Average Corrosion Rate Prediction Model for Buried Oil and Gas Pipelines Based on SSA-LightGBM

Average Corrosion Rate Prediction Model for Buried Oil and Gas Pipelines Based on SSA-LightGBM

Corrosion represents a major cause of damage and leakage in oil and gas pipelines, making accurate corrosion rate prediction critical for operational safety. This study establishes an average corrosion rate prediction model using the Sparrow Search Algorithm-optimized Light Gradient Boosting Machine...

Full description

Saved in:
Bibliographic Details
Main Authors: Weigang Fu, Haitao Wang, Kuankuan Zhang, Xia Wang, Kunlun Chen, Chunmei Sun, Zhengwei Wang, Liuyang Song, Niannian Wang
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Average corrosion rate
LightGBM
machine learning optimization
oil and gas pipeline
sparrow search algorithm
SPDA
Online Access:https://ieeexplore.ieee.org/document/11029024/
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Corrosion represents a major cause of damage and leakage in oil and gas pipelines, making accurate corrosion rate prediction critical for operational safety. This study establishes an average corrosion rate prediction model using the Sparrow Search Algorithm-optimized Light Gradient Boosting Machine (SSA-LightGBM). To enhance model accuracy, Stochastic Perturbation Data Augmentation (SPDA) was applied to expand the original dataset from 60 to 300 samples. The optimized SSA-LightGBM model was then trained, and its performance was rigorously evaluated through error analysis and stability assessment. Results demonstrate that the SSA-LightGBM model achieves exceptional accuracy and stability, with an R2 value of 0.9813 and stability index (Pstd) of 1.1701. Comparative analysis against LightGBM, GBRT, CatBoost, SSA-GBRT, and SSA-CatBoost models confirms that SSA-LightGBM delivers superior predictive performance in both accuracy and stability metrics.
ISSN:2169-3536

Similar Items