-workers over the years [4,5,7]. For the range of glass design stresses treated in most standards the monolith and “oil” laminate are equivalent. If we consider a laminate made from an interlayer with finite shear stiffness that contributes to the overall thickness we may expect further strength benefits over the equivalent monolith. In support of this point, Figure 1 also plots the measured glass stress development in a 3 mm glass / 2.29 mm PVB / 3mm laminate.

As can be seen from these data, the PVB-based laminate requires significantly greater applied pressure to generate the specified design glass stress, and therefore, displays greater strength behavior than the monolith. These observations and the body of published literature cited supports our argument that laminated glass is severely over-designed in most current glazing design practice for 4-side support, uniform pressure (wind) loading. This issue is discussed further in the companion paper in these proceedings.

Design Approach

As a guide to strength design with PVB-based laminates we have constructed charts that aid the selection of polymer type and thickness, glass thickness, for specified loading/support and rate/ temperature conditions. These charts have been computed using a procedure described in detail elsewhere [9]. Briefly, the procedure consists of: 1)establishing a constitutive model for the PVB (Butacite™) interlayer by dynamic mechanical analysis [10,11]; 2) carrying out finite element analyses of glass stress development [11]; 3) validating selected analyses against controlled loading experiments [9]; 4)combining stress analyses with a statistical (Weibull) glass breakage model [12-14] that incorporates a time-dependence for glass strength [15,16].

Design charts may then be constructed for specified glass breakage probability and laminate build of interest. Note that the approach of coupling large-deflection stress analysis with a Weibull breakage model and time dependent strength effects is the basis for the current ASTM design charts for monolithic glass[14].

An immediate question arises to the accuracy of finite element calculations in such complex laminate systems. Accuracy depends upon many factors, both physical, related to the accuracy of materials models

9

Page 10: Laminated Glass

developed and experimental boundary conditions imposed in a validation test, and numerical, related to details of the mesh sizeand element choice. We have performed many validation tests of our approach and have lookedat a range of shapes, sizes, support conditions and temperatures. One example of a validation experiment is shown in Figure 2.

In this experiment stress development was monitored directly at three locations on a laminate plate. The plate was loaded under uniform pressure and supported on four sides. Predictions from our finite element model of glass stress development are shown in Figure 2 along with measurements of maximum principal stress from strain gages. Agreement between model and experiment is below the measurement uncertainty of 5 %. We have seen a similar level of accuracy in all other validation tests carried out.

Figure 2. Glass stress development in a laminated plate at three locations, the plate was subjected to uniform pressure and supported on four sides. The curves are predictions of maximum principal stress from a finite element model.

An example of such a design chart computed for a 6 mm laminate is shown in Figure 3. This chart has been constructed in ASTM E1300 [17] format using the glass strength parameters specified in that standard [18]. In the USA 6 mm laminated glass is defined as 2.7 mm glass / 0.76 mm PVB / 2.7 mm glass. The chart plots allowable pressure loading contours as a function of plate dimensions for a glass breakage probability of 0.008. Note that the chart is constructed for a 3 s wind loading at 25°C. The chart contains all the common features of the established monolithic counterpart. For monolithic glass a series of such charts for common glass thickness is used in an iterative fashion to select the required glazing thickness.

10

Page 11: Laminated Glass

Figure 3. Proposed strength design chart for a 6 mm laminate. In the USA this is defined as: 2.7 mm glass / 0.76 mm PVB / 2.7 mm glass. Contours of allowable uniform pressure are plotted as function of plate size and shape. Note the probability of breakage (0.008), load duration (3s), temperature (25°C).

We have developed a set a charts that allows selection of laminate build for specified loading/support conditions, loading rate and temperature. Glass thickness, glass type, PVB thickness are variables in the laminate design and provide more options for construction of a laminate build to meet specified strength performance. These charts are currently being examined by the ASTM committee addressing glass strength. It is interesting to compare the strength design levels for a laminate with equivalent monolithic glass.

Figure 4. Comparison of strength behavior for 8 mm laminated glass to that of 8 mm monolithic glass. The design strength values computed for the laminate at 50°C have been divided by the strength values from an 8 mm monolithic chart. The figure plots these ratios for the same plate size/shape range in the design charts. Note that there is equivalence in strength between laminate and monolith over a wide range of plate sizes.

Figure 4 plots the strength ratio: (8 mm laminate strength/8 mm monolithic strength) for the range of plates sizes and shapes graphed in the standard ASTM chart. Note that this comparison has been made for strength properties at 50_C, which represents the upper use temperature deemed reasonable in high wind areas in desert climates in the USA. It can be seen that for a large range of plate sizes that the laminate and monolith are equivalent. Indeed under some conditions the laminate affords better strength performance and is probably due to a slight thickness benefit over the monolith and the fact that the laminate moves more rapidly to the membrane state. Slight reductions in strength are noted for the laminate at high aspect rations where bending stresses are expected to play a larger role in overall stress development. Such charts should encourage the engineer to take a more objective approach to laminated-glass design and not to use overly conservative penalty factors against laminated glass that are couched in the worst possible loading scenario.

Conclusion11

Page 12: Laminated Glass

We have argued that the design of laminated glass for strength performance is generally over conservative for large plates subjected to uniform loading and 4-side support. Both experimental and fracture analysis presented support this conclusion and is consistent with much previous work addressing this topic. We have proposed a method to generate design information for laminated glass that incorporates all the necessary features to treat laminate glass breakage. This design information is available in chart form and is being considered by various standards bodies.

References

1. Hooper, J.A., “On the Bending of Architectural Laminated Glass”, International Journal of Mechanical Sciences, 15 309-323 (1973).

2. Behr, R.A., Minor, J.E., Linden, M.P., Vallabhan, C.V.G., “Laminated Glass Units Under Uniform Lateral Pressure,” Journal of Structural Engineering, 111[5] 1037-50 (1985).

3. Behr, R.A., and Linden, M.P., “Load Duration and Interlayer Thickness Effects on Laminated Glass,”Journal of Structural Engineering, 112[6] 1141-53 (1986).

4. Vallabhan, C.V.G., Minor J.E., and Nagalla, S., “Stresses in Layered Glass Units and Monolithic Glass Plates,” Journal of Structural Engineering, 113[1] 36-43 (1987).

5. Das, Y.C. and Vallabhan, C.V.G., “A Mathematical Model for Nonlinear Stress Analysis of Sandwich Plate Units,” Mathematical Comput. Modelling, 11 713-19 (1988).

6. Behr, R.A., and Norville, H.S., “Structural Behavior of Architectural Laminated Glass,” Journal of StructuralEngineering, 119[1] 202-22 (1993).

7. Vallabhan, C.V.G., Das, Y.C., Magdi, M., Asik, M., Bailey, J.R., “Analysis of laminated glass units”, Journalof Structural Engineering, 119[5] 1572-1585 (1993).

8. Van Duser, A., Jagota, A., Bennison, S.J. “Analysis of Glass/Polyvinyl Butyral (Butacite®) Laminates Subjected to Uniform Pressure” Journal of Engineering Mechanics, ASCE, 125[4] 435-42 (1999).

9. Bennison S.J., Davies P.S., Jagota A., Van Duser A., Smith C.A., Foss R.V., “Structural Performance ofLaminated Safety Glass”, presented at Glass Tech Asia 2000, Singapore.

10. Ferry, J.D., Viscoelastic Properties of Polymers, 3rd edition, John Wiley & Sons, (1980).11. ABAQUS® version 5.8, (1998) Hibbit, Karlsson & Sorensen, Inc., Pawtucket, R.I. 02860-4847.12. Weibull, W., “A Statistical Distribution Function of wide Applicability,” J. Appl. Mech., 18 293 (1951).13.Davidge, R.W., “Mechanical Behavior of Ceramics,” chpt 9, p 132-39, Cambridge Solid State Science

Series, Cambridge 1979.14. Beason, W.L., Morgan, J.R., “Glass Failure Prediction Model”, Journal of Structural Engineering, 110 [2]

197- 212 (1984).15. Brown W.G., “A Practicable Formulation for the Strength of Glass and its Special Application to Large

Plates”, Publication number NRC 14372, National Research Council of Canada, Ottawa (1974).16. Reed D.A., Fuller E.R. (Jr.), “Glass Strength Degradation Under Fluctuating Loads”, Journal of Structural

Engineering, 111 [7] 1460-1467 (1984).17. ASTM Standard E 1300-97 “Determining Load Resistance of Glass in Buildings” in 1997 ASTM Annual

Book of Standards, American Society for Testing and Materials, West Conshohocken, PA.18. Norville, H.S., Minor, J.E., “Strength of Weathered Window Glass”, American Ceramic Society Bulletin, 64

[11] 1467-1470 (1985).

Strength Factor for Laminated Glass

12

Page 13: Laminated Glass

H. Scott NorvilleTexas Tech UniversityGlass Processing Days, 13–16 June ’99

Abstract

US building codes use type or strength factors to relate the design strength of various window glass types and constructions to that of monolithic annealed window glass. Currently, the highest strength factor for annealed architectural laminated glass under short term loading is 0.90. This strength factor holds despite voluminous test data and theoretical analyses that indicate laminated glass displays strength and behavior equivalent to that of monolithic window glass of the same type and nominal thickness. This paper discusses mechanisms in US design methodology that lead to conservatism in window glass design, reviews failure strength data for laminated glass, and presents new failure strength data for laminated glass. Consideration of the extensive test data coupled with a design methodology that results in glass with design strength in excess of design wind loading leads to the conclusion that the strength factor for laminated glass should be 1.0.

Introduction

Laminated architectural window glass (LAG) consists of two or more glass plies bonded together with elastomeric interlayers, usually polyvinyl butyral (PVB). In most applications, LAG consists of 2 glass plies having the same thickness bonded together with one PVB interlayer. The author, his associates, and others have reported destructive tests of LAG under uniform loading [1,2,3,4,5,6]. These tests indicate that LAG fractures at uniform loadings roughly equal to or larger than that at which monolithic glass, of the same glass type and having the same geometry, fractures. Investigations of layered plate and LAG behavior [3,5,7,8] support these experimental results.

The author serves on the ASTM task group that wrote and maintains Standard practice for determining the load resistance of glass in buildings: E1300-96 [9], which furnishes the basis for window glass design in the US. ASTM E1300- 96 [9] contains the most comprehensive glass design methodology developed to date. Its experimental basis lies in tests of weathered window glass [10,11]. Its theoretical basis comes from a finite difference stress analysis technique developed by Vallabhan [12] coupled with a statistical approach to theory of fracture that Beason [10] calibrated from the work of Brown [13]. ASTM E1300-96 [9] provides a scientifically rigorous and robust basis for window glass design.

ASTM E1300-96 [9] presents the designer with 12 non-factored load charts, one for each nominal thickness designation, that relate window glass geometry to design strength measured in terms of a 60-sec. duration uniform load associated with a probability of breakage of 8 lites per 1000. Figure 1 presents a non-factored load chart similar to that found in ASTM E1300-96 [9] for nominal 6-mm (1/ 4-in.) glass. ASTM E1300-96 [9] uses glass type or strength factors to relate the design strength of window glass types and constructions to the nonfactored load.

Laminated Glass Design Using ASTM E1300

Table 1 summarizes strength factors for monolithic window glass and LAG under loading having 60-sec. or shorter duration as presented in ASTM E1300-96 [9]. Table 1, and the remainder of this paper, uses the following abbreviations:

13

Page 14: Laminated Glass

Figure 1. Non-Factored Load Chart for 6-mm Glass.

“AN” denotes annealed, “HS” denotes heat strengthened, “FT” denotes fully tempered, “AR” denotes aspect ratio (ratio of long dimension to short dimension for a rectangular lite), and “b/t” denotes the flexibility ratio (short dimension of rectangular LAG divided by its nominal thickness).

Table 1. Selected Strength (Glass Type) Factors from ASTM E1300.

Perusal of Table 1 indicates that the strength factor for LAG with AR£2.0 and b/t >150 is 0.90 that for monolithic glass of the same type. For all other cases, AR > 2.0 or b/t £ 150, the strength factor for LAG is 0.75 that for monolithic glass of the same type.

LAG design consists of determining the nominal thickness for the dimensions of a fenestration for glazing that provides load resistance, i.e., design strength, in excess of a specified loading. To determine the load resistance of rectangular LAG under short duration loading having a particular nominal thickness, the designer uses the following procedure:

(1) Determine the non-factored load from the appropriate chart for the nominal thickness,(2) Calculate the AR by dividing the long dimension of the rectangular LAG by its short dimension,(3) Calculate the flexibility ratio, b/t, for the LAG,(4) Select the strength factor from Table 1 based upon AR, b/t, and glass type, and(5) Multiply the non-factored load by the strength factor to obtain the load resistance.

14

Page 15: Laminated Glass

If the load resistance from Step 5 falls below the specified design loading, the designer must try a larger nominal thickness or a different heat treatment. When he achieves a combination of thickness and glass type for which the load resistance exceeds the specified design loading, the LAG design is adequate. The designer achieves an optimum situation when a combination of LAG thickness and heat treatment provides load resistance equal to the specified loading.

Because window glass comes in discreet thicknesses, the designer will rarely, if ever, achieve this optimum situation. For a particular glass type, as the designer works towards achieving an adequate design one nominal thickness will provide load resistance below the specified load while the next larger thickness will provide load resistance higher than the specified loading. Almost always, the final combination of nominal thickness and glass type for LAG or any other glass construction will nearly always provide load resistance in excess of specified design loading. This fact constitutes one mechanism that provides conservatism in the window glass design process.

A second, less obvious, point in ASTM E1300– 96 [9] that leads to conservatism in design lies in the definition of the non-factored load, especially for AN monolithic glass or LAG fabricated with AN glass plies. Wind pressures in model building codes in the US derive from wind speeds most commonly taken from ASCE 7 [14]. ASCE 7 wind speeds are associated with 3-sec gusts. Principles from glass fracture theory [10,11,13] indicate that uniform pressure P acting for 60-sec causes the same damage to AN glass as a uniform pressure of 1.21P acting for 3-sec. A designer in the US who naively bases AN window glass or AN LAG design on ASCE 7 [14] without converting pressures associated with 3-sec gusts to equivalent 60-sec uniform loads selects a nominal thickness that provides a load resistance approximately 21% higher than necessary.

LAG Tests

References [1,4,5] report failure strengths of new symmetric LAG fabricated with PVB and compare LAG fracture strengths to fracture strengths for new monolithic glass of the same type and geometry. All references report LAG fracture strengths nearly equal to or in excess of monolithic glass having the same geometry and fabricated with the same window glass type. The load duration for all tests in these references ranged from 30-sec. to significantly longer durations. All these tests involved LAG having AR £ 2.0 and most were conducted at room temperature with some at higher temperatures.

Norville et al. [5] indicated that an increase in interlayer thickness leads to higher LAG strength. Haar and Norville [15] recently completed additional tests of LAG with a new PVB interlayer formulation from Solutia that further support the preceding statements. Table 2 provides data from these tests.

Table 2. Failure Strength Data from Haar and Norville.

All samples tested consisted of 20 or more individual specimens. An equivalent monolithic sample had specimens with the same rectangular dimensions and nominal thickness as the LAG sample. Loading rates for LAG and equivalent monolithic specimens were the same. Haar and Norville [15] conducted all tests at room temperature. Perusal of Table 2 indicates that mean LAG fracture strength uniformly and significantly exceeds fracture strength.

Discussion and Conclusions15

Page 16: Laminated Glass

Glass design methodologies for wind loading should result in glass thicknesses that provide sufficient strength to resist wind loading without fracture. ASTM E1300-96 [9] provides such a methodology. ASTM E1300-96 [9] presents a comprehensive design methodology that has a very strong scientific basis.

While it provides a strong and scientifically defensible design methodology, ASTM E1300-96 [9] fails to consider factors that lead to conservatism in the design process. ASTM E1300- 96 [9] also does not address completely the voluminous research, both theoretical and experimental, that indicates LAG strength is equivalent to or exceeds monolithic glass strength under short duration loading. At this time, ASTM E1300-96 [9] fails to explicitly consider that the wind pressures used in window glass design in the US derive from 3-sec. duration gusts rather than 60-sec duration gusts. Its approach thus results in overly conservative designs for LAG as well as AN monolithic glass.

In light of the voluminous test data, the author feels that the architects of model window glass design recommendations should use a strength factor of 1.0 for LAG under short term loading. A strength factor of 1.0 for LAG under short duration loading adequately reflects published experimental research and recent theoretical analyses.

References

[1] Minor J.E., and Reznik, P.L. (1990). “Failure strengths of laminated glass.” Journal of Structural Engineering, ASCE, Vol. 116, No. 4, pp. 1030–1039, April.

[2] Pilkington ACI (1971). “A practical and theoretical investigation into the strength of laminated glasses under uniformly distributed loading.” Lab. Rep. and Discussion, Pilkington ACI Operations Pty., Ltd., Dandenong, Australia.

[3] Behr, R.A., Minor, J.E. and Norville, H.S. (1993). “The structural behavior of architectural laminated glass,”Journal of Structural Engineering, ASCE, Vol. 119, No. 1, pp. 202–222, January.

[4] Norville, H.S., Bove, P.M., Sheridan, D., and Lawrence, S. (1993). “The strength of new heat treated window glass lites and laminated glass units,” Journal of Structural Engineering, ASCE, Vol. 119, No. 3, pp. 891– 901, March.

[5] Norville, H.S., King, K.W., and Swofford, J.L. (1998). “Behavior and strength of laminated glass,” Journal of Engineering Mechanics, Vol. 124, No. 1, pp. 46–53, January.

[6] Norville, H.S. (1990). “Breakage tests of Dupont laminated glass lites,” Glass Research and Testing Laboratory, Texas Tech University, Lubbock, TX, August, 101 pp.

[7] Duser, A.V., Jagota, A., and Bennison, S.J. (in press). “Analysis of glass/polyvinyl butyral (Butacite®) laminates subjected to uniform pressure,” Journal of Engineering Mechanics, ASCE.

[8] Vallabhan, C.V.G., Minor, J.E., and Nagalla, S. (1985). “Stresses in layered glass units and monolithic glassplates.” Journal of Structural Engineering., ASCE Vol. 111, No. 11, pp. 2416–2426, November.

[9] Standard practice for determining the load resistanceof glass in buildings: E1300-96. (1996), ASTM, W.Conshohocken, PA.

[10] Beason, W.L. (1980). “A failure prediction model for window glass,” Institute for Disaster Research, TexasTech University, Lubbock, TX, NTIS Accession No. PB81-148421.

[11] Norville, H.S. and Minor, J.E. (1985). “The strength of weathered window glass,” Bulletin of the American Ceramic Society, Vol. 64, No. 11, pp. 1467–1470, November.

[12] Vallabhan, C.V.G. (1983). “Iterative analysis of nonlinear glass plates,” Journal of Structural Engineering, ASCE, Vol. 109, No. 2, pp. 2416–2426.

[13] Brown, W.G. (1974). “A practicable formulation for the strength of glass and its special application to largeplates.” Publ. No. NRC 14372, Nat. Res. Council of Canada, Ottawa, Ontario, Canada.

[14] Minimum design loads for buildings and other structures: ANSI/ASCE 7-95, (1995), ASCE, Reston, VA.[15] Haar, D.W., and Norville, H.S. (in press). “The effects of interlayer formulation on laminated glass

strength,” Glass Research and Testing Laboratory, Texas Tech University, Lubbock, TX, 81 pp.

http://cat.inist.fr/?aModele=afficheN&cpsidt=2084530

16

Page 17: Laminated Glass

Affiliation(s) du ou des auteurs / Author(s) Affiliation(s)(1) Glass Res. and Testing Lab., Dept. of Civ. Engrg., Texas Tech Univ., Lubbock, TX 79409, ETATS-UNIS(2) Wilfred Baker Engrg., Inc., 8700 Crownhill, Ste. 310, San Antonio, TX 78209-1128, ETATS-UNISRésumé / Abstract

Published experimental data indicate that under most conditions laminated glass strength equals or exceeds the strength of monolithic glass of the same nominal thickness. To date, these experimental data exist without a theoretical model. This paper presents a theoretical, engineering mechanics model that accounts for factors that affect laminated glass behavior including temperature, thickness of the interlayer, and composition of the interlayer. It presents additional fracture strength data for laminated glass lites with a thicker interlayer than in previous tests. Both the theoretical model and the new fracture strength data indicate that laminated glass strength increases as interlayer thickness increases and that laminated glass strength decreases as temperature increases. Although an increase in temperature beyond 38°C (100°F) leads to a decrease in laminated glass strength, the theoretical model indicates that laminated glass possesses significantly more strength than layered glass, e.g., simply, two plies of glass with no shear transfer, at temperatures above 49°C (120°F).Revue / Journal Title

Journal of engineering mechanics   ISSN 0733-9399   CODEN JENMDT Source / Source

1998, vol. 124, no1, pp. 46-53 (25 ref.)Langue / Language

AnglaisEditeur / Publisher

American Society of Civil Engineers, Reston, VA, ETATS-UNIS  (1983) (Revue)Mots-clés anglais / English Keywords

Glass ; Laminated plate glass ; Sheet glass ; Mechanical properties ; Rupture strength ; Modeling ; Theoretical model ;Mots-clés français / French Keywords

Verre ; Verre feuilleté ; Verre vitre ; Propriété mécanique ; Résistance rupture ; Modélisation ; Modèle théorique ;Mots-clés espagnols / Spanish Keywords

Vidrio laminar ; Vidrio ; Resistencia ruptura ; Modelización ; Modelo teórico ;Localisation / Location

INIST-CNRS, Cote INIST : 572 A, 35400007723043.0070

http://www.saflex.com/pdf/en/Saflex_Structural_Guide.pdf

ENERGY EFFICIENT LAMINATED GLASS17

Page 18: Laminated Glass

CONSTRUCTIONS

Phillip S DaviesDu Pont (Australia) LtdPO Box 930, North Sydney NSW 2059, AustraliaGLASS PROCESSING DAYS, 13–15 Sept. ‘97

ABSTRACT

Laminated glass offers the potential for a vast range of options in regard to the energy performance and aesthetics of building glazing, while also furnishing enhanced safety, structural integrity, security and sound attenuation. Various laminated glass configurations can be used in a single glazing or insulating glass unit to achieve desired solar control and optical performance.

These include:• Laminated glass with coloured interlayers.• Laminated body tinted glass.• Laminated spectrally selective heat absorbing glass.• Laminated metallic coated glass.• Laminated Low-E glass.• Laminates combining the above products.

In this presentation the versatility of laminated glass in energy efficient building design is illustrated using examples of recent building projects. The application of Solar LAMTM computer software as an aid in project design is also demonstrated.

1. INTRODUCTION

Energy control in buildings is an important issue in view of the concerns about greenhouse gas emissions and global warming. As a result worldwide emphasis is increasingly being placed on design of energy efficient buildings. In the Asia Pacific region building regulations imposing energy requirements have been in force in Singapore since the early eighties. Similar regulations have recently been enacted in Hong Kong and Thailand.

In Germany a Heat- Conservation Ordinance was enacted for buildings in 1995, while in North America and Australasia there is considerable activity taking place on the development of energy ratings for windows.

Window design and the choice of glass are especially critical to a building is consumption of energy for lighting, heating and cooling, as well as to the health, well-being and productivity of a building ís occupants. The choice of the right glass for a building project is often a complicated decision involving many factors. There are many reasons, including energy performance, to select laminated glass over other glazing options for modern homes and buildings.

2. FEATURES AND BENEFITS OF LAMINATED GLASS

In addition to energy efficiency, laminated glass offers numerous features and benefits over other glazing options. These include:

[1] Safety - Human Impact

The PVB interlayer in laminated glass absorbs the energy of human impact, resisting penetration, and although the glass may break the glass fragments remain firmly bonded to the interlayer, minimizing the risk of injuries.

[2] Structural Integrity

In the event of glass breakage laminated glass remains intact in the façades of a building and there are no falling glass fragments. Wind and rain are prevented from entering the building.

18

Page 19: Laminated Glass

[3] Security

Laminated glass resists intrusion of burglars and vandals. If broken the laminated glass continues to provide a barrier until replaced at the owner’s convenience.

[4] Sound Reduction

Laminated glass is often selected for building projects on the basis of its improved sound attenuation compared to monolithic glass. The noise reduction performance of insulating glass units is substantially improved by the incorporation of laminated glass one or both lites of the unit.

[5] Screening of UV

The PVB interlayer in laminated glass can screen out over 99% of UV radiation, thus preventing the fading of carpets and furnishings caused by these damaging rays.

[6] Low Visible Distortion

Sharp reflected images are possible with curtain walls constructed of laminated annealed glass.

3. LAMINATED GLASS FOR SOLAR CONTROL

Various laminated glass configurations can be used to achieve low shading coefficients for solar heat gain reduction, either in a single glazing or in an insulating glass unit.

These include

• Laminated glass with coloured interlayers.• Laminated body tinted glass.• Laminated spectrally selective heat absorbing glass.• Laminated metallic coated glass.• Laminated low-E glass.• Laminated glass constructions combining the above products.

With the above scope of possible combinations, and with the available ranges of glass tints, interlayer colours and glass coatings, it can be seen that laminated glass offers the potential for a vast selection of options in regard to energy performance and aesthetics of building glazing.

The use of laminated solar control laminated glass in curtain walls and the faÁades of high-rise buildings has been a prominent feature of Australian cities for over 20 years. The first of Brisbane ís two AMP towers, constructed in 1976, is a 36-storey glass gold office building with a curtain wall of 8.38 mm SOLARSHIELD 220 Gold laminated annealed glass. The second tower, constructed in 1984, is identical in design with a curtain wall of COOLPANE S20 blue laminated annealed glass. The shading efficient for the glass in both curtain walls is 0.29. Thermal safety assessments, good design and good quality control of glass edges were factors that enabled annealed heat reflecting laminated glass to be used in both buildings for low visible distortion without risk of thermal stress fractures.

The sculptured glass curtain wall of the 40-level State Bank building in Sydney contains 18,000m2 of glass embracing 11,000 windows and glass spandrel panels. Laminated annealed glass with SOLARSHIELD S20 coating and bronze 0.38 mm interlayer was used for the spandrel panels and the outer lites of the doubled glazed vision units. A combination of two edged support structural glazing and annealed laminated glass provides thevisual appearance of a clean ribbon of glass running the full height of the building.

The curtain wall of 53-storey Bourke Place in Melbourne, the second-tallest building in Australia, incorporates 17,000 m2 of solar energy control annealed laminated glass with a SS22 coating as follows:

Lower levels of building:6 mm SUNCOOL SS22 on Clear/1.52 mm PVB/10 mm Clear Glass

19

Page 20: Laminated Glass

Upper levels of building:6 mm SUNCOOL SS22 on Clear/1.52 mm PVB/6 mm Clear Glass

The thicker laminated glass used in the lower levels of the building was specified to mitigate the noise from passing trams. A uniform appearance was achieved in the façade of the building by using 6 mm coated glass on the exterior lite of the laminated glass. The SolarLAMTM performance parameters of the 13.52 mm and 17.52 mm laminated glasses are as follows:

13.52 mm 17.52 mmVisible Light Transmittance: 20.9% 19.9%Shading Coefficient: 0.34 0.33Luminous Efficacy: 0.61 0.61U-value, watts/m2.K 5.67 5.69

* Luminous efficacy, also called Coolness Index, is the ratio of the visible light transmittance to the shading coefficient (Tvis/SC). The higher the number the better the glass filters the infra red from the solar spectrum while admitting visible light. This factor is a major guide to design relating to the degree of trade off between the reduction of the cooling load (SC) and the need (or not) for artificial light (Tvis).

4. LAMINATED GLASS WITH COLOURED INTERLAYERS

A variety of interlayer colours are available for solar control and building aesthetics - the darker the colour, the greater the solar absorbence. Coloured interlayers are often used in combination with metallic coatings. If the coating is on surface #3 of the laminate the observed colour of the interlayer is highlighted when viewed from the exterior of the building. In some cases, use of a coloured interlayer can result in less thermal stress than if a body tinted glass of similar colour was used in the laminate.

The Glen shopping centre in Melbourne features two large atrium domes and a gabled roof of 12.38 mm annealed laminated glass containing a SL20 coating with green interlayer. SolarLAMTM was used to estimate the performance parameters of the laminated glass. The laminated glass construction and performance parameters are:

Construction:6 mm SUNCOOL SL20 on Clear/0.38 mm green interlayer/6 mm clear glassPerformance Parameters:Visible Light Transmittance: 15%Shading Coefficient: 0.35Luminous Efficacy: 0.43U-value, W/m2K: 6.06

5. LAMINATED BODY TINTED GLASS

The use of laminated body tinted glass for solar energy control is exemplified in Malaysiaís PETRONAS Twin Towers, now the tallest buildings in the world. The curtain walls of the PETRONAS Twin Towers contain 56,000 m2 of 14.38 mm laminated annealed glass. Sunlight and heat were critical issues in the building design, and considerable time was spent refining the performance and aesthetics of the curtain wall in a full-scale exterior curtain wall mock-up. The design incorporates intelligent use of sunshades with laminated green heat absorbing glass. The construction and SolarLAMTM performance parameters of the laminated glass are as follows:

Laminated Glass Construction:6 mm Green Glass/0.38 mm B14 Clear PVB/ 8 mm Clear GlassPerformance Parameters: Visible Light Transmittance: 71%Shading Coefficient: 0.60Luminous Efficacy: 1.18

20

Page 21: Laminated Glass

U-value, W/m2K 5.89

6. LAMINATED SPECTRALLY SELECTIVE GLASS

In body tinted glasses reduction in solar transmittance is usually accompanied with reduced visible light transmittance. Within the last decade new body tinted glasses have been introduced which reduce unwanted solar heat gain and maximise desirable natural daylight transmittance. Examples are Pilkington LOF EverGreenTM and PPG AzurLite® glasses. A laminated glass construction combining a spectrally selective glass with a pyrolytic (hard coat) Low-E glass with the coating on laminate surface #4 has been termed a “Low-IR” laminated glass. An example is the 12.38 mm annealed laminated glass in the Revesby Workers Club in Sydney, which was installed on the clubís western and eastern elevations to reduce the building ís air conditioning load. This 12.38 mm annealed laminated glass has the following construction and SolarLAMTM performance parameters:

Construction:6 mm EverGreenTM/0.38 mm Clear PVB/ 6 mm K Glass

Performance Parameters:Visible Light Transmittance: 56%Shading Coefficient: 0.40Luminous Efficacy: 1.41U-value, W/m2K 4.01

7. LAMINATED METALLIC COATED GLASS

Metallic coated glasses control against solar heat gain by reflecting solar radiation as well as absorbing it. The most efficient metallic coatings for solar heat-gain reduction, while maximizing visible transmission are vacuum “sputtered” coatings. In laminated glass, with the coating on surface #2 or #3 of the laminate, these coatings are protected from possible damage that can occur during installation and window cleaning. There is a widening range of sputtered coatings, including coatings that are highly spectrally selective to provide significant solar heat gain reduction with high light transmission. The compatibility with PVB of specific metallic coatings must be checked prior to adoption of any coating in laminated glass.

The Hotel Nikko Hotel on Sydney’s DarlingHarbour is constructed over a freeway andutilizes 5,000 m2 of heat reflective laminatedwindow glass for solar energy and noise control.The 11.52 mm and 17.52 mm laminated glassincorporate TS30 coating with 1.52 mm PVB andclear annealed glass.

The SolarLAMTM performance parameters for these laminated glasses are:

11.52mm 17.52mmVisible Light Transmittance: 32.6% 30.4%Shading Coefficient: 0.44 0.41Luminous Efficacy: 0.74 0.74U-value, watts/m2.K 5.90 5.73

The atrium of a new corporate headquarters in Melbourne houses a flourishing rainforest. In addition to concern about solar heat gain the building designer needed the ideal level of natural lighting for rainforest growth. For the roof of the atrium an insulating glass unit was chosen in which the outboard lite is 12.76 mm Viracon VE laminated glass containing a spectrally selective coating. A further feature of this type of laminated glass construction is low visible reflectance, both exterior and interior. The IG unit construction and performance parameters are as follows:

Construction:

21

Page 22: Laminated Glass

12.76 mm VE1-2L #2 Laminate Heat-Strengthened Glass 13.2 mm Air Space6.76 mm Clear LaminatedHeat-Strengthened Glass

Performance Parameters:

Visible Light Transmittance 67%UV Transmittance (380 mm): <1%Visible Light Reflectance - Out 12%

- In 14%Shading Coefficient: 0.43U-value, W/m2K 2.44

8. LAMINATED LOW-E GLASS

Reduced U-value along with increased solar gain reduction can be achieved by incorporating a pyrolytic Low-E glass in laminated glass with the coating on surface # 4 of the laminate. Following is a comparison of the SolarLAMTM performance parameters of 6.38 mm clear glass with those of 6.38 mm laminated K Glass:

6.38mm 6.38mmClear K Glass

Visible Light Transmittance: 85.7% 79.8%Shading Coefficient: 0.91 0.79Luminous Efficacy: 0.94 1.01U-value, W/m2.K: 5.80 3.63

The energy specification for Hong Kong’s new airport at Chek Lap Kok was met by a single glazing of 17.5 mm laminated glass construction incorporating a grey tinted glass, a reflective coating and K Glass.

9. LAMINATED GLASS IN DOUBLE GLAZED UNITS

Solar controlling laminated glass can be used in double glazed units to achieve higher levels of solar and noise control than achievable in a single glazed configuration. For example, the glazing for the Qantas Airlines terminal extension in Sydney is a double glazed unit in which each lite is a high performance laminate. This glass has the following performance characteristics:

Visible Light Transmittance: 16%External Visible Reflectance: 20%Shading Coefficient: 0.16Luminous Efficacy: 1.0U-value, W/m2K: 2.0

10. THE SolarLAMTM PROGRAM

Because of the almost limitless possible laminated glass constructions, a computer program was needed to aid glazing design with laminated glass by enabling rapid estimation of the basic solar, thermal and optical performance of laminated glass constructions from the performance parameters of their component glasses, glass coatings and interlayers.

In response to this need, DuPont worked with Ignatius Calderone of Calderone & Associates, Glen Waverley, Australia to create a PC program called SolarLAMTM. The basic solar/optical properties calculated by SOLARLAM can be used as input to the Lawrence Berkeley Window 4.1 PC program to obtain the total window thermal performance indices. Alternatively, these can be imputted into a companion program to SolarLAMTM called SC_U_VAL to obtain the shading coefficient, U-value, solar heat gain coefficient and luminous efficacy for the laminated glass construction.

The program assumes that there is no reflection at the contacting surfaces between glass and interlayer (unless

22

Page 23: Laminated Glass

there is a reflective coating on that surface). Since the refractive index of the glass is very close to the refractive index of the interlayer, this does not introduce any significant errors.

The total transmittance of the laminated glass is calculated by taking the product of the transmittance of each component that is directly transmitted and adding the extra transmitted energy following the multiple internal reflections. Since the reflectance from an uncoated surface is only of the order of 4%, the internal reflections become negligible after the second reflections from an uncoated surface.

The outer surfaces, as well as any inner surfaces that have a reflective coating, are treated as individual components. The transmittance of the surface τs, is taken as 1 minus the reflectance, ρs, from the surface:

τs = 1 - ρs

Hence, for an uncoated outer glass surface, where the reflectance is 0.04, the transmittance is 0.96.

The transmittance of the glass and interlayer components is calculated for the appropriate thickness by using the extinction coefficient as follows. The transmittance of each component is the fraction, a, of each component available after absorption. This is also called the absorption coefficient:

a = e-kt

Where K is the extinction of coefficient and t is the thickness, or path length, through the component.

The total reflectance of the laminated glass I calculated by adding the reflectance of the outer surface, plus the product of the reflectance of the inner surface, and the transmittance of each component that the energy passes through, both as it approaches the reflective inner surface and as it leaves the surface. If a reflective coating isused the product of the reflectance of the coating and the transmittance of each component that the energy passes through, both as it approaches the reflective coating and as it leaves the coating, is also added.

SolarLAMTM Version 1.11 has the capability to incorporate only one glass coating. Also, it uses a “single band” approach for estimation of solar and optical properties. This approach is adequate for glazing systems comprised of layers whose properties do not change dramatically over the solar spectrum, however it introduces a noticeable error into results for systems with more than one spectrally selective layers. A new version of SOLARLAM is being developed to overcome these limitations.

11. THE FUTURE

Intense research is currently in progress around the world to develop advanced glazings with the aim of realising significant energy and environmental benefits. Numerous technologies are involved, including solar photovoltaic cells, electrochromic coatings, ceramic coatings and angle selective glazings, to name a few. The Japanese Government has announced that solar PV cells will be a major plank in its domestic effort to cut emissions. Even with today’s PV technology, a modern office block can be a standalone power station, generating its own electricity from solar faÁades, solar roof arrays and curtain walls.

12. SUMMARY

Laminated glass constructions offer numerous options for designers of energy-efficient buildings, while also providing part of the answer to modern day concerns about safety and security of people and property, and the wellbeing of building occupants. Safe, secure, noise reducing and energy saving, laminated glass not only provides solutions for today’s building needs, but will also be a key ingredient in advanced glazings, which will further improve the energy efficiency of buildings.

ACKNOWLEDGEMENTS

The author wishes to thank Ignatius Calderone for the explanation of the theory used in the SolarLAMTM Program provided in Section 8 of this paper. The author also wishes to express appreciation to many people from numerous companies who provided details of glazing constructions for various building projects used in this paper to illustrate the application of energy efficient laminated glass.

23

Page 24: Laminated Glass

Detection of Tensile Stresses Near Edges of Laminated and Tempered Glass

Alex S. Redner and Barbara R. HoffmanStrainoptic Technologies, Inc.Glass Processing Days, 18–21 June 2001

Abstract

In both tempered and laminated glass, manufacturing processes introduce residual stresses that develop during cooling, bending and lamination. Local surface tension can develop near the edges, making the glass vulnerable to cracks and failure. Methods of locating and measuring these tensile stresses are discussed and their effectiveness evaluated.

Introduction

Tempered, bent and laminated glass products develop residual stresses when subjected to a complex transient temperature gradient and bending. In tempered glass, the objective of the process is to develop a protecting layer of surface and edge compression. In bent and laminated products, surface compression is also present, although to a much lesser degree, and bending stresses are considerably higher. A well-known parabolic stress distribution typically develops as illustrated in Figure 1 below.

Figure 1. Stress distribution in thickness of tempered and annealed glass

In heat-strengthened and tempered products, minimum and maximum surface and edge compression are specified as shown in Table I.

The stress distribution near glass edges is considerably more complex than shown in the Figure 1 parabolic distribution, because the edge cooling rate differs from the central regions. Edge stresses are distributed as shown in Figure 2.

24

Page 25: Laminated Glass

Figure 2. Surface and mid-layer stress near edges

In annealed, bent and laminated glass, the major concern of the producer is to assure that there is no surface tension since a tensile stress could produce failure during installation and use. Automotive windshield glass is particularly susceptible to problems related to surface tension. The tensile surface stress will typically appear near glass edges, but can exist at any point. Surface tension can appear as result of high edge temper, non-uniformity of cooling rate, or bending stress introduced in forming. Quantitative analysis of the stresses that develop in tempering, bending and lamination were reported by Dr. S. T. Gulati [1] who strongly recommended stress measuring as a systematic and necessary part of quality control to eliminate potential of delayed fractures.

To assure freedom of surface tension, thorough process supervision is necessary and a good understanding of the way the tension develops is required. Necessarily, the ability to reliably measure surface and edge stress is required to identify the presence of tension and to permit prompt corrective action. Using modern technology, measurement of surface stress and edge stress is now feasible and economical.

Edge-Stresses

Surface stresses are related to the temperature gradient that results from cooling. Surface stresses are also related to bending in bent and laminated glass [1]. As shown in Figure 2, the edge of a glass plate is, in reality, a side-surface (E) with surface-compression comparable in size [2] to the surface (F) stress, assuming the cooling rate and the heat flow are approximately the same. Surface compression forces are balanced by tensions in the mid-plane. Superposition of the stress due to surface F cooling and edge E cooling results in a local increase of the mid-plane tension in the edge direction. Under certain conditions, the tensile stresses are pushed to the glass surface. In bent and laminated glass, bending introduces additional surface tension (Figure 2). Measuring the average membrane stress in transmitted light yields the results shown in Figure 3, indicating a peak average tensile stress at a small distance from the edge (typically 12-25 mm).

A large positive tensile average is an indication of an undesirable cooling rate and a potential problem. Dr. S. T. Gulati [1] indicated that a membrane tensile stress of 1000 psi (7 MPa) could lead to a delayed crack growth.

Strength and Deformation Behavior of Laminated Glass

25

Page 26: Laminated Glass

Dr. Tammy Amos & Dr. Stephen J. Bennison, Glass Laminating Solutions, E.I. DuPont de Nemours & Co. Inc. Wilmington, DE 19803 USAGLASS PROCESSING DAYS 2005

Abstract

After a decade of research, a clear picture of the strength and deformation behavior of laminated glass has now been established. In this contribution we show three methods for computing the stress and deflection behavior of laminates and compare the methods to validation experiments. Key to the accurate structural analysis of laminates is adequate characterization of the time-temperature nature of polymer interlayers. Particular emphasis is placed on how to treat time and temperature effects on strength and how different types of interlayer affect the performance of the laminate. Established and emerging standards for assessing the structural performance of laminated glass are discussed in context of the progress made in this subject. Availability of software packages to aid in the strength design of laminated glass is also discussed. The approaches presented enable efficient structural design of laminates allowing full performance potential to be realized.

Introduction

The strength performance, characterized by glass stress development and deformation behavior under a specified loading and support scenario is a primary design consideration that often dictates the final laminate construction in many applications. Traditionally, laminated glass strength has been analyzed for a worst-case loading/support condition [1], leading to much over-design of the product [2]. Over the past decade we (and others) have researched the structural behavior of laminated glass and have developed several methods of computing the strength and deformation response of laminates under varied loading/support scenarios and different polymer interlayers [3-9]. In this contribution, we illustrate three objective methods of computing the stress and deflection development and compare predictions of the methods to experiments.

Case Study – Biaxial Deformation of a PVB Laminate

Here we use results of one simple experiment to illustrate and validate the analysis methods presented. The experiment consists of loading a PVB laminate with a ring and supporting it on three balls to approximate a ring-on-ring loading condition [6]. This test is a very useful method for evaluating stress and deflection behavior of laminates and allows the experimenter to readily change variables, such as loading rate/duration, temperature, laminate build, interlayer type, etc. For the purpose of this study, we have evaluated glass stress development in a laminate by measuring directly the glass strain using a strain gage on the nominal tensile surface of the specimen, opposite the loading ring, as shown in Figure 1. The experiment consists of loading at various rates, recording the strain development, thence maximum principal glass stress assuming an equi-biaxial stress field at the glass surface. Figure 2 shows an example of the glass stress development, σm, as a function of applied load, F.

Analysis Methods

Method 1 – Finite Elements with Full Polymer Interlayer Model

This approach allows simulation of laminate deformation under complex loading/support scenarios and varying thermal/load duration histories. A three-dimensional model of the biaxial test shown has been analyzed. Figure 3 shows a typical finite element mesh used. The glass is modeled using 8-node brick elements with incompatible modes for accurate capture of bending. The PVB interlayer is essentially incompressible for typical temperatures, rates and is modeled using 8-node brick elements with incompatible modes and a hybrid formulation with pressure as an independent unknown. The finite element code ABAQUS® [10] has been used.

Soda-lime silica float glass is modeled as linear-elastic material with a Young’s modulus of 72 GPa and Poisson ratio of 0.22 [6]. The PVB interlayer is modeled as a linear viscoelastic material [6,7] with time and temperature effects related through the Williams-Landell-Ferry (WLF) equation [11]. The shear relaxation modulus, G(t), used in the viscoelastic description has been determined from dynamic mechanical analysis of Butacite® and the bulk modulus, K(t), has been determined using hydrostatic volumetric tests. The results of

26

Page 27: Laminated Glass

simulations of glass stress development as a function of loading are plotted as the solid line on Figure 2. The prediction line passes through the data points that record the glass stress development measured directly from the strain gage. Note this is not a “fit” to the data but a first principles prediction using the finite element method and knowledge of materials properties.

Many other validation experiments have shown that this approach is capable of accurately simulating deformation as function of temperature, rate and different geometry, such as pressure loading of a plate simply supported on all four sides (membrane-dominated stress state) [7,8]. Combining this stress analysis method with statistics of glass breakage [12,13] and static fatigue effects [14] provides the basis for the computation of the strength design charts for laminates contained in ASTM E1300-02 [15]. The approach described in this method is capable of simulating changes in thermal and time history of loading plus long-term creep behavior.

Method 2 – Finite Elements with Simple Polymer Interlayer Model

This approach also allows simulation of laminate deformation under complex loading/support scenarios but simplifies the time-temperature description of the interlayer. The method is still powerful in that complex geometries may be treated when the load duration and use temperature are well defined andconstant. We use the same model shown in Figure 3 but simply consider the polymer interlayer to behave with a fixed shear modulus, G, and Poisson ratio, ν. The question then becomes: what value of G and ν should be used in a simulation? Using our viscoelastic model established from measurements of shear and bulk elastic properties we can extract values of G and ν for different temperatures and load durations: Table 1. Note these values represent the end point states after relaxation at the temperature and load duration.Accordingly, to simulate the biaxial test we choose a G and ν value for the temperature of the experiment and the time duration. The simulated stress-force development using this approach is shown as the diamond symbols in Figure 2. The predictions closely match the observations.

Method 3 – Analytical Effective Thickness Approach Using prEN 13474

In this method we examine the use of an “effective thickness” approach to modeling laminates as proposed in prEN 13474-2. This developing standard proposes the use of analytic equations that determine the shear coupling between two glass plies through the interlayer [16]. The shear coupling depends on the interlayer shear stiffness, G, and glass properties and laminate geometry. To summarize there are two limits for the effective thickness: 1) thickness of one glass ply as the interlayer shear modulus approaches zero, i.e. “high” temperatures relative to polymer glass transition temperature, “long” load duration; and 2) total laminate thickness as the interlayer shear modulus attains a stiffness that allows efficient shear transfer, i.e. temperatures below the interlayer glass transition temperature, and “short” load duration. The effective thickness can then be used in standard solutions for stress analysis. The effective thickness of the biaxial test specimen was determined from the geometry shown in Figure 1 and the appropriate PVB shear modulus for the temperature and load duration of interest. This value was used in conjunction with standard formulae for estimating stress development in the biaxial test [17]. The calculations are plotted as square symbols in Figure 2. Again this approach yields good estimates of stress development during loading in this experiment. We have found from other studies that this approach yields acceptable predictions for the case of simply supported beams but does not work as well when treating plates supported on four sides (membranes). We believe the approach is accurate for cases where bending stresses dominate, such as one, two-sided support beams. The method is not recommended for the analysis of point loading and support conditions [18].

The Role of the Interlayer Stiffness Properties

Strength and deflection for a bending-dominated case are dependent on the modulus properties of the polymer interlayer. The effect of interlayer stiffness has been investigated by measuring stress development in beams using a three-point bend test. Figure 4 shows the results.

Here we have measured the load required to generate a specified glass stress and defined this load as the strength. For the cases studied, the load bearing capacity of a PVB laminate decreases slightly compared to a monolithic glass beam of equivalent thickness because the coupling leads to effective thickness behavior between the two limiting cases. However, when the PVB interlayer is replaced by a structural interlayer, SentryGlas® Plus, a stiff interlayer with an enhanced glass transition temperature, the strength increases substantially over the equivalent thickness monolithic glass. Laminates with this structural interlayer may be

27

Page 28: Laminated Glass

simulated using the properties in Table 2.

Enhanced structural performance can, therefore, be achieved with the use of the SentryGlas® Plus interlayer. Laminates are essentially stiffer and stronger than equivalent PVB laminates.

Available Software Tools

The approaches described provide comprehensive methods to analyze the deformation behavior of laminated glass. However, the engineer often is asked to analyze standard designs. Here we review two available software tools with simple user interfaces to aid the design process.

The first tool is “Window Glass Design 2004” and is based upon ASTM E1300-02 “Standard Practice for Determining Load Resistance of Glass in Buildings”. The Standards Design Group, Inc. (SDG) [19], produces the software and it is available at www.standardsdesign.com for download. The ASTM E1300-02 standard is one of the most developed for treating PVB laminated glass objectively. It contains design charts for determining the load resistance of rectangular panels supported on one, two, three or four sides subjected to uniform load. The charts were developed using the approach summarized in Method-1 above. The interface allows the user to essentially query arrays that contain values for every case treated in E1300-02 design charts. Furthermore, SDG has produced an enhanced version of the program with design charts for the structural interlayer laminates (Figure 5).

Figure 5 GUI for Window Glass Design 2004 software

28

Page 29: Laminated Glass

Window Glass Design 2004 allows rapid analysis of common loading and support scenarios for glass laminates used in facades and estimates performance at a temperature of 50ºC.

The second tool described here is The DuPont Strength of Glass Calculator™. It can be accessed at www.dupont.com/safetyglass/en and be used online through a browser. The tool allows the user to evaluate the glass stress and deflection of laminated glass for both 1- and 2-sided simple support conditions. Loading options include a line load or uniform load, which can be applied separately or in combination. The basis of the calculator is the “effective thickness” approach described in Method-3 above. The first window asks for a case as shown in Figure 6.

29