Quantitative Evaluation of Reinforced Concrete Slab Bridges Using a Novel Health Index and LSTM-Based Deterioration Models

Abstract

The Health Index (HI) serves as an essential tool for assessing the structural and functional condition of bridges, calculated based on the condition of structural components and the serviceability of the bridge. Its primary purpose is to identify the most deteriorated structures in an asset inventory and prioritize those in most urgent need of repair. However, a frequently cited issue is the lack of accurate and objective data, with the determination of the HI often being heavily reliant on expert opinions and engineering judgment. Furthermore, the HI systems used in most countries are dependent on the current state of bridge components, making it challenging to use as a proactive indicator for factors such as the rate of bridge aging. To address this issue, this study introduces a novel HI as a quantitative evaluation metric for reinforced concrete slab bridges and details the process of deriving the HI based on deterioration models. The deterioration models are derived by preprocessing the deterioration data of reinforced concrete (RC) slab bridges, wherein the relationship between time and deterioration is directly employed for training a long short-term memory model. The HI was validated through a case study involving six RC slab bridges, wherein accuracies of >93% were achieved, confirming that the proposed quantitative evaluation methodology can significantly contribute to maintenance decisions for bridges.

Keywords

• bridge maintenance
• artificial intelligence
• performance index
• LSTM
• deterioration model
• bridge evaluation