| In this paper, using the first-principles method and BP neural network, we study the mechanical properties of kaolinite and relevant application in civil engineering. The main research work and conclusions are as follows:(1)Using the first-principles method, we study the different pressure of the elastic constants, volume, and bond length of kaolinite. The calculated result show that, under low pressure condition, the volume, total energy, bond length have little change, while the elastic constants have considerably large vary before and after structure optimization as increasing pressure values.(2) The impact factors of sedimentation are classified in BP neural network, by training and simulation, the nonlinear relation is established between the influence factor and settlement, the error is very small by simulating between the predicted values and the measured value, which proves that the BP neural network is an effective method to predict the settlement of dilatability kaolinite clay foundation.(3) In order to improve the anti-carbonation property of concrete, metakaolin is added in the concrete carbonization test. The experimental results show that, along with the increasing ratio of metakaolin come to decreasing trend in carbonate depth, concrete carbonation resistance are also improved accordingly, when the total proportions of mineral admixtures dosage is 35%, including 15% metakaolin concrete anti-carbonization ability reaches a maximum value of 38.02%. These results can provide the new guide for improve the anti-carbonation property of concrete. |