In this paper, neural network technique(NNT) is applied to estimation of chloride diffusion coefficient, which has been used for strength evaluation and concrete mixing design. Rapid chloride diffusion test is performed for concrete with 3 different w/c ratios (0.37, 0.42, and 0.47) and various mineral admixtures such as GGBFS, FA and SF. Regarding 120 obtained diffusion coefficients, NNT is applied and diffusion coefficients are simulated. Utilizing the technique, chloride behavior in concrete is evaluated through FE model based on Multi-Component Hydration Model and Micro Pore Structure Formation Model. The simulated results are verified with the previous test results which have been exposed to sea water for 6 months.

• 권성준(한남대학교 건설시스템공학과 교수) | Kwon, Seung-Jun
• 이학수(한남대학교 건설시스템공학과 교수) | Lee Hack-Soo
• 김성준(한남대학교 건설시스템공학과 석사과정) | Kim Seong-Jun