Frans Scholten, Ernst van der Stok
Papers # 2016 Berlin
The residual quality of first generation PE50 gas pipes, installed in Netherland between 1961 and 1979, is still remarkably good. The only premature failures of a few PE50 pipes were exclusively caused by point loads. When under point loading, they started failing in practice after 8 to 26 years, in a brittle manner. The effect of point loads was modeled in order to help grid-owners to prioritize replacements.
Many Point Load Tests (PLTs) were performed using 8 mm indentation and a new Dehyton PL detergent on excavated, defect-free, 110 mm PE50 gas pipes. The influence of temperature (40, 60 and 80 °C), pressure and SDR (standard dimension ratio) was measured and later modeled using a multi-variate statistical model. The squared correlation coefficient (R2 ) was found to be favorably high: 0.940. For all three variables, the statistical probability (Pstat) of there being no correlation is favorably low (for all three variables <0.033), and lower than the requirement (<0.05). All three variables therefore have a separate effect on lifetime.
The PLT model shows that a lower gas pressure reduces the effect of a point load and that thin-walled PE50 pipes fail sooner than thick-walled pipes, probably because in a thin-walled pipe a growing crack needs to travel a shorter distance through the wall until a leak is formed.
The acceleration factor in the PLT with Dehyton PL detergent with respect to water is 16. With this factor, the PLT model predicts a lifetime for PE50 pipes at 3 and 4 bars and 10 °C of 5 to 35 years, which is remarkably close to practice, demonstrating the value of the PLT model.
When, for the sake of argument, the LPL in ISO 9080 for PE50 pipes without point loads is extrapolated a failure time of about 4,300 years at 4 bars and 10 °C is calculated.
The time to failure in the PLT for PE80 pipes is 9 to 10 times longer than for PE50 pipes.