Deformation-induced change to the mechanical properties of polyethylene

31 March 2015
Pean-Yue Ben Jar
A short-term test method to evaluate the performance of polyethylene pipes enables the effect of prior deformation on the tangent modulus to be measured and may prove useful for in-service monitoring.2

Polyethylene (PE) has recently been widely implemented as pipe material for fluid transportation, replacing more traditionally employed materials such as steel, concrete, and clay. Currently, more than 90% of newly installed gas pipe systems are made of PE. As the expected service lifetime of pipe is at least 50 years, information is required regarding the long-term mechanical properties of PE pipe. The current standard test for this purpose (ASTM D1598) requires results to be obtained for full-sized pipe products and takes more than a year to complete. As a result, this test is not suitable for early screening, in-service monitoring, or post-failure analysis. We intend to develop a short-term, coupon-based test method to meet these requirements.

Our belief is that damage accumulation, introduced under normal operation conditions, is the cause of long-term pipe failure. An example of normal operation is the squeeze-off process, which is used for pipe repair or maintenance. This process is designed to stop or reduce the gas flow by flattening the pipe using two metal bars, until the internal surfaces on opposite sides of a cross section meet. Although this process does not generate the kind of permanent deformation that is capable of affecting the immediate performance of the pipe, the local strain it introduces can be very high, particularly along the surface that contacts the metal bars and in the regions where the pipe wall is folded. Experimental studies have shown that pipe sections that have undergone the squeeze-off process are prone to crack development under internal pressure.1, 2

Our method employs two separate tests. In the first test, deformation is introduced. In the second, the influence of this deformation on the mechanical properties is characterized. We used cylindrical specimens of high-density PE (HDPE) with a central gauge section of 6mm diameter and 20mm length. In a previous study, I was able to show that deformation generated in this type of specimen is representative of that which is generated in standard specimens.3 The cylindrical test specimens have axisymmetric geometry, which makes it easier to determine changes in cross-sectional dimensions. The load and diameter of the gauge section, recorded during testing, are used to calculate the true stress (determined by an instantaneous load acting on the instantaneous cross-sectional area) and area strain (based on the logarithmic ratio of the original diameter to the deformed diameter) of the HDPE pipe. The maximum area strain introduced in the first test was 0.8 at a crosshead speed of 1mm/min and 0.5 at a crosshead speed of 5μm/min.

Figure 1 shows the residual strains from the first test (the deformation stage), which we measured at least one month later to ensure full recovery of the viscous deformation (a short-term phenomenon that does not contribute to long-term performance). The results suggest that the residual area strains generated at a speed of 1mm/min are greater than those generated at 5μm/min. After unloading, we were unable to detect any apparent necking (large amounts of strain localized in small areas) in the specimens that were subjected to an applied strain of below 0.5 in the first test.

Residual area strain of the high-density polyethylene specimens, measured at least one month after the first test, plotted as a function of the maximum strain applied. R2: Goodness of fit.

Figure 2 shows the tangent modulus of the specimens shown in Figure 1. These were measured during the second test (the characterization stage) at an area strain of 0.01 and a crosshead speed of 1μm/min. This low crosshead speed was used to minimize the influence of viscous behavior on the measured tangent modulus and to reduce the possibility of further damage generated from the second test.3

Values for the tangent modulus (measured at an area strain of 0.01 in the second test) as a function of the maximum area strain applied in the first test, obtained using two crosshead speeds (1mm/min and 1μm/min).

A change in the elastic modulus may indicate the presence of damage in materials. A damage parameter D, based on the ratio between the elastic moduli of the damaged and virgin speciments, has been proposed to quantify this damage.4 D increases from zero for no damage, to 1 for complete failure. Based on this concept, Figure 2 suggests that deformation introduced in the first test causes damage to the specimens, and that this damage is introduced before the yield point. However, it is not yet known whether this decreased elastic modulus reduces the fracture toughness. Therefore, an increase in this kind of damage may not cause further degradation of the mechanical properties.

Figure 3 shows that an increase in the level of area strain introduced during the deformation stage shifts the stress/original-strain curves, generated during the characterization stage, to the right and downwards, indicating a reduction in the load-carrying capacity. However, in the high-strain section of the curves (i.e., above the maximum area strain introduced in the first test), the amount of shift shown in Figure 3 is nearly constant, independent of the area strain introduced in the first test. This suggests that the reduction in load-carrying capacity is caused by the deformation introduced in the first test. We also found evidence that cyclic loading used to introduce this deformation leads to a shift of the stress/original-strain curves that may not occur in the same way as is shown in Figure 3. Using cyclic loading, the curve may actually be higher than that of the virgin specimen.5 It is therefore unclear whether the change in the tangent modulus shown in Figure 2 does indicate the presence of damage in the traditional sense, or is simply an indication of the change in PE's resistance to deformation at the low deformation level at which the tangent modulus is measured.

Plot of true stress as a function of original strain from the second test on specimens that have been subjected to a crosshead speed of 1mm/min. Numbers indicate the maximum area strain introduced to the specimens in the first test.

In summary, we have developed a method to evaluate the influence of prior deformation on the tangent modulus for HDPE specimens at an area strain of 0.01. Results show that the tangent modulus decreases as the strain introduced in the prior deformation increases. We are currently working to determine whether mechanical properties, such as fracture toughness, are affected by the deformation introduced in the first test.


Pean-Yue Ben Jar
University of Alberta

Ben Jar is a professor at the University of Alberta. He has studied the mechanical behavior of polymers and polymer-based fiber composites for more than 25 years.


  1. N. Brown and J. M. Crate, Analysis of a failure in a polyethylene gas pipe caused by squeeze off resulting in an explosion, J. Fail. Anal. Preven. 12 (1), pp. 30-36, 2012.

  2. P. Yaylaa and Y. Bilgin, Squeeze-off of polyethylene pressure pipes: experimental analysis, Polym. Test. 26 (1), pp. 132-141, 2007.

  3. P.-Y. B. Jar, Transition of neck appearance in polyethylene and effect of the associated strain rate on the damage generation, Polym. Eng. Sci. 54 (8), pp. 1871-1878, 2014.

  4. J. Lemaitre, A continuous damage mechanics model for ductile fracture, J. Eng. Mater. Technol. 107 (1), pp. 83-89, 1985.

  5. P.-Y. B. Jar, Effects of tensile loading history on mechanical properties for polyethylene, Polym. Eng. Sci., 2014. First published online: 28 November. doi:10.1002/pen.24042

DOI:  10.2417/spepro.005821

Footer Links (2nd Row)