SAMURAI - NIMS Researchers Database

HOME > Profile > KATAYAMA, Hideki

External affiliations

  • 東京理科大学 創域理工学部 先端化学科
  • 法政大学 生命科学部 環境応用化学科

Research

Keywords

材料加工・処理

PublicationsNIMS affiliated publications since 2004.

Books
Proceedings
Misc

Society memberships

日本鉄鋼協会, 表面技術協会, 腐食防食学会, Electrochemical Society, 電気化学会, 金属学会, 原子力学会

Research Center for Structural Materials
Title

Corrosion Risk Prediction for Steel Structures Using Hyperspectral Analysis

Keywords

Hyperspectral imaging, Carbon steel, Atmospheric corrosion, Corrosion products, SAM analysis, Corrosion risk prediction

Overview

Preventive maintenance of steel structures requires techniques for assessing both their current corrosion condition and subsequent corrosion progression. In this study, corrosion products formed on carbon steel specimens exposed to ordinary and sheltered outdoor environments at Choshi, Chiba and Miyakojima, Okinawa for up to 2.5 years were identified using hyperspectral measurements and Spectral Angle Mapper (SAM) analysis. Reference spectra of α-FeOOH, β-FeOOH, γ-FeOOH, and Fe₃O₄ were used to evaluate their spatial distributions and surface fractions. Although the surface fractions of γ-FeOOH and Fe₃O₄ showed no correlation with the increase in corrosion loss, γ-FeOOH predominated under ordinary exposure, whereas Fe₃O₄ predominated under sheltered exposure, indicating that these products could help distinguish between exposure environments. In contrast, the surface fractions of α-FeOOH and β-FeOOH showed positive and negative correlations, respectively, with the increase in corrosion loss over the subsequent year. These findings demonstrate the potential of optical information from surface rust for corrosion risk assessment.

Novelty and originality

● Corrosion risk assessment based on rust identification
● Evaluation of the spatial distributions and surface fractions of different corrosion products using hyperspectral data
● Verification of applicability under different outdoor exposure conditions
● Identification of potential indicators of subsequent corrosion progression

Details

image

Carbon steel specimens exposed to ordinary and sheltered environments at Choshi, Chiba, and Miyakojima, Okinawa, for 1.5 years were examined using a hyperspectral camera. SAM analysis was used to visualize the spatial distributions of surface corrosion products. Reflectance spectra were acquired over a wavelength range of 400–1100 nm, and each corrosion product was identified by evaluating spectral similarity at 500 and 760 nm using reference spectra of four known corrosion products. In Fig. 1, α-FeOOH, β-FeOOH, γ-FeOOH, and Fe₃O₄ are represented by blue, yellow, red, and gray, respectively. At both sites, γ-FeOOH was predominantly detected under ordinary exposure, whereas Fe₃O₄ was predominantly detected under sheltered exposure, revealing differences in surface rust associated with the exposure environment. The detection of Fe₃O₄ suggests an association with prolonged wet conditions. The surface fraction of each corrosion product was calculated from the number of pixels assigned to its corresponding color, enabling evaluation of the distributions and fractions of corrosion products that are difficult to distinguish visually.

The surface fraction of each corrosion product was then compared with the increase in corrosion loss over the subsequent year (Figure 2). For example, the increase was calculated as the difference in corrosion loss between specimens exposed for 0.5 and 1.5 years. A one-year comparison interval was used to reduce seasonal effects. While no clear correlations were observed for γ-FeOOH or Fe₃O₄, the surface fraction of α-FeOOH showed a positive correlation with the increase in corrosion loss, whereas that of β-FeOOH showed a negative correlation. These results suggest that the surface fractions of α-FeOOH and β-FeOOH may serve as indicators of subsequent corrosion progression. Further work is needed to improve the identification accuracy of mixed rust and validate the quantitative assessment of surface fractions. Validation across a wider range of environments is also necessary to clarify the relationship between corrosion products and subsequent corrosion progression.

image

Summary

Hyperspectral measurements combined with SAM analysis enabled identification of corrosion products on carbon steel after outdoor exposure and evaluation of their spatial distributions and surface fractions. The surface fractions of α-FeOOH and β-FeOOH showed positive and negative correlations, respectively, with the increase in corrosion loss over the subsequent year. These findings demonstrate the potential of optical information from surface rust to support corrosion risk assessment and preventive maintenance of steel structures.

この機能は所内限定です。
この機能は所内限定です。

▲ Go to the top of this page