Study On Shrinkage Performance Ofcement-based Materials At Lowwater-binder Ratio

The shrinkage performance of concrete is always one of the key and difficult problems in the field of concrete engineering.With the development of water reducing agent and the extensive use of mineral admixtures,the shrinkage and deformation properties of modern concrete are more complex.So far,there still exists controversy about the classification,definition,mechanism,research methods and models of shrinkage.The study in this thesis focused on the cement-based material at low water-binder ratio;and studied the early shrinkage,drying shrinkage and autogenous shrinkage of cement-based materials.The hydration heat of cement-based materials was studied by using TAM air 8-channel calorimeter,and the BP neural network model about autogenous shrinkage was established by using MATLAB software.In this paper,the non-contact displacement sensor was used to measure the early shrinkage.The length method was used to measure the drying shrinkage,and the corrugated tube method was used to measure the autogenous shrinkage.At the same water-binder ratio and the same strength grade,the early shrinkage(0～3d)of reference cement is larger than ordinary portland cement,but the drying shrinkage(3d～)is smaller than ordinary portland cement.When the water-binder is in the range of 0.35～0.25,the early shrinkage increases with the reduction of water-binder ratio,but the drying shrinkage decreases with the reduction of water-binder ratio.The addition of polycarboxylic series of water reducer can increase the early shrinkage and the drying shrinkage.The drying shrinkage is related to the humidity during curing.When the relative humidity(RH)is 60%,the Specimen show a shrinkage trend;when the RH is more than 90%,the Specimen show a slight expansion trend.The autogenous shrinkage of cement-based materials increases with the reduction of water-binder ratio.Polycarboxyllic series of water reducer can reduce the autogenous shrinkage,and the autogenous shrinkage decrease with the increase of the water reducer.But when the water-binder is high,such as 0.40,the addition of polycarboxylic series of water reducer will increase the autogenous shrinkage.When fly ash and mineral powder are added separately and the mixing amount is 20% and when the ratio of fly ash to slag powder is 2:1 to replace 30% cement the best effect of reducing autogenous shrinkage is obtained.When the water-binder ratio is in the range of 0.40～0.35,the hydration heat(0～3d)of cement-based materials decreases with decrease of water-binder ration of cement paste.And the addition of fly ash and mineral powder can reduce the hydration heat,but the effect of fly ash on reducing hydration heat is better than mineral powder.The model of 6-18-18-1 BP neural network was built by using the neural network toolbox of MATLAB.In the network,tansig function was chosen as the transfer function from the input layer to the hidden layer,and purelin function was chosen as the transfer function from the hidden layer to the output layer.The autogenous shrinkage data of 1487 groups were divided into two parts: the training samples and the verification samples.The linear relationship between the output value of the training sample and the actual value is good,and the correlation coefficient is 0.9991.The error range is less than 7% when the trained network model is used to the verification.