Study on carbonation life prediction of reinforced concrete structures based on improved support vector regression

The accurate prediction of the carbonation life of reinforced concrete structures can effectively prevent and avoid the risk of reinforced concrete structures. This paper proposes a carbonation life prediction method of reinforced concrete structures based on improved support vector regression. Firstly, 19 prediction model inputs are selected according to the analytic hierarchy process, mainly including cement content, water cement ratio, aggregate type, aggregate size, relative humidity, temperature, concrete vibration time, etc. Secondly, genetic algorithm is used to improve the support vector regression method. Finally, the carbonisation life prediction model of reinforced concrete structure is built and trained through the improved support vector regression, and the carbonisation life of reinforced concrete structure is predicted according to the trained model. The experimental results show that the proposed method has a smaller absolute error of only 0.222, the maximum RMSE value is only 0.336, and the fitting degree is closer to 1.