Numerical simulation of capillary flow and curing behaviour of healing agent in encapsulation-based self-healing concrete

Abstract

A self-healing concrete has emerged as a potential solution for tackling cracking issues in concrete. Self-healing works through the infiltration of healing agents into the cracks, followed by the curing process that prevents further crack penetration in the concrete matrix. As far as the literature is concerned, the effect of the curing reaction on its rheological properties has not been addressed adequately. In addition, a fluid flow model considering the capillary effect and curing reaction has not been established for flow behaviours within discrete cracks in the encapsulation-based self-healing concrete. Therefore, in this study, a coupled fluid flow and curing reaction model was proposed to simulate the concrete encapsulation system’s mechanics better. The healing agent flow was modelled using the Volume-of-Fluid (VOF) method. This study proposed using the viscosity function to describe the curing effect using the Castro-Macosko model. The dynamic mechanical analysis experiment cured and changed the cyanoacrylate's rheological properties. The fluid flow and curing reaction models were coupled in ANSYS Fluent in the form of self-developed user-defined functions. Parametric studies were carried out to determine the influence of healing agent rheological properties (surface tension, contact angle and viscosity) and crack geometries (planar, inclined and tapered) on the healing efficiency. The coupled model was validated against available experiment results and the model's capability to predict the healing agent's flow accurately and the curing process was shown. For flows in small cracks driven by capillary action, the simulated VOF outcomes with constant contact angles were in poor agreement with the experiment. The simulation results showed a better prediction of the capillary flow with the use of dynamic contact angles (DCA). For example, when validated against the modified Lucas-Washburn equation (LWE), the VOF predictions considering the velocity-dependent DCA have mean absolute percentage errors of between 3.1 – 5.3%, much lower than that of classical LWE with errors between 17.0 – 42.9%. The results indicated that a DCA influences the initial speed of the capillary flow and plays a vital role in the healing efficiency of fast-curing healing agents. Due to the curing reaction, the increasing viscosity arrests the capillary flow of the healing agent in a small discrete crack. DCA and viscosity control the infiltration speed of capillary flow via frictional dissipation and flow resistance, respectively. However, they do not affect the final equilibrium height in capillary rise. A higher frictional coefficient in the DCA model decreases the infiltration speed at the initial state of the capillary rise. In said capillary flow, the infiltration length of the healing agent depends on the capillary pressure, which is strongly influenced by the surface tension, equilibrium contact angle and crack widths. Based on the Young-Laplace equation, the capillary pressure is directly proportional to the surface tension force and inversely proportional to the crack width. A lower contact angle indicates good wettability and provides faster liquid spreading on a surface. Overall, this study has provided a new coupled self-healing model for predicting the transport and curing processes in encapsulation-based self-healing systems in concrete. The model can provide a better understanding of flow mechanisms and serves as a sound basis for future researchers to design a more efficient concrete self-healing system.