Various models have been proposed to describe the swelling behavior of buffer in high level waster repository. One of the most notable models, the Barcelona Basic Model (BBM), is a mechanical model that simulates the behavior of unsaturated ground and is widely applied to soils that undergo large expansion due to water. Among the BBM parameters of Kyeongju bentonite, which is found in Korea, there are no experimental data for parameters that describe the unsaturated state. Such hydromechanical properties should be characterized through experimental programs. However, such experiments are highly complicated and require long periods of time to produce an unsaturated state through different methods according to the suction range. Although there are several studies in which geotechnical parameters were obtained through a back analysis instead of direct experiments, few studies have employed machine learning methods for the identification of geotechnical parameters. In this study, instead of direct experiments, the results of a relatively simple swelling pressure experiment was compared to the numerical analysis results to propose a method of determining some of BBM parameters. Influential factors were identified by a sensitivity analysis and the values of the factors were estimated using an artificial neural network and optimization method. The obtained parameters were applied to the numerical model to estimate the swelling pressure growth, which was subsequently compared to the experimental value. As a result, it was found that there was no significant difference between the two swelling values.

• Minseop Kim(Korea Atomic Energy Research Institute, 111, Daedeok-daero 989beon-gil, Yuseong-gu, Daejeon) Corresponding author
• Seok Yoon(Korea Atomic Energy Research Institute, 111, Daedeok-daero 989beon-gil, Yuseong-gu, Daejeon)
• Seungrae Lee(Korea Advanced Institute of Science and Technology, 291, Daehak-ro, Yuseong-gu, Daejeon)