Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Jul 4;18(13):3177.
doi: 10.3390/ma18133177.

Prediction of Failure Pressure of Sulfur-Corrosion-Defective Pipelines Based on GABP Neural Networks

Li Zhu  1 , Yi Xia  1   2 , Bin Jia  1 , Jingyang Ma  1
Affiliations

Prediction of Failure Pressure of Sulfur-Corrosion-Defective Pipelines Based on GABP Neural Networks

Li Zhu et al. Materials (Basel). .

Abstract

This study systematically investigates the degradation and failure prediction of pipeline materials in sulfur-containing environments, with a particular focus on X52 pipeline steel exposed to high-sulfur environments. Through uniaxial tensile tests to assess mechanical properties, it was found that despite surface corrosion and a reduction in overall structural load-bearing capacity, the intrinsic mechanical properties of X52 steel did not exhibit significant degradation and remained within standard ranges. The Johnson-Cook constitutive model was developed to accurately capture the material's plastic behavior. Subsequently, a genetic algorithm-optimized backpropagation (GABP) neural network was employed to predict the failure pressure of defective pipelines and the corrosion rate in acidic environments, with prediction errors controlled within 5%. By integrating the GABP model with NACE standard methods, a framework for predicting the remaining service life for in-service pipelines operating in sour environments was established. This method provides a novel and reliable approach for pipeline integrity assessment, demonstrating significantly higher accuracy than traditional empirical models and finite element analysis.

Keywords: defective pipelines; failure prediction; neural networks; sulfur-containing corrosion.

PubMed Disclaimer

Conflict of interest statement

Author Yi Xia was employed by the China Electronic System Engineering Co., Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Figures

Figure 1
Figure 1
Uniaxial tensile test: (a) schematic diagram of specimen sampling; (b) geometry of the specimen; (c) final machined specimen; (d) tensile test setup.
Figure 2
Figure 2
Stress–strain curve of tensile test specimen.
Figure 3
Figure 3
Finite element model of tensile specimen.
Figure 4
Figure 4
X52 hardening curve.
Figure 5
Figure 5
Comparison of finite element simulation results based on the Johnson–Cook constitutive model and tensile test data.
Figure 6
Figure 6
Finite element model of defective pipes.
Figure 7
Figure 7
Comparison of blasting test results and finite element analysis results: (a) ITDS-2; (b) ITDS-5; (c) ITDS-6.
Figure 8
Figure 8
BP neural network structure diagram.
Figure 9
Figure 9
Flowchart of genetic algorithm optimization of BP neural network.
Figure 10
Figure 10
The BP neural network structure for predicting failure pressure of defective pipelines.
Figure 11
Figure 11
The MSE values associated with varying numbers of hidden layer neurons in the neural network model employed for failure pressure prediction.
Figure 12
Figure 12
Genetic algorithm optimization of BP neural network convergence curve.
Figure 13
Figure 13
Pearson correlation coefficients corresponding to 50 independent training runs.
Figure 14
Figure 14
Fitting effect diagram of BP neural network: (a) training set fitting effect diagram; (b) validation set fitting effect diagram; (c) test set fitting effect diagram; (d) overall fitting effect diagram.
Figure 15
Figure 15
Fitting effect diagram of GABP neural network: (a) training set fitting effect diagram; (b) validation set fitting effect diagram; (c) test set fitting effect diagram; (d) overall fitting effect diagram.
Figure 16
Figure 16
The impact of the weighted summation of individual parameter weights on the prediction of failure pressure in the neural network model.
Figure 17
Figure 17
Comparison of GABP and BP neural network prediction results.
Figure 18
Figure 18
Comparison of GABP and BP neural network prediction errors.
Figure 19
Figure 19
Comparison of failure stresses predicted by various evaluation methods and the GABP network.
Figure 20
Figure 20
BP neural network structure diagram for predicting sulfur corrosion rates.
Figure 21
Figure 21
The MSE values associated with varying numbers of hidden layer neurons in the neural network model developed for predicting the sulfur corrosion rate.
Figure 22
Figure 22
GABP network prediction corrosion rate fitting diagram.
Figure 23
Figure 23
The impact of the weighted summation of individual parameter weights on the prediction of sulfur corrosion rate within the neural network model.
Figure 24
Figure 24
Prediction process for residual bearing capacity of pipelines with sulfur corrosion defects.
See this image and copyright information in PMC

References

    1. Rubaiee S. High sour natural gas dehydration treatment through low temperature technique: Process simulation, modeling and optimization. Chemosphere. 2023;320:138076. doi: 10.1016/j.chemosphere.2023.138076. - DOI - PubMed
    1. Vosikovsky O., Macecek M., Ross D.J. Allowable defect sizes in a sour crude oil pipeline for corrosion fatigue conditions. Int. J. Press. Vessel. Pip. 1983;13:197–226. doi: 10.1016/0308-0161(83)90026-1. - DOI
    1. Li X., Ma X., Zhang J., Akiyama E., Wang Y., Song X. Review of Hydrogen Embrittlement in Metals: Hydrogen Diffusion, Hydrogen Characterization, Hydrogen Embrittlement Mechanism and Prevention. Acta Metall. Sin. (Engl. Lett.) 2020;33:759–773. doi: 10.1007/s40195-020-01039-7. - DOI
    1. Cai L., Bai G., Gao X., Li Y., Hou Y. Experimental investigation on the hydrogen embrittlement characteristics and mechanism of natural gas-hydrogen transportation pipeline steels. IOP Publ. Ltd. 2022;9:046512. doi: 10.1088/2053-1591/ac6654. - DOI
    1. Negi A., Barsoum I., Alfentanil A. Predicting sulfide stress cracking in a sour environment: A phase-field finite element study. Theor. Appl. Fract. Mech. 2023;127:104084. doi: 10.1016/j.tafmec.2023.104084. - DOI

LinkOut - more resources