Session: 02-06-01 Data-driven Models for Marine Structures

Submission Number: 156548

Experimentation and a Data-Driven Machine Learning Approach for Strength Prediction of Non-Uniformly Corroded Rebar 

Accurate prediction of the residual load-carrying capacity of corroded reinforced concrete (RC) structures is essential for assessing their service life and durability, particularly in aggressive environments. Traditional reliability analyses often rely on parameters such as uniform corrosion rates and mechanical strength of reinforcing steel, which are typically related to time or corrosion progression. To address this, a series of accelerated electrolytic corrosion experiments, both with and without additional loading, were conducted to investigate the degradation of rebars in RC structures. A total of 36 samples with varying degrees of corrosion were analyzed, focusing on current density, corrosion rate, morphology, ductility, compressive strength, and critical corrosion depth. Results revealed significant non-uniformity in corrosion depth across cross-sections and along rebar length, accompanied by a decline in concrete compressive strength as corrosion progressed. Monotonic tensile testing further demonstrated a marked reduction in yield and ultimate strengths, emphasizing the detrimental effects of corrosion on structural integrity.

To expand the analysis, a dataset of 1,423 samples, including natural exposure and accelerated corrosion results from both literature and the current study, was utilized to assess the effects corrosion on corroded rebar strength. Three machine learning methods were used to evaluate the influence of various factors on rebar strength degradation. Feature importance analysis was conducted using the Shapley Additive Explanation (SHAP) algorithm, quantified each input parameter's contribution to strength loss. Based on this analysis, a simplified predictive formula was developed to quantify strength loss based on three key factors. This newly user-friendly expression serves as an effective tool for forecasting rebar strength degradation, providing valuable insights for the maintenance and assessment of aging infrastructure in harsh environmental conditions.

Presenting Author: Bing Chen Shandong Jiaotong University

Presenting Author Biography: Bing Chen, ph.D.
Work:
College of Transportation and Civil Engineering, Shandong Jiaotong University.
Primary Research Areas:
Reliability of Coastal and Marine Engineering Structures
Offshore hydrodynamics and sediment transport