Research On Durability And Reliability Of RC Structures In Chloride Environment

Durability deficiency of RC structures caused by steel corrosion, which is induced by chloride ions penetration, is one of the most important topics in civil engineering field as such cases happened widely and cause big loss in the world. In this dissertation, some part of deterministic models are modified or expanded which can be used to predict RC structure's life corresponding to steel corrosion initiation with experimental study and numerical analysis. Three limit states after corrosion initiation are defined by experimental method. Based on Q-MCMC approach (Quasi Monte Carlo simulation of discrete Markov Chain), a stochastic model is proposed to calculate failure probability corresponding to every limit state of RC structures placed in chloride environment.In most current applications of Fick's second diffusion law, obstruction of steel presence has not been taken into account. In fact, that chloride cannot transport through steel will cause accumulation at front point of steel become rapid which makes steel corrosion initiation of RC structures take place in advance. With two-dimension finite element analysis and nonlinear LSM regression, this obstruction effect is embedded in the analytical solution of Fick's second diffusion law. This modified model can predict time to steel corrosion initiation accurately, conveniently and rapidly.Influences of water to cement ratio (w/c) and cement content on concrete durability and safety performance are analyzed in detail based on a large of experimental database. A CDC (initial chloride effective diffusion coefficient) predicting model and a strength predicting model are built up respectively according to experimental results and these two models will be useful to determine the optimal or minimum cement content on condition of that strength and durability reach the demand level.With electrochemical acceleration method, three limit states of RC structures after reinforcing bar corrosion initiation are defined which are lightdegree corrosion, middle degree corrosion, and serious degree corrosion. When crack width setting as the criterion of corrosion degree, influences of concrete strength, cover thickness, steel spacing and steel location on reinforcement corrosion ratio and cover cracking type are observed and generalized. Definitions of these limit states after corrosion initiation provide evidence for determination of every kind of life end when predicting with the deterministic model or stochastic one.A stochastic process, discrete Markov Chain, is first introduced to describe steel corrosion deterioration process of RC structures which is induced by chloride ions penetration. Adopting quasi-Monte Carlo method for random sampling and based on discrete Markov Chain, a stochastic model is proposed to calculate failure probabilities of RC structures when steel corrosion initiation or subsequent deteriorations are set as limit states respectively. Using this proposed Q-MCMC probabilistic model, calculation efficiency is promoted at large extent.