Assessment of the Reliability Level Embedded in Pipeline Design Codes

Author(s):  
Sviatoslav Timashev ◽  
Anna Bushinskaya

Recently a long time discussion among specialists about the meaning of the probabilities of failure (POF) produced by different reliability analysis methods surfaced in pipeline journals. This paper, which was a long time in the making, is a follow-up on the discussion and analyses the actual reliability level which was empirically embedded in codes for pipeline design [B31G, B31Gmod, Shell92 and Battelle (PCORRC)] and Building Standard (BR) Main Pipelines #2.05.06-85, using a real pipeline as an example. Assessment of the actual reliability level empirically embedded in BR is based on assessing the order of the quintiles of strength parameters (design values of tensile strength and yield strength of the pipe material) and load (internal pressure) on the pipeline. This approach allows direct connection of the deterministic safety coefficients used in the BR with the level of reliability of the pipelines associated with these coefficients. The actual reliability level, empirically embedded in international codes, is calculated as the probability that the limit state function (LSF) of ideal pipeline (without defects) is positive. LSF = Pf − Pop, where Pf is the failure pressure of an ideal pipe, which is estimated by any design code; Pop is the operating pressure. The failure and operating pressure are considered as random variables. The expression for this probability was obtained analytically and in closed form. Recommendations are also presented for choosing probability distributions and statistical parameters for random variables RVs. Extensive calculations permitted revealing the reliability levels which are actually present in the analyzed international pipeline design codes. In a nutshell, the paper proves that the international codes under consideration are very reliable, as they produce very safe designs of pipelines with very low POF, and, hence, provide large safety coefficients, and that the algorithm developed in the paper permits connecting the current level of pipeline degradation (in terms of POF), with its current safety coefficient, which, in this case, is a function of time. All calculating in the paper where performed using MathCAD. Illustrations of these calculations are also presented in the paper.

2007 ◽  
Vol 353-358 ◽  
pp. 2561-2564
Author(s):  
Ouk Sub Lee ◽  
Dong Hyeok Kim

The reliability estimation of pipeline is performed in accordance with the probabilistic methods such as the FORM (first order reliability method) and the SORM (second order reliability method). A limit state function has been formulated with help of the FAD (failure assessment diagram). Various types of distribution of random variables are assumed to investigate its effect on the failure probability. It is noted that the failure probability increases with the increase of the dent depth, the operating pressure and the outside radius, and the decrease of the wall thickness. Furthermore it is found that the failure probability for the random variables having the Weibull distribution is larger than those of the normal and the lognormal distributions.


1984 ◽  
Vol 106 (3) ◽  
pp. 362-368
Author(s):  
J. F. Kiefner ◽  
T. P. Forte

An analytical model is presented for predicting hydrostatic retest intervals in liquid pipelines which are subjected to frequent large pressure cycles. The model utilizes pressure cycle history, hydrostatic test history, and fatigue crack growth rate data for the pipe material to calculate time to failure for the largest possible defect which could have survived a previous hydrostatic test. An example problem is described which shows the value of maximizing the margin between test pressure and operating pressure in order to achieve long time intervals between tests.


Author(s):  
Zhangli Hu ◽  
Xiaoping Du

In traditional reliability problems, the distribution of a basic random variable is usually unimodal; in other words, the probability density of the basic random variable has only one peak. In real applications, some basic random variables may follow bimodal distributions with two peaks in their probability density. When binomial variables are involved, traditional reliability methods, such as the first-order second moment (FOSM) method and the first-order reliability method (FORM), will not be accurate. This study investigates the accuracy of using the saddlepoint approximation (SPA) for bimodal variables and then employs SPA-based reliability methods with first-order approximation to predict the reliability. A limit-state function is at first approximated with the first-order Taylor expansion so that it becomes a linear combination of the basic random variables, some of which are bimodally distributed. The SPA is then applied to estimate the reliability. Examples show that the SPA-based reliability methods are more accurate than FOSM and FORM.


Author(s):  
Ian Matheson ◽  
Wenxing Zhou ◽  
Joe Zhou ◽  
Rick Gailing

The reliability-based design and assessment (RBDA) methodology has gained increasing acceptance in the pipeline industry, largely due to a multi-year PRCI program aimed at establishing RBDA as a viable alternative for the design and assessment of onshore natural gas pipelines. A key limit state of buried pipelines that operate at elevated temperatures is upheaval buckling. The elevated temperatures generate large compressive axial forces that can cause Euler buckling susceptibility. The tendency to buckle is increased at vertical imperfections (i.e. a series of cold formed bends) that primarily occur due to topography. Upheaval buckling in itself is not an ultimate limit state but can lead to high strains, local buckling, high cycle fatigue, expensive remediation measures, and even loss of pressure integrity. The critical forces at which upheaval buckling occurs for typical hill-crest type imperfections present in onshore pipelines cannot be readily predicted using analytical methods. A parametric study is therefore undertaken using non-linear finite element analyses to generate a matrix of upheaval buckling responses. The critical force for the onset of upheaval buckling is then developed using a series of empirical relationships to capture the influences of all key parameters. An upheaval buckling limit state function is subsequently developed by comparing the critical buckling force with applied compressive force, which is a function of operating pressure and temperature differential between the operating and tie-in conditions. The limit state function can be readily implemented in a reliability analysis framework to calculate the pipeline failure probability due to upheaval buckling.


2005 ◽  
Vol 127 (4) ◽  
pp. 353-358 ◽  
Author(s):  
Bernt J. Leira ◽  
Trond Stokka Meling ◽  
Carl M. Larsen ◽  
Vidar Berntsen ◽  
Bernie Stahl ◽  
...  

Safety factors required to control fatigue damage of deepwater metallic risers caused by vortex-induced vibration (VIV) are considered. Four different riser configurations are studied: Cases I and II: Vertical tensioned 12in. risers suspended from a spar buoy at water depths of 500 and 1500m. Cases III and IV: Steel catenary risers suspended from a spar buoy, both at 1000m. For Case III, the riser diameter is 12in., while for Case IV it is 33in. For each riser configuration, relevant design and analysis parameters which are subject to uncertainty are identified. For these quantities, random variables are established also representing model uncertainties. Subsequently, repeated analyses of fatigue damage are performed by varying the input parameters within representative intervals. The results are applied to fit analytical expressions (i.e., so-called response surfaces) utilized to describe the limit state function and to develop the probabilistic model for reliability analysis of the risers. By combining the random variables for the input parameters with the results from the parameter variations, a relationship between the fatigue safety factor and the failure probability is established for each riser configuration.


2015 ◽  
Vol 61 (3) ◽  
pp. 133-147 ◽  
Author(s):  
A. Dudzik ◽  
U. Radoń

AbstractThe study deals with stability and dynamic problems in bar structures using a probabilistic approach. Structural design parameters are defined as deterministic values and also as random variables, which are not correlated. The criterion of structural failure is expressed by the condition of non-exceeding the admissible load multiplier and condition of non-exceeding the admissible vertical displacement. The Hasofer-Lind index was used as a reliability measure. The primary research tool is the FORM method. In order to verify the correctness of the calculations Monte Carlo and Importance Sampling methods were used. The sensitivity of the reliability index to the random variables was defined. The limit state function is not an explicit function of random variables. This dependence was determined using a numerical procedure, e.g. the finite element methods. The paper aims to present the communication between the STAND reliability analysis program and the KRATA and MES3D external FE programs.


Author(s):  
Robert I. Coote ◽  
J. Kyle Keith

Pipeline companies often reduce the pressure while performing maintenance activities and integrity excavations on in-service pipelines. Despite this practice, pipeline design codes, regulations and industry publications offer little guidance on what factors should be considered to determine how much, if any, the pressure should be reduced from operating levels during excavation activities. Also, it is not commonly understood what level of safety is introduced with these reductions and what historical operating pressure level should be used as the basis for the reductions. A literature survey and an interview process with CEPA member companies summarized common industry practices and determined factors to be considered when assessing if and how much of a pressure reduction is appropriate while excavating an operating energy pipeline.


2018 ◽  
Vol 140 (3) ◽  
Author(s):  
Dimitrios I. Papadimitriou ◽  
Zissimos P. Mourelatos

A reliability-based topology optimization (RBTO) approach is presented using a new mean-value second-order saddlepoint approximation (MVSOSA) method to calculate the probability of failure. The topology optimizer uses a discrete adjoint formulation. MVSOSA is based on a second-order Taylor expansion of the limit state function at the mean values of the random variables. The first- and second-order sensitivity derivatives of the limit state cumulant generating function (CGF), with respect to the random variables in MVSOSA, are computed using direct-differentiation of the structural equations. Third-order sensitivity derivatives, including the sensitivities of the saddlepoint, are calculated using the adjoint approach. The accuracy of the proposed MVSOSA reliability method is demonstrated using a nonlinear mathematical example. Comparison with Monte Carlo simulation (MCS) shows that MVSOSA is more accurate than mean-value first-order saddlepoint approximation (MVFOSA) and more accurate than mean-value second-order second-moment (MVSOSM) method. Finally, the proposed RBTO-MVSOSA method for minimizing a compliance-based probability of failure is demonstrated using two two-dimensional beam structures under random loading. The density-based topology optimization based on the solid isotropic material with penalization (SIMP) method is utilized.


Author(s):  
Michael Gardiner ◽  
Ross Michie ◽  
Gerardo Douce

Metrogas SA operates a natural gas distribution concession within the Greater Buenos Aires region of Argentina. In August 2007 a failure occurred on a section of the 22-bar system that dates from the early 1960s and, as such, was ‘inherited’ by Metrogas at privatization. The line pipe in this part of the system is spirally welded and at the failure point the spiral weld root was found to have been incomplete. Subsequent investigations showed that incomplete spiral welds were also present at other locations in the same section of the system. This paper describes some of the steps taken to investigate the incident of 2007 and to manage the threat from other defective spiral welds in the same pipeline section. We present a limit state model for through-wall failure of such features and show how this was used to help understand the incident. We also discuss modeling of uncertainties in parameters of the model and look at results from a probabilistic structural reliability implementation of the limit state function, which allowed the failure frequency of other defective spiral welds in this section to be predicted for various reductions of the operating pressure. Metrogas was then able to use these quantified reliability data to make a responsible, informed decision to keep the affected section in downrated service.


Author(s):  
Ho Hyun Lee ◽  
Hae Sung Lee

<p>This proceeding presents the calibration process of load and resistance factors for the design of cable members under a gravitational loads-governed limit state adopting optimization scheme. In reliability-based bridge design code, although the cable members show various behavior depending on the structural types of bridges, a proper reliability level should be satisfied by the load and resistance factors. A cable is a nonlinear component, thus tension of it also shows nonlinear characteristics. In this study, the limit state function is linearized, and the tension of each load component is normalized by total nominal tension. With the purpose of performing code calibration independent of structural types of bridges, the normalized tensions are parameterized by three load ratios. The target reliability indices of cable members are determined considering results of reliability analyses of existing cable-supported bridges in South Korea, and a target strength, which satisfies the target reliability indices exactly, is evaluated. Optimization problem to minimize an error between the target strength and nominal strength, which is calculated by the load and resistance factors, is defined, and optimal values of the factors are calibrated. Reliability analyses for the strength calculated from the optimal factors are performed and it is verified that the factors can lead to the design with a uniform reliability level.</p>


Sign in / Sign up

Export Citation Format

Share Document