Challenging Glass 2

Challenging Glass 2

I

Conference on Architectural and Structural Applications of Glass

Faculty of Architecture, Delft University of Technology

May 2010

1

Bos, Louter, Veer (Eds.)

Challenging Glass 2

Challenging Glass 2 Conference on Architectural and Structural Applications of Glass Faculty of Architecture, Delft University of Technology vmw.bk.tudelft.nl/challengingglass [email protected]

Support Challenging Glass 2 is supported by the Delft Centre of Materials (DCMAT), the Research School Structural Engineering, CUR Bouw & Infra, Kenniscentrum Glas (KCG), and the Dutch group of lABSE.

Cover credits Background: Front Inc. Photo ribbon, from back left to front right: JCDA Inc. (2x), Front Inc., Pascal Richet (2x), Werner Sobek, JCDA Inc., Werner Sobek, JCDA Inc., Werner Sobek.

Editors Bos, Louter, Veer

Organizing Committee Freek Bos Christian Louter Fred Veer

Scientific Committee Jan Rots, chair (NL) Fred Veer, secretary (NL) Jan Belis (B) Paulo Cruz (PT) Ulrich Knaack (NL) Rob Nijsse (NL) Mauro Overend (UK) Tanguy Rouxel (F) Jens Schneider (D) Geralt Siebert (D) Holger Techen (D) Bemhard Weller (D)

ISBN 978-90-8570-524-6

Printed by Wöhrmann Print Service, Zutphen, the Netherlands

Copyriglit © with the authors. All rights reserved. No part of this publication may be reproduced in any fonn, by print, copy, or in any other way, without prior written pennission from the respective author(s). The editors and authors are not responsible for the use which might be made of the following information.

ii

Contents

Foreword

Dear reader.

Proudly, we present the proceedings of the second Challenging Glass Conference (CGC2). Obviously, the economic difficulties of recent years have not thwarted the development of glass as an architectural and structural material. Almost 70 papers from around the globe show that efforts in both research and practice to fiilly utilize the possibilities of this material are relentless.

These proceedings stait with four key-note papers. Three of them present the latest advances in architectural design, fafade and structural engineering with glass. The fourth takes us outside the usual scope of the building industry, right into the heait of the material and its propeities.

The session papers cover a wide range of subjects related to the use of glass in buildings, from the fundamental questions of glass strength to the practical problems of construction. The authors constitute a mix of established names and young professionals. We trust this wi l l spark lively discussion and an exchange of ideas that wil l help the progression of glass in buildings in the coming years. We have done our utmost to provide the right setting.

Challenging Glass 2 owes its success to all its contributors. We would like to express our gratitude to the key-note speakers as well as to the other presenters and authors. Furthermore, we thank the Scientific and Organizing Committees, the supporting organizations, the advertisers and, of f course, all participants.

Welcoine to Challenging Glass!

Wishing you an inspiring and enjoyable conference,

Freek Bos, Christian Louter Organizing Committee

Jan Rots, Fred Veer Chairman and Secretary of the Scientific Committee

iii

Challenging Glass 2

7. References [ I ] Ngo, T.; Gupra, A.; Ramsay, J., Blast Loading and Blast Effects on Structures. Electronic Journal of

Structural Engineering (Special Issue: Loading on Structures), 2007, pp. 76-91

[2] Beveridge, A., Foreni;c/;7vei//ga/;'o« q/'iLvp/o^/on^, Taylor and Francis, London, 1998 [3] Baker, W., Explosions in Air, University of Texas Press, Austin, 1973 [4] Scholze, H. , Glass: Nature, Structure and Properties, Springer - Verlag, New York, 1991 [5] Beason, L.; Morgan, J., Glass Failure Prediction Model, ASCE Journal of Structural Engineering, 110

(2), 2007, pp. 197-212 [6] Amstock, J., Glass in Construction, McGraw Hi l l , New York, 1997

[7] Vallabhan, C ; Das, Y.; Ramasamudra, M. , Properties of PVB Interlayer Used in Laminated Glass, Journal of Materials in Civil Engineering, 4 (1), 1992, pp. 71-76

[8] Du Bois, P.; Kolling, S.; Fassnacht, W., Modelling of Safety Glass for Crash Simulation, Computational Material Science, 28, 2003, pp. 675-683

[9] Wei, J.; Dharani, L. , Fracture Mechanics of Laminated Glass Subject to Blast Loading, Theoretical and Applied Fracture Mechanics, 44, 2005, pp. 157-167

[10] Vallabhan, C , Iterative Analysis of Glass Plates, Journal of Structural Engineering, 109 (2), 1983, pp. 489-502

[ I I ] Vallabhan, C ; Das, Y.; Magdi, M . ; Asik, M . ; Baily, J., Analysis of Laminated Glass Units, Journal of Structural Engineering, 1 19 (5), 1993, pp. 1572-1585

[12] Wei, J.; Shetty, M . ; Dharani, L., Failure Analysis of Architectural Glazing Subjected lo Blast Loading, Engineering Failure Analysis, 13,2006, pp. 1029-1043

[13] Wei, J.; Shetty, M . ; Dharani, L. , Stress Characteristics of a Laminated Architectural Glazing Subjected to Blast Loading, Computers and Structures, 84, 2006, pp. 699-707

[14] Weggel, D.; Zapata, B.; Kiefer, M. , Properties and Dynamic Behavior of Glass Curtain Walls with Split Screw Spline Mullions, Journal of Structural Engineering, 133 (10), 2007, pp. 1415-1425

[15] Weggel, D.; Zapata, B., Laminated Glass Curtain Walls and Laminated G/O.M Lites Subjected to Low-Level Blast Loading, Journal of Structural Engineering, 134 (3), 2008, pp. 466-477.

[16] GSA, TS0F2003 Standard Test Method for Glazing and Window Systems Subject to Dynamic Overpressure Loadings, General Services Administration, Washington, 2003

[17] DoD, Structures lo Resist the Effects of Accidental Explosions. UFC 3-340-02, Unified Facilities Criteria, U.S. Army Corps of Engineers, Department of Defence, United States of America, 2008

[18] ASTM, ASTM F 2248-03 Standard Practice for Specifying an Equivalent 3-Second Duration Design Loading for Blast Resistant Glazing Fabricated with Laminated Glass, ASTM International, West Conshohocken, 2003

[19] Zienkiewicz, O.; Taylor, R., Finite Element Method for Solid and Structural Mechanics (6''' Edition), Elsevier, 2005

[20] Pica, A.; Wood, R.; Hinton, E., Finite Element Analysis of Geometrically Nonlinear Plate Behaviour, Computers and Structures, 11, 1980, pp. 203-215

[21 ] Reddy, J., Mechanics of Laminated Composite Plates and Shells, CRC Press, Boca Raton, USA, 2004 [22] Pica, A.; Hinton, E., Efficient Transient Dynamic Plate Bending Analysis with Mindlin Elements,

Earthquake Engineering and Structural Dynainics, 9, 1980, pp. 23-31 [23] Vallabhan, C ; Asik, M , ; Kandil, K., Analysis of Structural Glazing Systems, Computers and Structures,

65 (2), 1997, pp. 231-239 [24] McLellan, G.; Shand, E., Glass Engineering Handbook, McGraw H i l l , New York, 1984 [25] Seica, M . ; Krynski, M . ; Packer, J.A., Explicit Dynamic Modelling of Architectural Glazing Subject to

Blast Loading, Proceedings of ISIEMS13, Cologne, Germany, 2009 [26] SJ Software, SJ Mepla v 3.0, Aachen, Germany, 2007 [27] Krynski M . , Dynamic Response of Architectural Glazing Subject lo Blast Loading, M.A.Sc. Thesis,

University of Toronto, Toronto, Canada, 2008 [28] Nicholas, T., Tesile Testing of Materials at High Rales of Strain, Experimental Mechanics, 1980, pp.

177-185

384

^ Challenging Glass 2 - Conference on Archileclural and Structural Applications of Glass, TUDelft Bos, Louter, Veer (Eds.). TUDelft, May 20IO

Copyright © with the authors. All rights reserved.

Corrosion effects on soda lime glass

Fred Veer Faculty of Architecture, Delft University of Technology

[email protected], www.glass.bk.tiidelft.nl Yurii Rodichev

70.5". Pisarenko Institute for Problems of Strength ofNASU, Kiev, Ukraine, rym@ipp. Iciev. ua

Although soda lime glass is the most common used transparent material in architecture, little is known about the corrosion effects on long term strength and the interaction between corrosion and defects. Extensive testing on soda lime bars under different environmental conditions and different degrees of damage has resulted in a inore clear picture of the stress-conosion luechanisms involved. The effects of these on long tenn strength are discussed.

Keywords: Glass strength, stress corrosion

1. Introduction Soda lime glass is coinmonly used as it is a durable material. It is however susceptible to stress con-osion. A review of this is given by Haldimann et al. in [1,2]. Although there is considerable previous research, such as [3,4], there are still inany questions. One of them i f only the pH of the water is critical.

A ftindamental problem is the complex series of flaws that exist in cut and cut, ground and polished float glass. These significantly complicate the analysis of the results. Some of this is covered by Veer et al. in [5,6]. I f it is difficult to determine the basic strength, determining the added corrosion product is an added difficulty.

To avoid some of these problems it was decided to use Schott AR glass rods. These have the same chemical composition as fioat glass, but as there are not cut, ground and polished; the results from these tests should be more easy to interpret. Initial results by Veer and Rodichev are given in [7]. These initial results gave some indications about the corrosion mechanism but as the scatter in test data was still significant additional tests were deemed necessary. This includes tests on glass bars with quantifiable damage created using a diamond indenter.

2. Methodology Standard Schott Ar glass rods are cut down in to 250 mm long segments. These are tested in four point bending in a custom made rig on a Zwick ZlOO universal testing machine equipped with climate chamber. Distance between the bottom supports was 200 mm, distance between the loading rollers was 100 mm. Test speeds of 50 mm/min and 0.5 mm/min were used. A l l specimens were conditioned for the environment where

385

Challenging Glass 2

Some specimens were indented using a Knoop indenter with the length axis of the indenter perpendicular to the length axis of the specimen. The Knoop indenter was mounted in a Zwick zlO. The loading pattern is given in figure 2, and consists of four steps of 50,100,150 and 200 N respectively. Each held for 30 seconds with slow loading and unloading. This was done to create beach marker trails on the fracture surface to better study the initiation and growth of the crack. After indenting the specimens were kept at room temperature for one week before the four point bending tests.

oell

y-

320.527

Figure 2: Load sequence for glass indentation.

386


One series of specimens was pre-con-oded (water strengthened) by corroding them in water at 80°C for 24 hours. The specimens were then dried, left to lie m air foi 24 hours and 4 point bend tested in air at room temperature.

Fracture surfaces were examined using optical microscopy techniques. The results of this wi l l be published later due to space constraints.

3. Results There are four sets of resuhs. The first deals with one set of experiments comparing tes in a r with tests ,n demmeralised water. The second set compares tests m air with

n salt water and tests on pre-coiToded (water soaked) specimens. The third set eompares tests on undamaged bars in air with tests on bars with indenter dattrage m a,^ ThT fourth set compares tests on undamaged bars m water with tests on bars with indenter damage in water.

Table 1: Tests in air and demineralised water.

Specimen number Air fast (MPa) Air slow (MPa) Demineralised

water fast (MPa)

Deiuineralised water slow (MPa)

1 141.0 61.9 117.5 70.8

2 139.9 125.7 107.9 90.1

3 164.1 81.2 80.5 86.9

4 129.2 107.9 116.4 82.6

5 128.2 83.3 113.9 79.0

6 112.1 110.7 87.9 100.1

7 113.6 95.0 103.2 46.6

8 90.1 115.0 73.0 71.9

9 68.4 109.3 92.2 98.3

10 90.4 64.1 121.0 75.1

Average 117.7 95.4 101.4 80.1

Standard deviation /average

24.3% 23.0% 16.7% 19.5%

387

Challenging Glass 2

Table 2: Tests in air, salt water and pre-corroded specimens

Specimen Air fast (MPa) number

Air slow (MPa)

Precorroded slow (MPa)

Precorroded fast (MPa)

Salt water slow (MPa)

1 118.9 91.1 117.1 81,2 77,3

2 84.7 107.2 88.6 113,6 89,7

3 115.0 62.7 94.7 89,0 62,3

4 91.5 76.2 100,4 104,3 57,7

5 94.7 75.1 66,2 115,3 94,0

6 157.4 118.5 90,8 150,6 68,7

7 122.5 82.2 101.5 119,3 66,2

8 108.6 82.6 74,8 80,8 63,0

9 98.6 110.4 91.5 91,1 69,4

10 131.7 66.2 104.0 60,9 53,4

Average 112.4 87.2 93.0 100,6 70,2

Standard deviation /

average 19.4% 21.9% 15,7% 25,3% 18,8%

Table 3: Comparison of undamaged and indented bars in air

Specimen number Air fast (MPa) Air slow (MPa) Indented fast

(MPa) Indented slow

(MPa)

1 95.7 98.2 37,6 31,4

2 119.0 110.8 31,4 27,1

3 94.0 89.9 26,5 27,2

4 96.1 84.0 27,9 33,2

5 110.7 75,7 30,6 23,7

6 115.6 95.7 35,5 28,8

7 112.3 95.9 29,7 26,2

8 118.1 84.4 25,4 26,7

9 104.4 91.5 30,5 29,4

10 95.7 115.2 28,7 32.5

Average 106.2 94.1 30,4 28,6

Standard deviation /average

9.5% 12.8% 12,5% 10,6%

388


Table 4: Comparison of undamaged and indented bars in water at 20°C

Speciinen number Water fast (MPa) Water slow (MPa) Indented fast in

water (MPa) Indented slow in

water (MPa)

1 77,4 67,8 30,2 30,2

2 85,3 76,5 25,2 20,1

3 78,2 65,7 26,8 18,6

4 77,8 71.1 27,4 21,3

5 84,9 98,2 31,4 20,9

Average 80,7 75,9 28,2 22,2

Standard deviation / qvp.rape

5.0% 17,3% 9,0% 20,6%

Table 5: summary of results

Test series Mean failure stress Standard deviation/mean Number of tests

Air fast 1 117.7 24.3% 10

Air fast 2 112.4 19,4% 10

Air fast 3 106,2 9.5% 10

Air slow 1 95,4 23,0% 10

Air slow 2 87,2 21.9% 10

Air slow 3 94.1 12,8% 10

Demi water fast 101.4 16,7% 10

Demi water slow 80.1 19,5% 10

Salt water slow 70.2 18,8% 10

Pre-corroded fast 100.6 25,3% 10

Pre-corroded slow 93,0 15.7% 10

Air indented fast 30,4 12,5% 10

Air indented slow 28,6 10.6% 10

Water fast 80,7 5.0% 5

Water slow 75,9 17.3% 5

Water indented fas ;t 28,2 9,0% 5

Water indented slo w 22,2 20,6% 5

4. Discussion The resuhs are summarized in table 5. I f we look at the results it becoines obvious that the three fast and slow series in air, which were done with about a month between each successive series due to limited machine availability, do not coincide. Figure 3 shows a Weibull plots for the three fast series separately. In figure 4 the data is combined to give a single Weibull plot. There is no clear reason for the differences. It does make it clear

389

Challenging Glass 2

that considerable care must be taken in comparing data from different time periods. As all glass specimens were prepared beforehand and all specimens were cut in a single session an aging phenomenon could be responsible but there is no logical basis for this and this is discounter for now.

The data does however give a lot of useftil information. Figure 5 shows Weibull plots for the slow tests in sah and demineralised water. The salt water is clearly more con-osive. Normal water seems to be in the middle in terms of average bending strength as is seen in table 5. Thus not only the temperature and pH of the water are important, the electrical conductivity also plays a role.

Figure 6 shows Weibull plots of fast tests in air for indented and normal specimens. Figure 7 shows the same for the slow tests in air. The strength of the indented bars corresponds with the strength of 10 inm thick annealed float glass which has been cut, ground and poHshed, [5].

The indentations apparently do not improve the predictability of the strength. Although the standard deviations are lower, the Weibull plots show non-linearity comparable to the non-indented bars. The indentations however decrease the strength considerably. The specimens also break into only two pieces, while non indented specimens usually produce four to six fragments. This is to be expected as the fracture energy is much lower. The results might be taken to suggest that indentation damage reduces stress corrosion susceptibility in air. The difference in average bending strength between the fast and slow tests on normal specimens is some 20%, while for the indented specimens this is only some 6%. This neglects the fact that indented specimens fail at much lower stress levels and thus much faster in constant displacement tests and there is thus less time for corrosion to take place. The indented specimens in water show a more significant strength loss than the normal specimens. Presumably immersion in water allows for much more rapid corrosion of the already severely damaged specimens as the corrosive agent is readily available. Supply of the corrosive agent is thus a critical determining factor in the stress corrosion of glass.

Soaking glass in water is commonly assumed to increase the strength. The pre-corrosion that takes place is supposed to make the "cracks" less sharp and thus lower the stress concentrations. Figures 9 and 10 show Weibull graphs for the pre-corroded and normal specimens. No increase in strength due to pre-corrosion is visible, i f anything a decrease is found. This implies that damage only occurs when the specimens are stressed while exposed to a corrosive environment.

A last point deals with reliability. Indenting the specimens should in theory give more predictable results as the specimens fail from a similar macro-flaw. Figure 11 shows the tests data for the fast and slow tests in indented specimens in air. Figure 12 gives a micrograph of an indentation. The indentation is clearly not regular or smooth. Some small geometrical differences might cause deviations. Some increase in Weibull linearity is observed compared to the normal specimens in figures 3 and 4, but there is still no clear single Weibull line. This implies that even after indentation there might be some differences in failure. Fractographic analysis might give some answers, [8]. One answer might be that the bars are less homogeneous than float glass and thus contain other sources of failure besides surface damage.

390


Long term strength of glass is clearly dependent on a number of variables. Even the strength of glass without macro damage is decreased due to corrosion in air. Direct exposure to water wi l l accelerate the process. Although heavily damaged glass in air does not seem to show rapid strength loss due to corrosion, direct exposure to water wi l l cause more rapid strength loss. As the structure and thus the strength of the regular glass bars is similar to the strength of the surface of float glass this implies that regular float glass that suffers surface damage in services wi l l also degrade in strength especially when regularly exposed to (salt) water. Strength values for the surface of less than 20 MPa are possible.

Figure 3: Weibull plots of the three series of fast air tests.

i

Figure 5: Weibull plot of slow tests in salt and demineralised water.

Figure 7: Weibull plot of slow tests in air on normal and indented specimens.

Figure 4: Weibull plot of all fast air tests in a single series.

Figure 6: Weibull plot of fast tests in air on normal and indented specimens.

Figure 8: Weibull plot of fast test in water on normal and indented specimens.

391

Challenging Glass 2

Figure 9: Weibull plot of results of slow tests in air Figure 10: Weibull plot of results of fast tests in air on nonnal and pre-corroded specimens. on normal and pre-corroded specimens.

Figure 11: Weibull plot of slow and fast tests on Figure 12: Micrograph of indentation, indented specimens.

5. Conclusions From the resuhs it is concluded that:

• corrosion of glass is faster in salt water and nonnal water than in demineralised water.

• the electrical conductivity of the water plays a role in the coiTosion mechanism • there is no evidence that water soaking increases the strength of glass • indenting the glass severely reduces the strength, but does not increase the

predictability of the strength • indented specimens tested in air have less strength loss due to increased

corrosion than non-indented specimens. Presumably the fast fracture does not allow for much corrosion in air.

• tested indented specimens under water suggests there is more corrosion damage in slow bending than for non-indented glass. This suggests that in indented glass the stress corrosion mechanism is reagent supply driven.

392


6. References [1] M . Haldimann, A. Luible, M . Overend, Structural use of glass, Structural engmeer.ng documents 10

r21 M Hafdi^üann, Fracture strength of structural; glass elements- analytical and numerical modelling , testing and design, PhD thesis Ecole Polythechnique federale de Lausanne, 2006

[3] A Fink, Bin beitrag zum einsatz von floatglass als dauerhaft tragender Konstrukt.onswerkstoff m PhD thesis Technische HochschuIeDanustad 2000 , „f the American

[4] ^.U.msAtrhom,L}\.Bo\z, Stress corrosion and static fatigue of glass,iaama\ ot the American

ceramic society , vol 53, p 543, 1970 151 F A Veer Strength ofglass.a non transparent value, Heron, vol 52, no 1, iUU / 6 F A Veer., Y . M . Rodichev, ne strength of glass, hidden damage. Proceedings thts conference

[7] FA.Veer, Y . M . Rodichev, Glass failure, sciencefiction, science fact and hypothesis. Proceedings GPD

[8] Y ^ M ' S c h l v . F.A.Veer, Fractography of indented gte...Proceedings glass processing days 2011, Tampere Finland.

393

Documents

Challenging Glass 2