@Article{sdhm.2026.081186,
AUTHOR = {Zhichao Liu, Jun Zhang, Dongling Yu, Libing Jin, Bingquan Song},
TITLE = {Research on the Chloride Ion Penetration Resistance of Manufactured Sand Concrete Based on WOA-Adam Hybrid Optimized BPNN},
JOURNAL = {Structural Durability \& Health Monitoring},
VOLUME = {},
YEAR = {},
NUMBER = {},
PAGES = {{pages}},
URL = {http://www.techscience.com/sdhm/online/detail/27011},
ISSN = {1930-2991},
ABSTRACT = {The chloride ion penetration resistance of manufactured sand concrete (MSC) critically determines the durability of marine concrete structures. However, its accurate prediction is challenging due to high uncertainty from complex influencing factors. To address this, a back-propagation neural network model optimized by a hybrid Whale Optimization Algorithm and Adaptive Moment Estimation strategy (WOA-Adam-BPNN) was developed to predict the electrical flux. The model was trained and tested on 245 experimental datasets covering eight key parameters and validated across four typical mix proportions. Results show that the WOA-Adam hybrid strategy effectively combines global search capability with adaptive convergence, significantly enhancing model performance. The proposed model achieved a mean absolute percentage error (MAPE) of 4.01%, a root mean square error (RMSE) of 59.57, and a coefficient of determination (R<sup>2</sup>) of 0.9879, significantly outperforming both the traditional Adam-BPNN and the genetic algorithm-optimized Adam-BPNN (GA-Adam-BPNN) models and the WOA-XGB hybrid model (WOA-XGB). Moreover, it maintains a stable prediction error within 5% across common engineering parameter ranges. This study provides a reliable reference for the mix proportion design and durability assessment of MSC.},
DOI = {10.32604/sdhm.2026.081186}
}