Abstract
In recent decades, the use of ethylene-tetrafluoroethylene (ETFE) membranes has been increasing in the building industry for lightweight transparent structures. The nonlinear behavior of this material has garnered significant attention from researchers in recent years. However, there remains uncertainty regarding whether the material is orthotropic or isotropic, leading to conflicting conclusions in the literature. This work aims to clarify this question.
To achieve this, uniaxial tensile tests were conducted on NOWOFLON ET 6235 foil at temperatures ranging from 16 °C to 32 °C, covering material orientations from –90° to 90° relative to the machine direction in increments of 15°. The results indicate that there is no significant difference in behavior based on cut orientation at small strains. However, a slight orthotropic behavior becomes evident after the initial inflection point on the engineering stress-strain curve, as confirmed by ANOVA analysis.
1 INTRODUCTION
The use of ethylene-tetrafluoroethylene (ETFE) has grown increasingly popular over the past few decades. Currently, much of the research focuses on the nonlinear behavior of this material. This paper aims to address the uncertainty surrounding whether ETFE behaves as an orthotropic or isotropic material, as existing studies present conflicting conclusions on this issue.
ETFE is a polymer composed of a partially fluorinated, semicrystalline copolymer produced through a polymerization process. Following synthesis, extrusion and blow molding technologies are utilized to form the thin foil. This manufacturing process leads to orthotropic behavior in the resulting foils typical of polymers (Ward 2016), as the extrusion direction promotes molecular alignment of the polymer chains, and the blow technology affect it on the second step. Generally, a measurable difference is observed between the machine direction (MD) and the transverse direction (TD) in the mechanical behavior; however, research indicates that this difference is not significant (Barthel et al. 2003, 72–75; Galliot–Luchsinger 2011, 356–365). On the other hand, De Focatiis and Gubler (De Focatiis–Gubler 2013, 1423–1435) identified orthotropic behavior using Cauchy stress, while Surholt et al. noted orthotropy beyond the yield point of the stress-strain curve (Surholt et al. 2022, 3156). Given that a complex viscoelasto-plastic model can be cost-prohibitive, it would be advantageous to verify whether an isotropic approach is adequate for ETFE.
As noted by Petraccone et al. (Petraccone et al. 1992, 22–26), X-ray diffraction tests reveal that below 0 °C, the ETFE polymer exhibits an orthorhombic crystal structure. However, as the temperature increases to 100 °C, the crystal structure gradually transitions to a hexagonal phase. Nye (Nye1985) introduced the relationship between a material’s stiffness matrix and its crystal structure. Consequently, the transformation from an orthorhombic to a hexagonal crystal structure is reflected in the shift of the stiffness matrix from an anisotropic condition to an orthotropic one, as demonstrated in Nye’s work.
The objective of this study is to thoroughly investigate the isotropic and anisotropic mechanical behavior of ETFE foils. The direction-dependent behavior was analyzed using statistical methods, taking into account the variations induced by factors altered during the experiments. While there is a wide range of methods available for analyzing relationships between responses and variables (Montgomery, 2012), this study specifically employs one-way factorial Analysis of Variance (ANOVA).
2 TEST PROCEDURES
The NOWOFLON ET 6235 foil was used for the measurements (‘NOWOFLON ET Architecture’). The samples had a width of 50 mm and a length of 300 mm (Fig. 1a). The nominal thickness of the foil was 250 μm, the weight was 435 g/m2. The average measured thickness of the specimens was 259.5 μm and the standard deviations were 6.22 μm.
a) Specimen dimensions for the direction-dependent investigation. N represents the direction of the tensile force. The grey dotted area indicates the region sprayed with a black pattern for Digital Image Correlation (DIC) measurements. The evaluated area after the measurements is marked by the bold dashed square at the center of the specimen. b) Cut orientations from the rolled foil. The strips for the first set were cut at 15° increments, as indicated by the dashed lines. The strips for the second set, in MD, TD, and 45°, are marked by dashed lines
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
Two coordinate systems were used in the analysis. The first coordinate system is defined by axes 1–2, where axis 1 represents the direction of uniaxial loading, and axis 2 is perpendicular to axis 1. The second coordinate system is aligned with the material, defined by the machine direction (MD) and the transverse direction (TD). The loading angle is determined by the angle between axis 1 and the MD axis, denoted as φ (Fig. 1).
To determine the direction dependence, two sets of specimens were prepared. The first set consisted of strips cut at ±15° increments from 0° to 180°, as shown in Fig. 1, resulting in twelve different directions, with three specimens for each direction at three different temperature levels. The second set included strips cut in three directions: MD (0°), TD (+90°), and +45°, with 20 specimens for each direction at a single temperature of 24°C. In total, 168 specimens were prepared, and all measurements were rounded to two decimal places.
Figure 1a also illustrates the direction of the tensile force (N). The grey dotted area represents the region sprayed with black paint in a dotted pattern for Digital Image Correlation (DIC) measurements. The area of evaluation is indicated by a bold dashed square at the center of the specimen. Figure 1b shows the cut orientations of the strips in MD, TD, and at 45°.
The Digital Image Correlation (DIC) system was used for the measurements (‘Correlated Solutions Digital Image Correlation’, 2024). Two cameras (Type: Basler acA1920-155um) were used to capture stereo images, each equipped with 25 mm focal length lenses (Type: Basler MEGA M35.5 1) (Fig. 2).
Experimental setup: (1) Test specimen clamped into the material testing machine, (2) DIC camera system, (3) Spotlight aimed at the white wall to improve lighting conditions, (4) Thermometers, and (5) Thermocouples for monitoring temperature and vapor content
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
The image capture frequency was set to 1 Hz. The images were acquired using the VICSnap 9 software (‘Correlated Solutions Digital Image Correlation’). The captured images were then processed with VIC-3D 8.0 software. Strains were calculated at 9 points from an area of approximately 10×10 pixels, arranged in a 3×3 square pattern. Figure 1a shows that the engineering strain field was computed from the displacement field using an 8-tap B-spline interpolation (Sutton et al. 2009).
The temperature levels were selected based on the cooling and heating capabilities of the air conditioner (Gree Comfort X 3.5 kW) in the testing room. The selected temperature levels were 16°C, 24°C, and 32°C (±1°C). The room was conditioned for 24 hours prior to testing to ensure a stable temperature. This time the specimens were stored in the room laying on a steel plate.
The ZWICK Z150 universal tensile testing machine (Fig. 2) was used to apply tension to the specimens. A strain rate of 0.166 %/s was selected, and a 5 N preload was applied to prevent wrinkling in the foils. One rate was used, as the direction dependency was on the focus, not certain parameters. The data capture frequency of the tensile testing machine was set to 10 Hz. The test was terminated upon the onset of plastic deformation. For the diagrams, engineering stress (σeng) and strain (εeng) were calculated using the equations below:
where 𝑁 is the tensile force, ℎ is the thickness and 𝑤 is the width of the measured region of the specimen.
where 𝐿 is the recorded length, 𝐿0 is the initial length of the specimen from the DIC system.
3 THEORY
To investigate the direction dependency of ETFE film, it is necessary to determine the characteristic points of the ETFE foil’s behavior under uniaxial tensile loading (Fig. 3). This study focuses on the characteristic points below the yielding point of the ETFE stress-strain curve, as defined by Surholt et al. (Surholt et al, 2022, 3156).
Generalized engineering stress-strain curve of ETFE foils obtained from uniaxial tensile tests performed at a constant strain rate. The figure highlights the characteristic stress levels on two different scales: (a) up to the ultimate stress level σu, and (b) up to the yield stress σy
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
The first region is the linear viscoelastic region, which ends at the stress σip at an inflection point where the material stiffness decreases by 80% of its maximum. The second region is the nonlinear viscoelastic region, which starts with σip and ends with the yield stress σy of the material. The value of σnve was determined by identifying the strain value where the difference between the measured stress and the straight line defined by the slope σnve exceeds 0.05 MPa.
The yield stress was determined based on EN ISO 527-1, which defines the yield point from the engineering stress-strain diagram, where strain increases without a corresponding increase in stress. In this study, the yield stress was defined as the point where the stress-strain gradient was less than 0.001.
To characterize the ETFE stress-strain curve, three additional points were identified: σ1, the stress at 1% strain; σ10, the stress at 10% strain; and σnve, the stress at the boundary between the first and second parts of the curve, where the stiffness modulus of the nonlinear viscoelastic region (Enve) is defined. The Young’s modulus of the linear elastic region, E0, was calculated from the gradient of the regression line fitted between 0.3% and 0.5% strain, while the stiffness of the nonlinear elastic region, Enve, was calculated in the same way but between 5.0% and 5.2% strain.
One-way analysis of variance (ANOVA) was used to compare several means (Montgomery 2012). This method is commonly employed in experiments to test the equality of population means for specimens subjected to different treatment effects. Measurements with the same cutting direction are defined as groups. Since more than two groups are investigated in this study – where a two-sample t-test would be insufficient – the ANOVA is applied for comparison. The observations are assumed to follow a linear statistical model:
where Yij is the measured stress value from the experiment at the i-th direction and j-th specimen; ndir and nobs are the total number of directions and observations per direction, respectively; β0 is the overall mean of the population; βi is the effect of the cutting direction as the deviation from the overall mean; and ρij is a random error component. The objective is to investigate the effect of the cutting direction on the desired parameter βi:
The total of the observations under the i-th direction is defined as:
It is important that each direction has a similar number of observations. The average of the observations under the 𝑖𝑖-th direction is:
then the sum of all observations is:
the mean of all observation is:
where ndir is the total number of observations of the experiment.
For testing of equality between groups of directions, the following hypotheses can be defined
Here, if the null hypothesis (𝐻0) is true, it means that all the observations from each direction follow the same normal distribution with mean 𝛽 and variance 𝑠2. This suggests no significant difference between the directions, implying that the material is isotropic. Conversely, if the null hypothesis is false, the material likely behaves differently in the corresponding directions, indicating anisotropic behavior.
In ANOVA, Degrees of Freedom (DoF) determine how much independent information is available to estimate variability, given how many parameters are estimated. In ANOVA one parameter, the sample mean is examined. The DoF between groups measures how much variation exists between group means, here the DoF is ndir. The DoF within groups represents the remaining variability within each group after accounting for group differences and now it is ndir (nobs–1). The total amount of variability in the data set is ndir ∙ nobs–1.
To compare the variance between group means to the variance within groups, an F-statistic with the probability distribution of the F-distribution, having ndir–1 and nobs–1 degrees of freedom, was considered. The F-statistic is the ratio of the variances between groups and within groups. A higher F-value indicates a greater difference between groups relative to the variation within groups. To use the F-statistic, it is assumed that the stresses at specific strains for each group are approximately normally distributed, the variances of the groups are homogeneous, and the observations are independent of each other. The F-statistic is then defined as:
where MSDirections is the mean square of stresses at a specific group of directions, MSE is the mean square of error:
where SSDirections is the stresses at directions sum of squares:
The error mean square is defined as:
where SS𝐸 is the error of the sum of squares defined as:
The P-value would be the probability beyond F0 in the Fαs,ndir–1,ndir (nobs–1) distribution. H0 should be rejected, and H1 is true when the P-value is smaller than a predefined significance value. This means that the material is anisotropic. In this case:
where F0 is calculated from Eq. 11, and αS is the level of significance. The expression on the right side represents the critical value determined from F-distribution corresponding to αs, based on the DoF of the numerator and denominator of variances in Eq. 11. F-distribution is a continuous right-skewed probability distribution commonly used for ANOVA analysis. The critical value from the F-distribution is computed using built-in function in MATLAB. If F0 < Fαs,ndir–1,ndir (nobs–1), the corresponding P-value of the F-statistic is larger than the determined significance level. This suggests that it is probable that H0 is true, indicating that the material behavior is isotropic. The significance level in this study was set to 0.05, which is widely used in hypothesis testing, as noted by Montgomery (Montgomery, 2012).
4 RESULTS
All the uniaxial tensile tests described in Sec. 2 revealed the characteristic stress-strain curve with three distinct regions: viscoelastic, nonlinear viscoelastic, and viscoplastic behavior (Fig. 3). Before conducting a comprehensive analysis of direction dependency, the stress-strain curves in the MD and TD were examined at three temperature levels.
With the inclusion of the other investigated directions, besides MD and TD, as presented in Figs 4–6, the overall curves and characteristic points exhibit similar trends but are highly variable between the directions and stresses. The only exception is the viscoelastic region in Fig. 6, where the curves in each direction align closely at the given temperatures. The stresses at the inflection point σip show similar results, though with increased standard deviations. Beyond this point, the curves for each direction begin to diverge. In most cases, the average stress-strain curve in MD is among the less stiff, while TD tends to be among the stiffest. However, no significant pattern was observed between the other directions in this set of measurements. The 45° direction is below or close to MD at 24°C and 32°C in the nonlinear viscoelastic region. At 24°C, the nonlinear viscoelastic-plastic region resembles the region before, while at 16°C, the characteristic value curves are above TD, and at 32°C, the 15° direction is close to TD (Figs 4–6). How-ever, if MD and TD are considered the weak and strong directions of the material, the other directions should fall between these two curves.
Engineering stress-strain curves of ETFE foils as a function of specimen cut orientations at a temperature level of T = 16 °C. The horizontal projection of these stress values is shown on the vertical plane at εeng = 30%
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
Engineering stress-strain curves of ETFE foils as a function of specimen cut orientations at a temperature level of T = 24 °C. The horizontal projection of these stress values is shown on the vertical plane at εeng = 30%
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
Engineering stress-strain curves of ETFE foils as a function of specimen cut orientations at a temperature level of T = 32 °C. The horizontal projection of these stress values is shown on the vertical plane at εeng = 30%
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
The means of the characteristic stress values in the nonlinear viscoelastic-plastic region, along with their COV, suggest isotropic behavior. A horizontal line fits closely to the specific stress values in each direction. However, in MD, TD, or their neighboring angles, local maxima and minima appear on the stress-versus-direction plots (Fig. 7). The variance in the mean stresses and the fluctuating COV indicate that uniaxial tests with three samples per direction do not provide sufficient data to definitively determine direction dependency.
Influence of cut orientation on the stress at specific temperatures for ETFE foils. The vertical axis represents the level of characteristic stresses, while the horizontal axis shows the temperature. At each temperature, the bars corresponding to the characteristic stresses are ordered from left to right as follows: –75° (dark grey) to 90° (white) cut orientations
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
The magnitude of stress change in different directions due to temperature variation is also investigated in Figure 8. In each direction, the stresses decrease. However, the impact of temperature change on the stresses varies across different directions.
Influence of temperature on the stress at specific cut orientations for ETFE foils. The vertical axis represents the level of characteristic stresses, while the horizontal axis shows the cut orientation. At each cut orientation, the bars representing characteristic stresses are ordered from right to left as follows: 16 °C (blue) to 32 °C (red)
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
Between 16 °C and 24 °C, the largest stress change occurred in the 45° direction, with variations ranging from 10% to 27.5% across all stress levels. The smallest change, around 5%, was observed in the MD. However, from 24 °C to 32 °C, the maximum stress difference shifted to MD, with approximately 10% variation across all stress levels in the viscoelastic region. Local maxima were observed around the ±45° directions, while the minima, around 5%, were found in the ±15° directions.
The direction dependence shows that no significant differences or clear order were observed. However, the number of specimens was small, with only 3 specimens per direction at all temperature levels.
The graphical analysis of the larger sample (20 specimens) from the uniaxial tensile test was conducted only at a temperature of 24 °C. The results show similar characteristics as before, but they provide a better representation of the directional material behavior, highlighting the assumed orthotropic behavior. The softest and stiffest directions remain MD and TD, respectively. However, in this case, the 45° direction falls between the two curves starting from the onset of the nonlinear viscoelastic-plastic phase (Fig. 9). This represents a significant deviation from the small sample experiments at 24 °C, where the 45° direction exhibited the softest material behavior. The stress values of σ1 and σip are extremely close to each other in all directions, while from σnve, the difference begins to increase, as shown in Table 1.
Comparison of the effect of cut orientation on the mean characteristic stress values at 24 °C from the stress-strain diagrams determined by uniaxial tensile tests, with twenty specimens tested for each direction. The blue underlined values indicate the smallest mean value, while the red italicized values represent the largest mean value for a specific stress level in each direction
| Temperature | Parameter | Directions [°] | |||
|---|---|---|---|---|---|
| 0 (MD) | 45 | 90 (TD) | |||
| T = 24 °C | σy | mean [MPa] | 17.68 | 18.05 | 18.55 |
| COV [%] | 3.23 | 2.7 | 2.99 | ||
| σ10 | mean [MPa] | 15.55 | 15.82 | 16.21 | |
| COV [%] | 4.14 | 3.24 | 3.22 | ||
| σnve | mean [MPa] | 13.07 | 13.22 | 13.34 | |
| COV [%] | 4.17 | 2.95 | 3.23 | ||
| σip | mean [MPa] | 11.78 | 11.93 | 11.89 | |
| COV [%] | 4.33 | 2.7 | 2.99 | ||
| σ1 | mean [MPa] | 8.23 | 8 | 8.09 | |
| COV [%] | 6.58 | 6.55 | 7.84 | ||
Engineering stress-strain curves of ETFE foils as a function of specimen cut orientations at a temperature of T = 24 °C. The vertical axis represents the engineering stress [MPa], and the horizontal axis represents the engineering strain (%). The curves were obtained from uniaxial tensile tests conducted at a constant strain rate, with twenty specimens tested for each direction at 24 °C
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
The ANOVA analysis described in Sec. 3 was applied to the test in the next step. The parameters of interest, τi, are the mean characteristic stress values of ETFE foils under three different orientations: MD, TD, and 45°. The null hypothesis assumes the equality of these means, and the test statistic is given by Eq. 11. The computed values are summarized in Table 2. The computed test statistic shows a monotonic increase from 0.95 to 6.59 σ0 for the stress values. The ANOVA analysis reveals no significant differences between the elastic Young’s modulus E0 and the stress at 1% strain σ1, as the P-value is higher than 0.05, leading to the acceptance of H0. This suggests that the direction does not significantly affect the material’s behavior, indicating that the material is more likely to behave as isotropic at this stage. However, after the inflection point, the resulting P-value for stresses in the viscoelastic-plastic region significantly drops, almost to zero at σy. Since the P-value is much smaller than αs = 0.05, there is strong evidence that H0 is not true, indicating that the cut orientation of the ETFE foil affects its material behavior and tensile stresses.
One-way ANOVA table for the effect of orientation on the different characteristic values obtained from the uniaxial tensile tests, with twenty specimens tested for each direction (MD, TD, 45°) at a temperature of 24 °C. A significance level of αs = 0.05 was used to evaluate the null hypothesis (H0)
| Parameter | Degrees of freedom | Sum of Squares | Mean Square | F0 | P-Value | |
|---|---|---|---|---|---|---|
| E0 | Groups | 2 | 1835.5 | 917.7 | 0.5 | 0.6052 |
| Error | 57 | 103246.1 | 1811.3 | |||
| Total | 59 | 105081.5 | ||||
| Enve | Groups | 2 | 848.6 | 424.3 | 24.6 | 1.94E-08 |
| Error | 57 | 981.4 | 17.2 | |||
| Total | 59 | 1830 | ||||
| σ1 | Groups | 2 | 0.2 | 0.1 | 0.7 | 0.4836 |
| Error | 57 | 9.3 | 0.2 | |||
| Total | 59 | 9.6 | ||||
| σip | Groups | 2 | 0.2 | 0.1 | 0.7 | 0.4836 |
| Error | 57 | 9.3 | 0.2 | |||
| Total | 59 | 9.6 | ||||
| σnve | Groups | 2 | 0.7 | 0.3 | 1.6 | 0.2025 |
| Error | 57 | 12.1 | 0.2 | |||
| Total | 59 | 12.7 | ||||
| σ10 | Groups | 2 | 4.4 | 2.2 | 7 | 0.0019 |
| Error | 57 | 18 | 0.3 | |||
| Total | 59 | 22.4 | ||||
| σy | Groups | 2 | 7.5 | 3.8 | 13 | 0.000023 |
| Error | 57 | 16.6 | 0.3 | |||
| Total | 59 | 24.1 | ||||
An investigation of the P-value throughout the entire strain history was conducted by performing an ANOVA analysis on discrete stress-strain points. In Fig. 10, the stress-strain curve from the uniaxial tensile test is plotted along with the corresponding P-value. In the elastic region of the curve, the P-value increases to a maximum value of 0.88 at approximately 1.6% strain and 11.01 MPa stress. After this point, the P-value begins to drop, starting around the first inflection point (2% strain at 11.87 MPa stress), as previously investigated. The P-value reaches the chosen significance level of 0.05 at approximately 6% strain, in the middle of the nonlinear viscoelastic-plastic region. Although rejection of H0 is possible below αs = 0.05, this curve suggests that the molecular change to direction-dependent behavior begins earlier in the elastic region, as the probability of accepting H0 significantly decreases after the first inflection point. A discrepancy is observed in the small strain region (<1%), where the P-value drops close to zero, suggesting that H0 could also be rejected. This is likely explained by measurement errors, as the specimens settle into place at the start of testing, causing larger deviations in the measured values.
Results of the ANOVA analysis of stresses for different directions up to the yield point are presented. The primary vertical axis (left) shows the probability of the null hypothesis (H0) from the ANOVA analysis (black continuous line – dimensionless) for each measured stress value. The secondary axis (right) represents the engineering stresses (green dashed lines, in MPa). The shaded grey area indicates the region where the null hypothesis (H0) was rejected, suggesting a significant difference in measurements between the groups corresponding to the cut orientations. The horizontal axis represents the strains up to ϵeng = 25%. The experimental results for the stress-strain curve were obtained using 20 specimens for each orientation
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
It is also essential to verify that the observations follow a normal distribution and are independent of each other. To begin, the normal probability plot of the measured E0 in different directions should be examined, where the data points are plotted against the quantiles of the normal distribution converted into probability values. As shown in Fig. 11, the probability values closely follow a straight line and fall within the confidence bands of the plot, indicating that the measurements follow a normal distribution. The continuous and dashed straight lines represent the expected values for normal distribution between the 25% and 50% percentiles and for the entire population, respectively. The curved lines denote the edges of the 95% confidence band for each direction.
Normal probability plot showing the distribution of initial Young’s modulus (E0) for measurements in different cut orientations, obtained from ANOVA analysis. The horizontal axis represents Young’s modulus [MPa], while the vertical axis represents the corresponding probabilities, determined from the quantiles of the normal distribution in a) MD (crosses), b) 45° direction (triangles), c) and TD (asterisks)
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
The independence of the measurements was verified by plotting the residuals versus time in Figure 12. The plot shows no significant pattern, indicating that the measurements are independent of each other.
Plot of residuals for Young’s modulus (E0) from ANOVA analysis versus the order of observation. The horizontal axis represents the order of observation, while the vertical axis shows the residuals (MPa). The random scatter of residuals around zero indicates that the measurements are independent of each other, with no discernible patterns or trends
Citation: Építés – Építészettudomány 53, 1-2; 10.1556/096.2025.00141
5 CONCLUSION
This paper takes a further step in clarifying the direction dependency of the ETFE membrane, addressing the controversial results reported in previous studies. A systematic laboratory test was conducted using 168 specimens. One group of specimens was analyzed at three different temperatures and twelve different directions, with three specimens for each direction. The other group was analyzed under the same temperature conditions but in three directions, with twenty specimens for each direction. ANOVA analysis was employed to assess the direction dependency.
Both groups of tests show the same trend: the material behaves isotropically in the elastic regime, while anisotropic behavior emerges beyond the yield point. This phenomenon can be explained by significant changes in free volume at higher strain levels, which cause the polymer molecules to reorganize into a more structured, chain-like pattern under load. This behavior is akin to a collection of disorganized threads becoming increasingly aligned when subjected to tension.
According to the measurement results, the MD direction is the softest, while the TD direction is the stiffest. However, the difference is not significant, as shown in Fig. 9 and Table 2. The stresses corresponding to specific elongations differ by less than 5%.
According to the measurements on the NOWOFLON ET 6235 foil the small stress levels (such as those used in serviceability calculations), isotropic models can be applied without any restrictions. However, beyond the yield point, orthotropic models are more suitable. That being said, the isotropic model introduces errors of no more than 5%. Other materials can differ, but we can expect comparable behavior, as the production technology is almost similar in each factory.
ACKNOWLEDGEMENTS
The authors acknowledge support of GRABOPLAN Kft. (9027 Győr, Bútorgyári u. 4.) who provided the ETFE foil for this research.
The DIC system was supported by the European Union and the Hungarian Government in the framework of Competitive Central-Hungary OP (Project ID: VEKOP-2.3.3-15-2017-00017).
REFERENCES
Barthel, Rainer–Burger, Norbert–Saxe, Klaus: Dachkonstruktionen mit ETFE-Folie. DBZ Deutsche Bauzeitung (2003) 72–75.
Correlated Solutions Digital Image Correlation: Correlated Solutions. https://www.correlatedsolutions.com (Accessed 27 March 2025).
De Focatiis, Davide S. A.–Gubler, Lorenz: Uniaxial Deformation and Orientation of Ethylene–Tetrafluoroethylene Films. Polymer Testing 32 (2013) 8. 1423–1435. DOI:
Galliot, Cedric–Luchsinger, Richard H.: Uniaxial and Biaxial Mechanical Properties of ETFE Foils. Polymer Testing 30 (2011) 4. 356–365. DOI:
Montgomery, Douglas C.: Design and Analysis of Experiments. 8th edition. Wiley, Hoboken, NJ 2012.
NOWOFLON ET Architecture: Nowofol. https://www.nowofol.com/en/products/nowoflon-et-architecture (Accessed 11 January 2025).
Nye, John. F: Physical Properties of Crystals. Their Representation by Tensors and Matrices, Clarendon Press, Oxford 1985.
Petraccone, Vittorio–De Rosa, Claudio–Guerra, Gaetano–Iuliano, Mauro–Corradini, Paolo: Fourier-Transform Analysis of Models for the Orthorhombic Crystal Phase of the Alternating EthyleneTetrafluoroethylene Copolymer. Polymer 33 (1992) 1. 22–26. DOI:
Surholt, Felix–Uhlemann, Jörg–Stranghöner, Natalie: Temperature and Strain Rate Effects on the Uniaxial Tensile Behaviour of ETFE Foils. Polymers 14 (2022) 15. 3156. DOI:
Sutton, Michael – Orteu, Jean-José – Schreier, Hubert: Image Correlation for Shape, Motion and Deformation Measurements. Basic Concepts, Theory and Applications. Springer, Berlin 2009. DOI:
Ward, Ian M.: Mechanical Properties of Solid Polymers. 3rd edition. Wiley, Hoboken, NJ 2016.
