| Unlined reinforced concrete (RC) shearwalls are often used by the U.S. Department of Energy to house radioactive materials, thus providing a 'tertiary barrier'. When subjected to lateral earthquake loading, these stiff structural members are susceptible to large tensile and shear forces along their base and diagonal planes, resulting in the development of cracks. Following an earthquake, dangerous gases (contained within the tertiary barrier) may leak into the environment through these cracks, lifting pollution to unsafe levels. The gas leakage rate through the damaged wall is therefore of primary concern. In this research, the primary objective is to develop a methodology to analytically predict the leakage rate of air through RC shearwalls, when subjected to a broad range of lateral demands.;The methodology presented includes both experimental and analytical studies. In each, two relations being investigated are: (1) damage versus lateral load, and (2) gas leakage rate versus damage. The relation between damage and lateral load is obtained by loading the model specimens to prescribed demand levels either during physical tests or simulated via finite element (FE) analysis. Gas leakage rate at different damage levels is measured using leakage rate tests in the experiments, or numerically predicted via flow rate (FR) analysis in the analytical studies.;An experimental program testing the gas leakage rate through an RC shearwall, when subjected to lateral demands beyond its normal design basis, is conducted. The experimental setup and procedures, as well as the results obtained are described. These experiments not only contribute additional results to the sparse literature, as well, they can be used to correlate and verify the numerical models used in analytical studies.;Based on the experimental results obtained by the author and those in the literature, a number of correlation studies are performed to acquire guidance for selecting the input parameters to the FE model. This is done by carefully comparing numerical and experimental lateral load-deformation and damage behaviors. Using this validated model, an FE model of the experimental specimen is developed and used to capture the crack strains under prescribed lateral demands. It is found that, although the FE mesh size affects the maximum strain, it will not affect the simulation of flow rate through the shearwall if an interpolation technique is adopted. However, of importance to accurate flow rate simulation is the crack spacing. By extending the theory proposed by Bazant and Oh (1983), average crack spacing is simulated using the crack strains obtained from the FE analysis. Knowing crack strains, one can estimate crack widths and this, coupled with crack length and thickness, provides the necessary characteristics to describe the crack patterns in the wall.;By applying four available leakage rate estimation formulas in the literature to the numerically determined crack patterns, the air leakage rates at different damage levels can be predicted. It is found that two of them (Rizkalla and Suzuki-2's methods) generally provide reasonable results, within the typical design load levels. The formula proposed by Nagano overestimates the flow rate due to the assumption of frictionless flow, and Suzuki-1's method consistently underestimated the flow rates.;Using the validated FE model and the verified leakage rate estimation formula, the flow rate of air is then analytically predicted for a broader range of loading conditions (lateral and axial loading), structural configurations (length, height, thickness, boundary elements, and the steel ratios in the panel and boundary elements), and the concrete strength, at different differential pressures. In total, forty eight FE models are constructed and subjected to a lateral cyclic displacement protocol. The relation between the shearwall lateral drift ratio and the air leakage rate is established. It is found that the influence of differential pressures is decoupled with other parameters. Only one pressure differential condition is necessary to be simulated, while the flow rate at other differential pressures can be obtained via interpolation. Among the parameters considered in this study, the boundary element size have the most significant effect on the flow rate. This work concludes with a design example to demonstrate the usefulness of the normalized flow rate plots developed. |
