No abstract

While the use of environmental factors in the analysis and prediction of failures of buried reticulation pipes in cold environments has been the focus of extensive work, the same cannot be said for failures occurring on pipes in other (non-freezing) environments. A novel analysis of pipe failures in such an environment is the subject of this paper. An exploratory statistical analysis was undertaken, identifying a peak in failure rates during mid to late summer. This peak was found to correspond to a peak in the rate of circumferential failures, whilst the rate of longitudinal failures remained constant. Investigation into the effect of climate on failure rates revealed that the peak in failure rates occurs due to differential soil movement as the result of shrinkage in expansive soils.

Asbestos cement (AC) pipes are among the oldest assets in many water supply networks. With increasing failure rates and cost consequences, asset management tools are required to pre-empt unplanned failures and schedule future replacement at the most cost-effective time during service. This paper develops a physical probabilistic failure model for AC pipes under combined internal pressure and external loading. Uncertainty in the degradation process is accounted for using the Weibull extreme value probability distribution. Monte Carlo simulation is then used to estimate the probability of pipe failure as ageing proceeds. In the final stage of the model, a cost/benefit analysis is conducted to determine optimal scheduling of future inspection and replacement activities.The end result is a potentially useful asset management methodology in which both the probable physical lifetime and the economic lifetime of an AC pipe can be estimated. C t probable failure costs for new pipe asset at time t (AU$) d discount rate for future cash flows D goodness-of-fit test statistic F(s R ), F(t) cumulative probability distributions for degradation rate and pipe lifetime f(s R ), f(t) probability density functions for degradation rate and pipe lifetime F m bedding factor H soil cover depth (m) h(t) hazard (failure rate) function N number of trials in Monte Carlo simulation NPV net present value (AU$) P internal pressure of pipe (Pa) p c internal pressure required for failure with no external load (Pa) s standard deviation s 0 original, as-produced tensile strength (MPa) SEM standard error of mean s f residual tensile strength at failure (MPa) s R degradation (strength loss) rate (MPa/year) T MAX maximum physical lifetime of pipe (years) w external load (N m 21 ) w c external load required for failure with no internal pressure (N m 21 ) Z AC known cost of failure in existing pipe asset (AU$) Z NEW known cost of failure in new pipe asset (AU$) a scale parameter for Weibull extreme value probability distribution g soil unit weight (kN m 23 ) dt time step for analysis (years) h shape parameter for Weibull extreme value probability distribution

A new mathematical model is presented that describes the shrinkage curve of environmentally stabilized soils. Environmentally stabilized soils are defined as those that have undergone a sufficient number of wet–dry cycles to reach a stable soil structure. The model is applicable for fitting to soils that exhibit all typical zones of soil shrinkage behaviour and to those soils that do not exhibit structural and (or) residual shrinkage behaviour. The fitting parameters in this model are directly related to features of the shrinkage curve and have direct relevance to soil mechanics theory and practice. The model is applied to data from 20 different datasets reported in literature with excellent fitting achieved. This model is extended to incorporate the effect of net stress, creating a surface describing soil volumetric behaviour in response to changes in water content and net stress.

The soil water retention curve (WRC) represents the relationship between soil suction and water content. Previous equations for representing the WRC do not include a direct relationship between features of the curve and equation fitting parameters. We present a new method for developing equations for representing empirical data that ensures that features of the curve are directly related to fitting parameters in the equation. The method also ensures that the equation has a simple derivative and is readily integrated, an important property if the equation is to be used as part of a numerical model. This method has been applied to the WRC, producing a new equation that is able to accurately represent the entire WRC using a single equation. The direct relationship between the main features of the curve and fitting parameters ensures that the fitting parameters have clear theoretical meaning. Additionally, this direct relationship allows fitting parameters, such as the air-entry value, to be calculated using pedotransfer functions. The equation has been fitted to a wide range of experimental data and has been demonstrated to achieve an excellent fit for sand, silt, and clay soils across the entire suction range from the minimum measured suction value to 10^ kPa.Abbreviations: WRC, water retention curve.