Time Series Prediction And Analysis Of Carbonation Depth Of Concrete

The carbonization of reinforced concrete is key point in the study of the durability,and accurate prediction of carbonation depth has guiding significance to the life prediction and durability design of concrete structure.This paper summarizes the mechanism,the influence factors and the prediction models of carbonation,points out the problems in the existing models and analyses the forecasting results of existing model. After that a method of predicting the carbonation depth using times series model is proposed.It attributes all the influencing factors of concrete carbonation to time factor,and using the time series model established by the measured value to forecast the carbonation depth in the future.This method based on bayesian theory combining prior information and sample information to establish the time series model, and with the increasing sample size of measured carbonation depth,the model can be updated.The updated model is more accord with the rule of the concrete carbonization,so its predicting results are more accurate.In order to verify the practicability of the method,this paper starts research from the following two aspects:(1)Firstly prove the feasibility of this method theoretically.Using the existing model obtain a set of carbonation depth values to establish time series model,forecast the carbonation depth in the future time with this model and compare the prediction results and the theoretical value.(2)Design accelerated carbonation test to verify the applicability of the method. In order to get a set of sample information reflects the natural law of carbonation,this paper designed three groups of concrete specimen for conducting accelerated carbonation test, using times series model to analysing the test results.Research and analysis show that the established time series model is simple, resulting in good convergence, and can be updated with the measured carbonation depth. So it has higher precision of prediction.