Research On Concrete Durability Based On Data Mining

In the process of using concrete,the durability of the concrete is often insufficient,which seriously threatens the safety of the building structure.Therefore,research on the durability of concrete has important theoretical and practical significance.This article first analyzes and studies the main influencing factors of concrete durability from the perspective of concrete raw materials and mix ratio.Based on the main influencing factors,the main indicators of concrete durability are predicted;on this basis,the optimization design of the mix ratio for concrete durability and other multiple objectives is studied;the durability and life of reinforced concrete are analyzed and predicted for chloride ion penetration.This paper studies the durability of concrete from three aspects: durability prediction,mix proportion optimization and life prediction.Firstly,based on the analysis of the influence mechanism of raw materials on concrete durability,the index system of durability factors is constructed,and the random forest algorithm combined with backward elimination method is introduced to eliminate the unimportant indexes.The optimal index system is obtained,in which there are 9 characteristic indexes.Frost resistance and impermeability are selected as two evaluation indexes of concrete durability.In view of the complex and nonlinear relationship between durability and characteristic index,support vector machine algorithm is introduced to predict the frost resistance and impermeability durability of concrete,and a prediction model with high accuracy and good generalization performance is obtained.On this basis,a new idea of combining support vector machine and genetic algorithm is proposed for mix proportion optimization.The support vector machine prediction model is used to replace the traditional function,and it is used as the fitness function of genetic algorithm.Taking anti-freezing,impermeability and cost as optimization objectives,the multi-objective optimization of genetic algorithm is used to obtain the mix ratio to meet the engineering practice.In order to predict and evaluate the service life of concrete,a bridge of Songtong project is taken as the object.A life prediction model based on chloride ion diffusion theory and suitable for salt freezing environment is selected,and convolution neural network is introduced to accurately predict chloride ion permeability coefficient.Combined with the life model,the service life of the bridge is predicted.Some measures to improve the service life of concrete are put forward.