Highly public test results show that, in some cases, the average strength of newly manufactured enamel glass is lower than that of similar uncoated glass [1,2,3,4]. Therefore, some people [1,4] suggested that the strength reduction factor should be included in ASTM E1300 "Standard Practice for Determining the Load Resistance of Glass in Buildings" [5] for the design of enamel glass.

The test results provided here indicate that although the average strength of newly manufactured enamel glass appears to be reduced, the coefficient of variation is also significantly reduced. In this paper, it is shown that the reduction of the coefficient of variation is combined with the fact that ASTM E1300 is based on the fact that the strength of the glass in use is reduced to compensate for the significant reduction in the average strength of newly manufactured enameled glass.

These results show that, for the heat strengthened glass examined, it is not necessary to add the strength reduction factor of enameled glass to ASTM E1300. The behavior of enamel glass is complex, and a lot of additional research is required to justify the change to ASTM E1300 through technical considerations.

Recently, regarding the proposed changes to ASTM E1300, the uniform load resistance of enamel-coated glass panels has become an interesting topic. Of particular interest are reports that, compared to uncoated, newly manufactured glass of the same type, the breaking strength of enamel-coated glass is reduced [1,2,3,4].

Although it is clearly pointed out in the text of ASTM E1300 that the glass thickness selection criteria provided are based on the strength of the glass in use rather than the strength of the newly manufactured glass [5], it may affect the use of ASTM E1300 for the design of enamel coatings. Layer glass panels. This controversy is mainly limited to enamel-coated heat strengthened (HS) glass.

Much of the controversy surrounding the use of ASTM E1300 to specify the uniform load resistance of enamel-coated glass is apparently due to the lack of understanding that the glass thickness selection chart provided in ASTM E1300 is a result of internal performance. Serve glass, not freshly made glass.

The results of the strength test of the glass in use conducted by Beason [6] show that the strength of the glass in use is significantly lower than that of the newly manufactured glass. Since Beason conducted preliminary tests, additional tests have confirmed that the strength of glass in use is actually significantly lower than that of newly manufactured glass [7]. According to reports, the in-service strength is reduced by up to 50% compared with newly manufactured glass [7,8].

The fact that the strength of glass in use is significantly lower than that of newly manufactured glass was one of the main driving forces for the initial development of ASTM E1300. The "Importance and Use" section of ASTM E1300 clearly states that the use of this practice assumes that "the surface condition of glass is typical of glass that has been used for many years, and is due to slight wear on the exposed surface" [5].

Therefore, if you want to challenge the applicability of ASTM E1300 to enamel-coated glass, it must be based on the comparison of the strength of enamel-coated glass with the use strength of uncoated glass defined in ASTM E1300, and the strength of uncoated glass that is not newly manufactured. strength.

As mentioned above, most of the current disputes about the strength of ceramic enamel coated glass are focused on the performance of HS enamel coated glass. Therefore, it is very important to have a clear understanding of the ASTM E1300 treatment of HS glass. For most of the 20th century, the glass thickness selection criteria proposed by the United States included the strength factor of HS glass at 2.0. At that time, the glass design community was already very clear that the 2.0 HS strength factor was based on linear stress analysis and conservative assumptions that the residual surface compression level of thermally strengthened glass would meet the minimum required value of 3,500 psi.

When ASTM E1300 included the glass thickness program for HS glass, there was ample discussion and debate within ASTM to determine the best way to achieve this goal. As part of these discussions, the results of research conducted by Beason [9] are considered. These results include a glass thickness selection chart corresponding to the glass in use with a minimum residual surface compression of 3,500 psi.

The results show that the accepted HS strength factor of 2.0 provides a reliable lower limit for all glass size ranges provided in ASTM E1300 [9]. In addition, studies have shown that for many glass geometries, higher HS intensity factors can be justified [9]. In addition, the residual surface compression during the manufacture of almost all HS glass is significantly greater than the required minimum of 3,500 psi, so it is inevitable to conclude that the treatment of HS glass by ASTM E1300 is very conservative from a strength point of view- view.

Those involved in the inclusion of HS glass in ASTM E1300 fully understand this fact. When making statistical inferences about the probability of breakage (POB) of glass, the two most important factors are central tendency and dispersion. The central tendency is usually quantified by the mean, and the dispersion is related to the standard deviation. In the long history of glass design procedures, what is particularly important is the ratio of standard deviation to average.

This statistic is often called the coefficient of variation (COV), and for a particular type of glass, it tends to remain relatively constant regardless of the geometry of the test. Historically, glass designers conservatively assumed that the COV of annealed (AN) and HS glass was approximately 20-25% and 15%, respectively [10].

The design strength associated with the specified POB decreases as the average value of the strength data decreases, and the design strength of the specified POB increases as the COV decreases. The glass thickness selection standard proposed in ASTM E1300 is based on the glass failure prediction model (GFPM) developed by Beason [11].

People often mistakenly believe that GFPM is nothing more than the application of the well-known two-parameter Weibull probability distribution function [12] as an alternative to the normal distribution function, which is widely used in the long-standing glass design program in the United States. However, GFPM is actually based on the relatively vague statistical failure theory of brittle materials also developed by Weibull [13].

GFPM contains two parameters m and k, which represent the distribution and severity of defects on the surface of the glass plate. It must be emphasized that these two parameters represent the occurrence and severity of surface defects [11], not the inherent material properties of glass, and can be determined by testing small glass samples in a controlled laboratory environment.

It can be seen that the m surface defect parameter corresponds more directly to COV, while the k surface defect parameter corresponds more directly to the average value [14]. When ASTM E1300 was introduced, the value of the m surface defect parameter was set to 7 to reflect the historical understanding of the COV commonly used in AN glass design, and the value of k was adjusted to reflect the reduced average strength in service [14].

As mentioned above, these surface defect parameters represent the severity and distribution of surface defects, not the type of glass. Since there is almost no scientific reason to believe that HS glass will cause different types of surface damage in use than AN glass, it is assumed that the number of surface defects associated with AN in use is the same as serving HS glass.

Therefore, as long as the residual surface compression is properly handled, it is reasonable to use the in-service surface defect parameters developed for AN glass to simulate HS glass. If GFPM is properly extended to the treatment of HS glass, it can be shown that as the level of residual surface compression increases, the relevant COV of HS glass decreases in a manner consistent with the above-mentioned historical design understanding. The surface defect parameters included in ASTM E1300 represent the characteristics and distribution of surface defects, which are selected as typical in-service exposures during the ASTM consensus process [14].

The exact value chosen is not the result of curve fitting a particular data set. On the contrary, ASTM E1300 surface defect parameters reflect the centralized trend and dispersion of large amounts of data developed by the industry and public interest groups, and incorporate the collective judgment of people participating in the ASTM consensus process [14].

It is now well known that when ASTM E1300 is introduced, if the newly manufactured glass is tested and analyzed to obtain the best surface defect parameters, these values ​​will be very different from those used in ASTM E1300. Nevertheless, ASTM E1300 surface defect parameters have been fully discussed and determined through the ASTM consensus process, and a set of obviously conservative glass thickness recommendations have been provided, maintaining strong continuity with the historical glass design process in the United States [14].

As mentioned above, regarding the treatment of HS glass, ASTM E1300 has built-in multiple levels of conservativeness. The inherent conservativeness of HS glass in ASTM E1300 makes it extremely unlikely that the design strength of enamel-coated glass will be lower than the strength shown in ASTM E1300.

The most straightforward procedure to determine whether there is a problem with selecting the minimum thickness of enamel-coated HS glass using ASTM E1300 is to test a representative sample of a full-size enamel-coated HS glass plate and compare the determined design load with the corresponding design ASTM E1300 The load given in. ASTM E1300 either overestimates the strength of enamel-coated HS glass or underestimates the strength of enamel-coated HS glass.

Any other comparison procedure involves prediction and extrapolation. The first set of data discussed in this article was collected by the author. These data are obtained by subjecting two sets of 40 x 60 x 1/4 inch glass panels to continuous supports on the four sides of the glass panels to sustain linearly increasing uniform lateral loads, resulting in failure. The test setup is shown in Figure 1.

One set of samples involved uncoated HS glass, and the other set of samples involved enamel-coated HS glass with an average thickness of 0.224 and 0.226, respectively. The enamel coating on the second group of glasses is a full coating on the entire glass surface. All tested glass samples were taken from the same batch of fresh glass. The glass plate is tested with the coated side in tension.

The main test statistic of interest is the 3-second equivalent continuous breaking load of each glass plate. The standard understanding of "static fatigue" is used to convert the measured failure strength data into an equivalent 3-second duration [8]. Statistical analysis is used to determine the equivalent average 3-second failure load, related standard deviation and COV, and the equivalent 3-second continuous failure load, corresponding to a POB of 8 liters per 1,000 liters.

These analyses were performed on coated and uncoated samples. The calculation of the equivalent 3-second continuous fault load corresponding to every 1,000 8 lite POB is done using the standard normal distribution assumption. In addition, the residual surface compression of each sample was measured. These data are listed in Table 1 [12].

The information provided in Table 1 is very interesting. First, it can be observed that the average 3-second continuous equivalent failure load of the enamel-coated samples is about 43% smaller than that of the uncoated samples. This decrease in strength is comparable to the decrease in strength of the glass in use discussed above [7, 8].

The COV of the uncoated sample was determined to be 8.1%, and the COV of the coated sample was determined to be 2.9%. This is a huge difference. As mentioned above, for design purposes, the COV of HS glass has historically been assumed to be close to 15%. This means that the measured COV of enamel-coated HS glass is approximately 5 times lower than the COV that has been assumed in the historical design of uncoated HS glass. Very low COV is often an emerging feature of enamel-coated HS glass, supported by extensive testing [1,2,3,4].

As mentioned above, assuming that the data is normally distributed, statistical techniques are used to estimate the load corresponding to each 1,000 8 lite POB. The results show that the design load of uncoated HS glass is 403 psf, and the design load of enamel-coated glass is 288 psf. This means that the design load of enamel-coated HS glass is about 29% lower than that of uncoated glass.

Although this comparison is meaningful, it has nothing to do with the use of ASTM E1300 to specify the appropriate enamel-coated HS glass thickness. If you consult the glass thickness selection chart provided in ASTM E1300, you will find that for a 40 x 60 x 1/4 inch AN glass plate, the in-service design load corresponding to 8 liters/1,000 POB is about 55 psf [5]. Then, if a factor of 2.0 is applied to this value, as indicated for HS glass, it can be determined that the ASTM E1300 design load for a 40 x 60 x 1/4 inch HS glass plate is 110 psf [5].

If this information is combined with the data provided in Table 1, it can be shown that the measured design load of 40 x 60 x 1/4 inch HS glass is approximately 3.7 times the ASTM E1300 requirement. This is because of the inherent conservatism discussed above. In addition, it can be seen that although the enamel-coated HS glass is weaker than the newly manufactured transparent HS glass, it is still 2.4 times stronger than the strength required to comply with ASTM E1300.

It's hard to imagine how this could be considered a problem. The second set of full-size data discussed here involves a 38 x 76 x 1/4-inch glass panel with four continuous lateral supports that can withstand increasing lateral pressure until it fails [4]. These tests involve five different sets of samples. The first set of glass samples is uncoated HS glass.

The other four groups of glass samples have different enamel coating patterns. The second group has a dot pattern covering 40% of the surface, the third group has a linear array covering 50% of the surface, the fourth group has a hole array covering 60% of the surface, and the fifth group has 100% coverage. The uniform coating. Berger et al. presented complete details of samples and test methods elsewhere. [4].

Berg et al. The processing data provided includes the average equivalent 3-second breakage load, the relevant standard deviation and COV, the percentage of strength reduction compared to uncoated glass, and the equivalent 3-second continuous failure load corresponding to every 1,000 8 liters of POB. Coating samples. Berger et al. completed the calculation of the equivalent 3-second continuous failure load corresponding to 8 liters of POB per 1,000. By calculating unique m and k for each data set. These data are listed in Table 2. [4]

If you refer to the glass thickness selection chart provided in ASTM E1300, you will find that for a 38 x 76 x 1/4 inch AN glass plate, the in-service design load corresponding to 1,000 8 liters of POB is about 41.8 psf [5]. Then, if a factor of 2.0 is applied to this value, as indicated for HS glass, it can be determined that the ASTM E1300 design load of the 38 x 76 x 1/4 inch HS plate is approximately 83.6 psf [5].

This means that the design load of 38 x 76 x 1/4 inch uncoated HS glass is approximately 3.25 times that of ASTM E1300. In addition, it can be seen that although the enamel-coated HS glass sample is weaker than the newly manufactured uncoated glass, the reported design load ratio of the enamel-coated HS glass sample is compared with the design load required by ASTM E1300 The range is 2.02 to 2.58.

Therefore, in the worst case, the strength of enamel-coated HS glass is more than 2.0 times the strength required by ASTM E1300. Again, it is hard to imagine how this could be considered a problem.

The above are the full-scale test results of two independent sets of enamel-coated HS glass panels to evaluate their load resistance and compare these results with the information provided in ASTM E1300. The test involves a set of samples tested by the author and a set of samples tested by Berger et al. [4]. All the tested glasses were freshly manufactured HS glass. For the two sets of glass plates tested, it was determined that the average strength of the newly manufactured enamel-coated HS glass was 29% to 48.3% lower than the similar average strength of the uncoated HS glass.

Since the middle of the last century, all glass design procedures commonly used in the United States have incorporated central tendency and dispersion into the process to estimate design stress or design load to meet the specified POB. Generally, the design intensity corresponding to the specified POB, such as 8 lite per 1,000, is proportional to the average intensity and inversely proportional to the COV of the intensity data. Therefore, as the average strength decreases, the design strength decreases, and as the COV decreases, the design strength increases.

Regarding the performance of enamel-coated HS glass, an observation that needs to be emphasized, regardless of the manufacturer or test organization, the accompanying COV of enamel-coated HS glass seems to be significantly lower than that of uncoated HS glass, and much lower A hypothetical COV is traditionally designed for uncoated HS glass. This trend compensates to a certain extent for the decrease in the average strength of enamel-coated glass compared to uncoated glass.

The basic assumption contained in ASTM E1300 is that the strength of glass refers to the strength of the glass in use, not the strength of newly manufactured glass [5]. Therefore, the design strength of enamel-coated glass must be evaluated based on the design strength in use obtained by ASTM E1300 rather than the measured strength of newly manufactured glass.

The full-scale board test results provided here show that the measured design load range of the enamel-coated HS glass related to 8 liters/1,000 POB is about 2.0 to 2.73 times that of ASTM E1300. Therefore, it is concluded that the enamel-coated HS glass types studied by the two groups of independent researchers introduced here meet all the requirements of ASTM E1300, and the reduction factor is not guaranteed.

The only unresolved issue in the data provided here is the effect of in-use exposure on the long-term strength of enamel-coated HS glass. Based on the authors’ collective experience, it is clear that the in-use strength reduction contained in ASTM E1300 is mainly the result of accumulated mechanical damage during years of use. Such damage includes scratches, pits, abrasions, etc., which are caused by cleaning, human contact, wind-borne debris, etc. In addition, the author is not aware of any reliable evidence that other non-contact in-service exposures, such as ultraviolet radiation, will significantly reduce the strength of the glass.

Regardless of the reason for the decrease in the average failure strength of enamel-coated glass, it is obvious that once the enamel coating is applied, it will provide a protective barrier on the coated surface, which will make it more difficult for the glass surface to accumulate additional mechanical damage.

In addition, most enamel-coated HS glass is used for spandrel applications, where the coated surface is protected from mechanical exposure in use, which is believed to be the cause of the in-use strength reduction contained in ASTM E1300. Currently, there is no credible evidence that the surface of enamel-coated glass will experience an additional decrease in average strength with exposure in use.

Finally, before the type of enamel-coated glass panel discussed here does not comply with ASTM E1300, the strength of the enamel-coated HS glass must be reduced by another 50% due to exposure in use. This seems unrealistic. Based on the information provided here, it can be concluded at this point that there is no technically reasonable reason for the design of enamel-coated HS glass to withstand uniform lateral pressure in ASTM E1300 to include a strength reduction factor.

[1] Natividad, K., Fonseca, JC, Erickson, B., Morse, SM and Norville, HS (2016), "Bending test of heat-treated glass and ceramic frits", GlassCon Global, pp. 165-172. [2] Maniatis, I. and Elstner, M. (2016), "Investigation on the Mechanical Strength of Enamel Glass", Glass Structure. Eng., Springer, pp 277-288. [3] Elstner, M. and Maniatis, I. (2016), "Enamel Glass-New Research GlassCon Global, pp. 69-76. [4] Bergers, M., Natividad, K., Morse, SM, etc. (2016), "Comprehensive testing of thermally strengthened glass with ceramic frits", Glass structure. Eng., Springer, pp 261-276. [5] ASTM Standard E1300, (2012), "Determining the Standard Practice for Load Resistance", E 1300, ASTM International, West Conshohocken, Pennsylvania. [6] Beason, WL, (1980), "Window Glass Failure Prediction Model", Ph.D. Thesis, Texas Tech University, Lubbock. [7] Norville, SH and Minor, JE, (1985), "The Strength of Weathered Window Glass", Bulletin of the American Ceramic Society, Vol. 64, Issue 11, pp. 1467-1470. [8] Beason , WL and Lingnell, AW, (2000), "Emerging Uses of Window Glass", Emerging Materials for Civil Infrastructure—The Latest Technology, ASCE, Reston, Virginia, pp. 190-216. [9] Beason, WL , (1993), "Development of a failure prediction model for heat-treated glass", final report submitted to Cardinal IG, W. Lynn Beason, Ph.D., PE, Engineering Consultant, College Station, Texas.[10] ASTM Standard E997, (1984), "Standard Test Methods for the Performance of Glass Structures in Exterior Windows, Curtain Walls, and Doors Under the Influence of Destructive Methods of Uniform Static Load", E 997-84, Philadelphia, Pennsylvania. [11] Beason , WL and Morgan, JR, (1984), "Glass Failure Prediction Model", Journal of Structural Engineering, ASCE, Vol. 110, Issue 2, Pages 191-212. [12] Ostle, B. and Mensing, RW ( 1975), Research Statistics, 3rd Edition, Iowa State University Press, Ames, Iowa. [13] Weibull, W., (1939), "Statistical Theory of Material Strength, Hander NR 151, Stockholm Royal University of Technology. [14] Beason, WL and Norville, HS, (1989), "Development of a New Glass Thickness Selection Program", Proceedings of the 6th National Wind Engineering Conference, Volume 2, University of Houston, Houston, Texas.

Via Del Lavoro Snc 64023 Mosciano Sant'Angelo TE Italy

Building 8, SIRIM Complex, No. 1, Persiaran Dato' Menteri, Section 2, PO Box 7035, Darul Ehsan, 40700 Shah Alam Selangor Malaysia

No. 160, Yichuan Road, Jiaonan City, Qingdao City, Shandong Province, China 266000

Log in or register to post a comment