Dissertation > Excellent graduate degree dissertation topics show

Reliability Assessment of Corroded Pipeline Based on Correlation of Defects

Author: ZengHaiLong
Tutor: LiShuXin
School: Lanzhou University of Technology
Course: Chemical Process Equipment
Keywords: Pipeline Reliability Monte Carlo simulation Related Artificial Neural Networks Probabilistic assessment
CLC: TE988.2
Type: Master's thesis
Year: 2010
Downloads: 241
Quote: 1
Read: Download Dissertation

Abstract

Oil and gas piped compared to other modes of transport due to losses caused by pipeline failure is a more safe and economical transportation methods. However, it should be noted that the larger pipeline failure of even a single event, such as the cracking of the pipe, but also may cause serious influence on the economy and the environment. Metal pipe wall thinning due to corrosion caused by the most common cause of pipeline integrity missing situation. Now it seems that there are many internal corrosion of the pipe wall thinning mechanism. Back in the 70s of the last century, the forecast on pipeline branch of the American Gas Association sponsored corroded pipeline burst failure pressure standard. This standard usually refers to the the B31G standard, the standard has been part of the ASME B31G pressure piping design specifications. ASME B31G an important role to play in the evaluation of corrosion in pipeline integrity, but later found that the results of the evaluation are too conservative and seem ambiguous in the corrosion of pipeline inspection. At present, the remaining strength of corroded pipelines evaluation using failure pressure model has been published, such as the Modified B31G, the Battelle and DNV-99. Forecast containing the dynamic corrosion defects in pipelines residual strength is a deterministic method that estimated the seriousness of each corrosion defect is to determine the value of the load and resistance parameters. However, the influence of the corrosion the piping load and resistance parameters has a great deal of uncertainty based on the following reasons: (i) defect size measurement uncertainty; (ii) pipe manufacturing uncertainty factors; (iii) pipeline operating conditions do not uncertainty. Reliability of a method for determining the probability of system performance metrics. Uncertainties arise from inadequate information, the reliability of the method can measure this uncertainty. The pipe can be seen as containing a number of shortcomings in the system, the failure mode is determined by the operating pressure and corrosion conditions, the pipe system may be different failure modes failure. With a pipe, due to the corrosive environment and operating conditions are the same, corrosion tend to occur in many positions rather than at some point. Since there is a correlation between the common cause of pipeline corrosion, pipeline defects. Correlation will affect the reliability of the results and the pipe Pipe overall failure. In this article, the direct Monte Carlo method is used to evaluate the reliability of the pipeline. Consider the effects of pipeline corrosion defects reliability probability analysis method. Using this method, including related defects corrosion pipeline failure probability, and compared with the traditional method. The results show that the simple assumption defects independent of each other would be a conservative estimate. The probability calculation results between Ditlevsert second upper and lower bounds, and at the same time that the method is effective and feasible. The sensitivity analysis of a combination of the impact of the combined effects of uncertainties and model considering all the variables of the formulas used. A fuzzy neural network method was used to evaluate the reliability of the oil and gas pipeline corrosion. Network training pipeline magnetic flux leakage tools to detect corrosion defects data, which is used to characterize the situation with the actual service time growth pipeline corrosion. This method is based on the simulation of probability analysis results to construct artificial neural network estimated the reliability of the pipeline, the use of supervised training algorithm to initialize the network weights so that the neural network to predict the oil and gas pipeline failure probability. The neural networks use 11 pipeline and operating parameters as input vector, the output vector is corroded pipeline failure probability. The method has a strong versatility, can be used to make decisions related engineering systems maintenance problems.

Related Dissertations

CLC: > Industrial Technology > Oil and gas industry > Oil machinery and equipment and automation > Corrosion and Protection of machinery and equipment > Corrosion and Protection of Oil and Gas Storage and Transportation Equipment > Pipeline corrosion and protection
© 2012 www.DissertationTopic.Net Mobile