A novel stacking ensemble learner for predicting residual strength of corroded pipelines

npj Materials Degradation, Год журнала: 2024, Номер 8(1)

Опубликована: Авг. 26, 2024

Abstract Accurately assessing the residual strength of corroded oil and gas pipelines is crucial for ensuring their safe stable operation. Machine learning techniques have shown promise in addressing this challenge due to ability handle complex, non-linear relationships data. Unlike previous studies that primarily focused on enhancing prediction accuracy through optimization single models, work shifts emphasis a different approach: stacking ensemble learning. This study applies model composed seven base learners three meta-learners predict using dataset 453 instances. Automated hyperparameter tuning libraries are utilized search optimal hyperparameters. By evaluating various combinations meta-learners, configuration was determined. The results demonstrate model, k-nearest neighbors as meta-learner alongside learners, delivers best predictive performance, with coefficient determination 0.959. Compared individual also significantly improves generalization performance. However, model’s effectiveness low-strength limited small sample size. Furthermore, incorporating original features into second-layer did not enhance likely because first-layer had already extracted most critical features. Given marginal contribution accuracy, offers novel perspective improving findings important practical implications integrity assessment pipelines.

Язык: Английский

---

Online fault detection and localization of multiple oil pipeline leaks using model-based residual generation and friction identification

International Journal of Dynamics and Control, Год журнала: 2024, Номер 12(8), С. 2615 - 2628

Опубликована: Фев. 16, 2024

Язык: Английский

---

Single and multiphase flow leak detection in onshore/offshore pipelines and subsurface sequestration sites: An overview

Journal of Loss Prevention in the Process Industries, Год журнала: 2024, Номер 90, С. 105327 - 105327

Опубликована: Май 13, 2024

Leaks may occur in existing pipelines, even when designed with quality construction and appropriate regulations. The economic impact of oil spills natural gas dispersion from leaks can be huge. Failure to detect pipeline promptly will have an adverse on life, the economy, environment, corporate reputation. Therefore, early detection leaks, their location, size high sensitivity reliability are important for efficient hydrocarbon transportation through a pipeline, both onshore offshore applications. Although several studies been conducted leak using various techniques, recent literature that comprehensively investigates summarizes different multiphase techniques could not found. this paper provides comprehensive review wellbores, subsurface sequestration wells. This is done by studying flow Computational Fluid Dynamics (CFD), Mechanistic, Machine Learning models, digital twin as well sub-surface sites. A investigation revealed few related integrated experiments, computational fluid dynamics, mechanistic implementing extended real-time transient monitoring machine learning. type systematic deemed more useful field Furthermore, new set recommendations provided last section which shows how experimental, mechanistic, CFD simulation data used drive statistical approach based modern deep learning techniques. allows precise understanding events such size, orientation leak, without sending remotely operated underwater vehicle or aircraft scan whole ocean.

Язык: Английский

---

Design and Study of EMAT-MFL Based Hybrid Sensor for Defect Detection in Ferromagnetic Materials

IEEE Transactions on Instrumentation and Measurement, Год журнала: 2025, Номер 74, С. 1 - 8

Опубликована: Янв. 1, 2025

The single nondestructive testing (NDT) techniques face the problem of low resolution when simultaneously detecting different types and locations defects. Composite inspection methods, which combine NDT techniques, have been widely studied due to their high defect detection accuracy. However, most existing composite methods require complex sensor systems or rely on signal-processing algorithms. Therefore, this article proposes a novel hybrid based electromagnetic acoustic transducer (EMAT) magnetic flux leakage (MFL) mechanism, allows accurately detect defects in ferromagnetic materials. This employs straightforward EMAT-MFL configuration, only requires permanent magnet generate requisite field for EMAT MFL components. In addition, it adopts unique orthogonal butterfly coil design cracks. obtained results show significant frequency difference between signals, demonstrates independence lack interference. eliminates potential issue signal aliasing decoupling. can fully use advantages technologies cracks top bottom surfaces Furthermore, wall thinning be detected with maximum error 4.48%. feasible approach miniaturizing sensors increasing performance.

Язык: Английский

---

Enhancing Pipeline Reliability: A Structural Integrity Management Approach Using Minimal Cut Set Method and Importance Measures

IntechOpen eBooks, Год журнала: 2025, Номер unknown

Опубликована: Янв. 10, 2025

Gas pipelines are fundamental structures for transporting energy resources. Their integrity is constantly threatened by failures caused potential punctures or ruptures, leading to gas releases, which can have significant consequences the installation, people, and environment. Various methodologies been proposed improve Pipeline Structural Integrity Management (PSIM) processes. In this work, a model estimating probability of release failure using Quantitative Fault Tree Analysis (QFTA) approach. The Minimum Cut Set (MCS) technique applied along with assessment Importance Measures (IM) provide an accurate estimation rate (λ) identification most critical basic events. This information be used support actions in Risk-Based Inspection (RBI) Reliability-Centered Maintenance (RCM) eliminate, control, mitigate risks. was validated comparing results obtained through Monte Carlo Simulation data from official databases pipeline incidents/accidents similar models published literature. proved capable accurately (λ), closely matching database values more convergent than those achieved reference study also provides guidelines correct effective application PSIM routines.

Язык: Английский

---

The application of structured light for external subsea pipeline inspection based on the underwater dry cabin

Applied Ocean Research, Год журнала: 2025, Номер 155, С. 104431 - 104431

Опубликована: Янв. 22, 2025

Язык: Английский

---

Application of machine learning to leakage detection of fluid pipelines in recent years: A review and prospect

Measurement, Год журнала: 2025, Номер unknown, С. 116857 - 116857

Опубликована: Янв. 1, 2025

Язык: Английский

---

Zero-shot pipeline fault detection using percussion method and multi-attribute learning model

Mechanical Systems and Signal Processing, Год журнала: 2025, Номер 228, С. 112427 - 112427

Опубликована: Фев. 6, 2025

Язык: Английский

---

Analyzing the Innovation Progress in Global Oil and Gas Pipeline Transportation

Journal of Pipeline Systems Engineering and Practice, Год журнала: 2025, Номер 16(2)

Опубликована: Фев. 8, 2025

Язык: Английский

---

On explosion limits of hydrogen–oxygen mixtures with a catalytic platinum surface

Fuel, Год журнала: 2025, Номер 391, С. 134773 - 134773

Опубликована: Фев. 21, 2025

Язык: Английский

---