Marine concrete is often exposed to multiple aggressive ions, so mechanical deterioration cannot be reliably interpreted using single-ion durability concepts. This study investigates ocean-oriented concretes incorporating high contents of mineral admixtures under coupled sulfate/chloride/carbonate/magnesium actions and develops a pore-structure-based D–P dual-parameter framework linking microstructural descriptors to macroscopic peak stress and peak strain. Three binder systems were designed: ordinary Portland cement concrete (OPC), cement–silica fume concrete (CSC, 20% silica fume), and cement–silica fume–fly ash concrete (CSFC, 20% silica fume + 50% fly ash). Specimens were immersed for 12 and 24 months in four representative binary-salt solutions. Porosity evolution and pore-size-class distributions were quantified by low-field NMR, while pore complexity was characterized using multi-scale fractal dimensions. The results show that mineral admixtures generally refine the pore system and improve the integrity of fine pores; CSFC exhibits the most robust microstructural stability across the tested environments, whereas CSC shows a pronounced degradation of fine-pore structure under CE4. A second-order response surface model built on Z-score normalized fractal dimension (D) and porosity (P) achieves reliable predictability for peak strain (R2 = 0. 85) and peak stress (R2 = 0. 79). Global Sobol sensitivity analysis reveals distinct controlling mechanisms: peak strain is predominantly governed by porosity (SP = 85. 9%), whereas peak stress is controlled by the combined effects of porosity, pore complexity, and their interaction (SP = 42. 4%, SD = 19. 8%, S₃ × = 37. 8%). Local sensitivity mapping further identifies high-sensitivity regimes at extreme pore states, providing mechanistic guidance for mixture optimization. Overall, the proposed D–P framework quantitatively bridges pore volume/geometry evolution and mechanical degradation, offering a practical predictive tool for durability-oriented design of marine concretes under multi-ionic attack.
Building similarity graph...
Analyzing shared references across papers
Loading...
Xiu-Cheng Zhang
Fujian Agriculture and Forestry University
Ying Peng (26452)
Fujian Agriculture and Forestry University
Building similarity graph...
Analyzing shared references across papers
Loading...
Zhang et al. (Sat,) studied this question.
synapsesocial.com/papers/69a67eb2f353c071a6f0a207 — DOI: https://doi.org/10.3390/fractalfract10030160