Pipeline corrosion is a significant challenge in oil and gas transportation, leading to economic losses and environmental hazards. Traditional detection methods are time-consuming and labor-intensive, necessitating the development of more efficient prediction models. This study introduces a hybrid model that combines nonlinear feature expansion (NLFE) and Northern Goshawk optimization (NGO) with an extreme learning machine (ELM) to predict the corrosion rate of natural gas pipelines. This model addresses the limitations of existing methods by enhancing data processing capabilities and improving prediction accuracy. NLFE captures feature relationships within the data, while NGO optimizes ELM by avoiding local minima and premature convergence. This hybrid method was validated using a dataset from a pipeline in Mexico, which contains 106 samples and 10 influencing factors, including oxidation-reduction potential, pipeline operation time, soil pH, grounding potential, and soil resistivity. The results indicate that the NLFE-NGO-ELM model outperforms traditional methods regarding prediction accuracy and robustness. Through an innovative combination of advanced feature augmentation and optimization techniques, this model effectively reduces the impact of noise and redundant information, capturing the intrinsic characteristics of the original data. The contributions of this study include addressing the sample size limitation through nonlinear feature expansion, optimizing the simulation with NGO, and validating the model’s feasibility using various performance metrics and Taylor diagrams. The NLFE-NGO-ELM hybrid model provides a promising solution for accurately predicting natural gas pipeline corrosion rates, enhancing pipeline safety and operational efficiency. This study lays the foundation for future research and practical applications in pipeline integrity management.

You do not currently have access to this content.