Related Experiment Video
Updated: May 28, 2026
07:12
Determining Tribocorrosion Rate and Wear-Corrosion Synergy of Bulk and Thin Film Aluminum Alloys
Published on: September 11, 2018
Machine Learning-Based Pitting Rate Classification and Prediction for 316L Stainless Steel in NaClO3 and NaCl
Cheng Zhang1, Jiaxin Yao1, Zhe Zhang1,2
1School of Chemical Engineering and Technology, Tianjin University, Tianjin 300350, China.
Materials (Basel, Switzerland)
|May 27, 2026
Summary
This study introduces a machine learning approach using a Feedforward Neural Network (FFNN) to classify pitting severity grades in 316L stainless steel. The method effectively addresses data imbalance for improved prediction in corrosive industrial environments.
Area of Science:
- Materials Science
- Corrosion Engineering
- Computational Materials Science
Background:
- 316L stainless steel is crucial for hydrogen production but susceptible to pitting corrosion in high-temperature chloride environments.
- Sodium chlorate (NaClO3), an electrolysis byproduct, complicates pitting corrosion behavior.
- Accurate prediction of pitting severity is vital for operational safety in industries using 316L steel.
Purpose of the Study:
- To develop a machine learning model for classifying pitting severity grades of 316L stainless steel.
- To address challenges of small sample sizes and imbalanced data in corrosion prediction.
- To provide a reliable method for evaluating pitting corrosion risks.
Main Methods:
- Utilized experimental pitting data of 316L stainless steel.
- Employed Adaptive Synthetic Sampling (ADASYN) to overcome data imbalance.
- Developed and evaluated a Feedforward Neural Network (FFNN) for pitting rate classification.
Main Results:
- The FFNN model demonstrated effective classification of pitting severity grades.
- ADASYN significantly improved the model's performance by mitigating data imbalance.
- The developed model showed feasibility for predictive evaluation of 316L stainless steel pitting corrosion.
Conclusions:
- The proposed FFNN model, enhanced with ADASYN, offers a viable solution for pitting severity grading.
- This approach provides a novel method for predicting and managing corrosion risks in industrial settings.
- The study validates the use of machine learning for classifying material degradation states.
