Weibull analysis, extrapolations and implications for condition assessment of cast iron water mains

Abstract

The applicability of Weibull statistics to the condition assessment of cast iron water distribution pipes has been considered. The effect of Weibull modulus, characteristic strength, sample size and mode of loading (tension or flexure) on the strength of cast iron water distribution pipes is investigated. The strength distribution of cast iron samples cut from sections of five different water distribution pipes recovered from the ground have been characterized. Strengths have been measured in flexure, at two different temperatures (ambient and 0 degrees C), and in tension at ambient temperature using two different sample sizes. It is shown that characteristic strength values in flexure decrease with increasing size of graphite flake and that there is no significant difference between the results at the two temperatures investigated. For samples of the same volume tested in tension and flexure, the reduced strength measured in tension is consistent with Weibull predictions. However, the strength of large samples tested in tension was not significantly different from the small samples, perhaps because the samples were of the same thickness and conventional Weibull scaling is not applicable. Finally, using a method which treats a large pipe as an assembly of small samples, the strength distributions from the small samples tested in tension are used to make a prediction of the strengths of 1 m span sections of pipe loaded in three-point bending, which were reported in previous work. The predicted pipe strengths are close to the lower end of the measured pipe strength distribution. Overall, this work suggests that Weibull analysis is a useful tool to examine the strength distribution of removed from cast iron water pipes and so has the potential to contribute in the assessment of asset condition

Thermally induced strains and stresses in cast iron water distribution pipes: an experimental investigation

“…(A value of 117 GPa is at the upper end of the typical range of 80 to 120 GPa for these materials). Moreover the measured tensile stress at failure for this pipe is consistent with residual strength measurements, and associated Weibull-scaling arguments presented, in previous work on degraded pipe samples (Belmonte et al 2007(Belmonte et al , 2008. Clearly a pipe with a greater degree of degradation will require a smaller temperature difference to trigger fracture.…”

“…A number of researchers have reported on the degradation of strength of cast iron pipes as a result of graphitic corrosion and its relationship to residual performance (Yamamoto et al 1983;Sheikh et al 1990;Conlin and Baker 1991;Rajani and Makar 2000;Atkinson et al 2002;Seica and Packer 2004). In particular, recent studies (Belmonte et al 2007(Belmonte et al , 2008) have used Weibull-based methods to investigate the strengths in tension and flexure of small samples of cast iron extracted from cast iron water distribution pipes. It has been shown that the 5 of 28 strength is degraded significantly as a result of graphitic corrosion in service and…”

Fatigue of metals under nonuniform stressed state. part 2. methods of the analysis of research results

“…The d c value determined by relation (40) is substantially higher than that in the axial loading case.…”

What does NDE need to achieve for cast iron pipe networks?

“…Further, because of the combination of the size of the pipe and its interaction with the local environment, it is possible to comment on the behaviour of the main a significant distance (multiple casting lengths or 'sticks' of pipe, typically 3-5 m long) either side of a sampling point (Belmonte et al, 2007(Belmonte et al, , 2008(Belmonte et al, , 2009. The size and spread of graphitic corrosion in distribution mains can be seen in Figure 4.…”