Controlling And Optimization Research On Concrete Shrinkage Deformation Based On Artificial Neural Network

With the area and size of modern building becoming bigger and bigger, there are many super long structures. The concrete cracks are the main problems of the super long structure, and the shrinkage of concrete is the main reason to cause cracking.Based on these facts, the subject analyzes the mechanism of concrete shrinkage and correctly distinguishes concrete shrinkage types. Simultaneously shrinkage tests of concrete mixing both fly ash and slag powder in natural environments are conducted. According to the results, effects on shrinkage of concrete such as mineral admixture and environmental conditions are analyzed. According to the experimental study, the shrinkage of concrete is predicted by the domestic and foreign models, and the model of concrete shrinkage based on BP neural network is suggested.After compared with the experimental results, the predicting results of the domestic and foreign models are not accurate. Test results of BP neural network model agree well with the experimental results and the network model has considerable sensitivity. The laws of the single factor effect upon shrinkage strain are reflected by the suggested model. Simultaneously the primary and secondary factors and the optimal mix proportion of concrete are successfully predicted by orthogonal numerical simulation method, it can provide a reference for the shrinkage prediction of concrete for the engineering structure.