Paper No. IPC2022- 87258

This paper explores using statistical and machine learning models, like logistic regression and random forest, to quantify pipe failure probability in distribution pipelines. Trained on historical incident records, model performance is compared using lift charts. The study discusses each model's strengths, limitations, and data integration. Case studies show how training data quantity and external datasets impact effectiveness. The results aim to guide operators in developing machine learning models for pipeline risk assessment and integrity management.

Read More

Paper No. IPC2022-87255

Structural reliability calculates failure probabilities in pipelines to manage risks like corrosion and third-party damage. Traditional Direct Monte Carlo (DMC) simulation is computationally intensive. This paper introduces Subset Simulation, which reduces computational demands by decomposing rare event probabilities. Applied to pipeline failure models, it is validated against DMC, showing improved efficiency and reliability in assessing pipeline integrity.

Read More

Paper No. IPC2022-87148

Co-locating pipelines with AC transmission lines creates electrical hazards during power line faults, which electrify nearby pipelines. This paper presents a method to assess the impact of these faults using spatial and historical data. The model estimates fault frequency, calculates fault currents, and identifies high-risk areas, helping prioritize locations for site analyses and mitigation system design.

Read More

Paper No. IPC2022-87145

This paper addresses challenges in pipeline facility reliability assessments when essential data is missing. Traditional methods often use conservative assumptions, inflating risk assessments. The paper proposes using K-means clustering to group equipment data and fill gaps with reasonable estimates. This method balances conservatism and accuracy, prioritizing equipment with confirmed attributes while considering those with limited data. Practical applications, including dimensionality reduction and sensitivity analysis, are also discussed.

Read More

Paper No. IPC2022-87066

Gas distribution systems don't fit traditional pipeline risk models. This paper presents predictive methods to prioritize mitigation and replacement, using a model expanded from a city to Saskatchewan. It links failure rates with asset characteristics through logistic regression and machine learning, enhancing integrity planning. The framework also classifies leaks by hazard levels, assessing failure probability and consequences for improved operational responses.

Read More

Paper No. IPC2022-87060

The API’s third edition of the 1163 Standard (2021) enhances ILI tool performance validation, focusing on Level 3 validation with real-world data. It addresses issues like truncation, where measurements below a threshold are not reported. This paper presents a model to estimate ILI tool performance despite truncation, validated through simulations and field data, improving pipeline integrity assessments and risk analyses.

Read More

Paper No. IPC2022-86833

A risk assessment framework for SoCalGas’s underground gas storage sites in California, aligned with API RP 1171 and regulations, evaluates 80+ failure mechanisms using a dynamic fault tree. It prioritizes high-risk threats, while simpler models are used for less critical ones. The modular design incorporates factors like monitoring and wellhead spacing, enhancing integrity management activities such as erosion and pressure monitoring.

Read More

Paper No. IPC2024-134136

This paper presents a probabilistic model to assess erosion risk in underground gas storage well laterals, following 49 CFR 192.12. Using DNV's guidelines and Monte Carlo simulations, it estimates wall thickness loss, failure probability, and time to failure, accounting for operating uncertainties. The model also identifies erosive environments and recommends mitigation strategies, emphasizing well-specific assessments for effective erosion management.

Read More

Paper No. IPC2022-86734

This paper presents a risk assessment framework for SoCalGas’s underground gas storage sites, compliant with API RP 1171 and updated regulations. It evaluates accidental impact threats to wellheads and piping from excavations, vehicle collisions, lifting operations, and aircraft crashes. Using well-specific data and expert judgment, the models provide risk estimates for individual wells, aiding mitigation efforts and emphasizing the importance of tailored safety assessments in gas storage operations.

Read More

Paper No. IPC2022-86794

This paper highlights the need for ongoing mechanical integrity evaluations of aging gas storage wells due to corrosion risks. It proposes a probabilistic corrosion analysis to determine optimal reinspection intervals that minimize risk, using well configuration and loading data. A Bayesian approach refines corrosion growth rate distributions, improving accuracy. The findings show how tailored inspection and maintenance plans can manage integrity risks for individual wells.

Read More

Paper No. IPC2024-133999

This paper presents a Bayesian model to estimate the service life of pipeline coatings, vital for preventing external corrosion. By combining prior knowledge with integrity dig data, the model quantifies the probability of coating durability over time. It improves corrosion growth rate estimates and supports probabilistic assessments, aiding better decision-making in pipeline maintenance and safety.

Read More

Paper No. IPC2024-134143

This study introduces a machine learning approach to predict pipeline stress as a percentage of the Specified Minimum Yield Strength (% SMYS), addressing data gaps. The models help classify segments for integrity management, reducing manual reviews. By identifying high-stress segments, especially those over 30% SMYS, and using models like random forest and XGBoost, the approach improves accuracy with feedback from manual reviews, demonstrating machine learning's potential in pipeline assessments.

Read More

Paper No. IPC2024-133658

This paper introduces a non-linear Bayesian sizing model to improve in-line inspection (ILI) tool validation, addressing limitations of the traditional linear approach. While API 1163 uses Bayesian methods, it assumes consistent performance across all flaw sizes, leading to inaccuracies, especially in thin-wall pipelines. The study highlights height sizing errors, crucial for fitness-for-service assessments of crack-like flaws, and includes a case analysis to demonstrate non-linear ILI performance. Although focused on crack-like flaws, the model is also applicable to metal loss ILI results

Read More

Paper No. IPC2024-134143

This study introduces a machine learning approach to predict the stress levels of steel pipelines as a percentage of the Specified Minimum Yield Strength (% SMYS), addressing incomplete data issues. The models estimate % SMYS to assist in classifying pipeline segments for integrity management and regulatory compliance, reducing the need for manual record reviews. Using regression and classification techniques, the study identifies high-stress segments, particularly those over 30% SMYS, for prioritized review. Models like random forest and XGBoost are trained with cross-validation, and feedback from manual reviews improves accuracy. Overall, the study highlights machine learning's potential in pipeline integrity assessments despite data gaps.

Read More

Paper No. IPC2024-130792

This paper presents a framework for assessing risks at SoCalGas natural gas storage wells, exceeding regulatory requirements. It evaluates life safety and environmental impacts using a model for ignition probabilities and an adapted PIR model. The methodology estimates risks and methane release quantities, combining safety, environmental, and societal cost metrics to assess mitigation measures, like downhole safety valves, and improve safety assessments.

Read More

Paper No. IPC2024-133369

This paper introduces a framework for calculating reliability benchmarks for onshore HVL pipelines, developed by Flint Hills Resources and Integral Engineering. The benchmarks assess life safety and environmental performance using PHMSA data, event trees, and population density analysis. The study provides a method to evaluate HVL pipeline reliability, focusing on individual and societal risks.

Read More

Paper No. IPC2024-133954

This paper enhances crack-like anomaly assessment in pipelines by addressing limitations of semi-elliptical crack models, which often underestimate burst pressure. Building on the PRCI MAT-8 model, it introduces a procedure to convert non-ideal crack depth profiles into equivalent semi-elliptical cracks with longer estimates. An algorithm aligns over 17,500 ILI profiles to quantify measurement uncertainties, addressing gaps in vendor specifications. This approach maintains a conservative bias while ensuring accuracy similar to typical ILI length sizing methods.

Read More