Prediction of moisture migration and pore pressure build-up in concrete at high temperatures

Abstract

Prediction of moisture migration and pore pressure build-up in non-uniformly heated concrete is important for safe operation of concrete containment vessels in nuclear power reactors and for assessing the behaviour of fire-exposed concrete structures. (1) Changes in moisture content distribution in a concrete containment vessel during long-term operation should be investigated, since the durability and radiation shielding ability of concrete are strongly influenced by its moisture content. (2) The pressure build-up in a concrete containment vessel in a postulated accident should be evaluated in order to determine whether a venting system is necessary between liner and concrete to relieve the pore pressure. (3) When concrete is subjected to rapid heating during a fire, the concrete can suffer from spalling due to pressure build-up in the concrete pores. This paper presents a mathematical and computational model for predicting changes in temperature, moisture content and pore pressure in concrete at elevated temperatures. A pair of differential equations for one-dimensional heat and moisture transfer in concrete are derived from the conservation of energy and mass, and take into account the temperature-dependent release of gel water and chemically bound water due to dehydration. These equations are numerically solved by the finite difference method. In the numerical analysis, the pressure, density and dynamic viscosity of water in the concrete pores are calculated explicitly from a set of formulated equations. The numerical analysis results are compared with two different sets of experimental data: (a) long-term (531 days) moisture migration test under a steady-state temperature of 200 °C, and (b) short-term (114 min) pressure build-up test under transient heating. These experiments were performed to investigate the moisture migration and pressure build-up in the concrete wall of a reactor containment vessel at high temperatures. The former experiment simulated the effect of long-term steady-state liner temperature during normal operation, and the latter simulated a situation where an accident resulted in sudden, short-term heating to approximately 400 °C. Finally, concrete spalling is simulated by the numerical analysis; and the results show how the moisture content and pore pressure distributions in concrete exposed to fire change with time and temperature. The numerical analysis can predict the time, position and temperature at which spalling occurs.