The Study Of Normal Section Calculation Models Of Recycled Concrete Based On Ga Optimizing BP Neural Network

This paper is a continued research of"analysis on mechanical properties and microstructure of recycled concrete"which is funded by Natural Science Foundation of Heilongjiang province.The utilize of recycled concrete is at preliminary stage all over the world. At present, recycled concrete is used mostly in bearing components or structure. One reason is that there is still not having uniform calculation method of recycled concrete component. The research of recycled concrete is not further enough which is a serious problem of the application of recycled concrete.This paper analyzes the research of the basic mechanical properties of recycled concrete which was done by scholars all over the world, based of which uniaxial compressive stress-strain model of recycled concrete is discussed.The domestic and foreign research of mechanical properties computation models of of recycled concrete beam is summarized. The calculation models of recycled concrete flexural members which are based on optimized BP neural network are established through the training and test of the sample data.Under different recycle aggregate replacement rates and different strength grades,cracking load, ultimate bearing capacity, maximum crack width and midspan deflection are calculated by these models,and the Standard calculation method is amended by these results.