136

Susannah Turner, Jack Davies, Fraser Gray, Tim Turner Highgrade Associates, Newcastle upon Tyne, UK

Deterministic integrity assessment uses characteristic values as inputs to a limit state equation. The output is a deterministic prediction of, for example, failure pressure or limiting anomaly dimensions for failure. We deal with uncertainties in the assessment by choosing appropriately conservative, characteristic inputs, and applying a safety factor or margin of safety to the result. This allows us to make integrity management decisions which are considered sufficiently safe and conservative. Different safety margins may be selected based on potential failure consequence, such that a deterministic assessment is not necessarily insensitive to risk. However, these assessments do not attempt to quantify probability of failure and may be criticised for leading to acceptance of an unquantified level of risk.

Probabilistic assessments using structural reliability techniques would claim to quantify uncertainty, giving a calculation of the probability of failure. A probabilistic assessment may essentially use the same limit state equation as a deterministic assessment, but with input parameters as distributed variables, including parameters to account for the uncertainty of the limit state model itself. The calculated failure probability is then compared with industry or company criteria to make integrity and risk management decisions.

This paper considers the sources of uncertainty in probabilistic assessments and discusses the extent to which calculated failure probabilities represent a true measure of the probability of failure. Case study comparisons are shown between deterministic assessment limits and the associated failure probability levels calculated using structural reliability techniques, for pipeline corrosion metal loss assessments. Observations are made regarding the comparisons and conclusions are drawn with respect to how this understanding can lead to improved decision making.

“If a man will begin with certainties, he shall end in doubts; but if he will be content to begin with doubts, he shall end in certainties.” Francis Bacon 1605

Read/Download PaperGradeAvailable after the event