LAUSR

Abstract Uncertainties significantly influence the durability-related experiments and field studies, which can hardly be addressed by deterministic approaches. This work aims at developing a stochastic numerical framework to disentangle the influence of material uncertainty on the case of leaching. To ensure the robustness of the numerical framework, a realistic stochastic reactive-transportation model is developed, which consists of a novel sampling algorithm and a comprehensive deterministic model. By using the proposed sampling algorithm, a more effective and efficient sampling process can be achieved without compromising the randomness of the uncertain properties. Besides, realistic mechanisms of leaching are considered by the deterministic model, including the simultaneous processes of ionic transportation, chemical reactions and material degradation. By performing the stochastic leaching analyses, numerical results suggest the overwhelming influence of the physical uncertainty on long-term leaching, while the impact of chemical uncertainty is more evident in terms of short-term leaching. It is also revealed that the root-time relation as determined from short-term experiments is inappropriate for long-term predictions. Thus, a modified relation is developed based on the stochastic leaching analysis, which generates accurate predictions for both the short-term and long-term leaching.