Font Size: a A A

Research On The Models Of Water-to-Cement Ratio,Void Status And Strength Based On Neural Network

Posted on:2019-11-17Degree:MasterType:Thesis
Country:ChinaCandidate:Y Q LuoFull Text:PDF
GTID:2381330602968485Subject:Engineering
Abstract/Summary:PDF Full Text Request
The relationship between water-to-cement ratio,void status and strength is the fundamental problem of concrete as engineering material,and also the main line in the history of concrete research.However,concrete is a multi-layered,heterogeneous polyphase and compound material.With the development of science and technology,the addition of mineral admixtures and chemical admixtures has made the development of concrete from the traditional four-component to the modern six,also a more complex nonlinear relationship between its mesostructure and macroscopic properties.In addition,as a kind of new permeable base material,the mix ratio design of porous concrete lacks convenient and effective guidance.BP neural network models are designed to analyze and solve the problem,the applicability of classical models in modem times and the development law of void status and strength.At last,the BP model is applied to the mix ratio design of porous concrete,which have strong theoretical and practical significance.The main conclusions are as follows:A set of weighting scheme based on data year and source is proposed.Based on this,a sample database is established and randomly divided into training,verification and test groups according to a certain proportion.Six BP network models of water-to-cement ratio,void status and strength are trained and verified.And a virtual laboratory has eonstructed based on this.Based on the virtual laboratory,the applicability of classical model of concrete are reviewed,the results show that the classical models describing the relationship between porosity and strength,such as Balshin power formula,Hasselman linear formula,Ryshkewitch exponent formula,etc,shows a good correlation only at low porosity(p?9%),in which the expression of Ryshkewitch exponential model is 44.23e-7·39p,and the coefficient of correlation is 0.9339.the concrete strength model such as Abrams formula,Bolomey formula and its improved formula show a good correlation in four-component concrete,but not well in modem six-component concrete(mixed with superplasticizers and mineral admixtures).In addition,a series of laws on the water-to-cement ratio,void status and strength are obtained under the deduction of the virtual laboratory.A new model is obtained by combining the BP model of water-to-cement ratio and strength of porous concrete and the BP model of water-to-cement ratio and void status of porous concrete,it is applied to the mix ratio design of porous concrete permeable base of a high grade Irighway and achieves ideal results.
Keywords/Search Tags:Cement concrete, strength, void status, water-to-cement ratio, neural network, mix ratio design
PDF Full Text Request
Related items