Research On Prediction Of Concrete Carbonization Based On Support Vector Machine And Genetic Algorithm

Reinforced concrete structure is widely used in the national capital construction because it has good fire resistance and integrity and its cost is not very high. But as the reinforced concrete buildings'service life increasing, structural performance continues deteriorating, durability issues is increasingly outstanding, and draws more and more attention, so it is very urgent to solve durability issues. Concrete'durability is the characteristic of itself working ability in usual circumstance, using circumstance and inner factor, or ability of resisting outer circumstance and destruction of inner action. So durability assessment and life prediction for the existing reinforced concrete structure provide foundation for existing structures'maintenance, reinforcement or demolition, at the same time the research results can also be designed for the durability of the structure of reference with the great significance.Important parts of durability study is concrete carbonization among factors which effects durability issues, concrete carbonization destroy alkali environment of concrete, lead to steel corrosion and concrete structure invalidation. This thesis used SVM and GA forecasting network model of concrete carbonization depth is establishing to verify the impact of factors of concrete carbonation.The research step of thesis includes:(1)Concrete durability's background significance and research situation is introduced at home and abroad. And this research are analyzed carbonization mechanism, factors which affects durability issues, various theoretical experiences neural network model and mitigation measure of concrete carbonization.(2)The thesis introduces SVM and GA, and mainly describes principle, mathematic model, algorithm of SVM and SVR; theoretical foundation, operation and characteristic of GA.(3)By considering carbonization influence factors which included water-cement ratio cement consumption carbonation time and so on, combining with GA and SVM, comprehensively using their advantage, genetic algorithm optimization SVM prediction parameter, forecasting network model of concrete carbonization depth is established. It show that forecasting model of concrete carbonization depth combines GA and SVM, and has well forecasting effect and recognition precision through analysis the examples'data, compared experience theory model with BP forecasting network model. It is feasible to use the optimized SVM model through the method of GA to predict. This research on prediction of concrete carbonization based on support vector machine and genetic algorithm has a very good value.