Reliability Assessment Of Corroded Pipeline Based On Correlation Of Defects

Pipelines are a safe and economics means for transporting oil and natural gas with the number of failures, defined as a loss of product, being relatively low compared to other means of transportion. However, it should be pointed out that even a single major failure, such as a rupture, will have a significant financial and environmental impact. Metal loss due to corrosion is one of the most common situations leading to the loss of pipeline integrity. There are various mechanisms for internal corrosion which may produce local reducions in wall thickness.In the early 1970s, a criterion for the prediction of the burst pressure of corroded pipeline was developed through research sponsored by the American Gas Association Pipeline Research Committee and the pipeline industry. This criterion, commonly referred to as the B31G criterion, has been embodied in pipeline design codes that are part of the ASME B31G Code for pressure piping. While ASME B31G has been very useful in evaluating the integrity of corroded pipe, it has been found to be overly conservative and ambiguous to apply in many situations encountered during pipeline inspections. At present, the remaining strength of corroded pipeline assessment is carried out using several already published failure pressure models, such as Modified B31G, Battelle and DNV-99 etc.The prediction of the remaining strength of pressure pipelines containing active corrosion defects is carried out using deterministic methods. These methods estimate the severity of each individual corrosion defect using nominal values for both the load and the resisrance parameters. However, it is well known that the load and resistance parameters have important uncertainties which reslut from:(â…°) the measurement of the dimensions of defects,(â…±) the manufacture of the pipe and (â…²) the operating conditions of the pipline. Reliability is a probabilistic measure of assurance of the performance of a system. Uncertainties are due to incomplete information, and reliability approaches account for such uncertainties. The pipeline is assumed to be a series system. The failure mode is considered to be controlled by the stresses due to internal pressure and the presence of corrosion. Pipeline system may fail in different modes of failure. As a result of environmental exposure and operaton, corrosion tends to appear at several locations instead of a single point. Because of common causes of corrosion over the pipeline segments, it is expected that some degree of spatial correlation exists. Correlations can affect the sequcence of failure combinations and the reliability of the pipeline. In this dissertation, The direct Monte Carlo method was used for assessing the reliability of pipeline. A method of probabilistic analysis is presented to consider effect of correlation of corrosion defects on steel pipeline failure. Failure probability of pipeline with correlated corrosion defects is modeled and compared with results from conventional method. The results showed that assumption of independent corrosion defects will lead to conservative results.The calculated results are within Ditlevsen bimodal bounds.This indicates the validity and feasibility of the proposed method. The combined effect of both the model sensitivity of prediction to variables and the uncertainty of variables should be taken into account in sensitivity analysis.A fuzzy artificial neural network (ANN)-based approach is proposed for reliability assessment of oil and gas pipelines. The proposed ANN model is trained with field observation data collected using magnetic flux leak-age (MFL) tools to characterize the actual condition of aging pipelines vulnerable to metal loss corrosion. The approach is to transform a simulation-based probabilistic analysis framework to estimate the pipeline reliability, using supervised training to initialize the weights so that the adaptable neural network predicts the probability of failure for oil and gas pipelines. This ANN model uses eleven pipe parameters as input variables. The output variable is the probability of failure. The proposed method is generic, and it can be applied to several decision problems related with the maintenance of aging engineering systems.