Research On Non Structural Components' Masonry Mortar Detection Methods Which Is Based On Artificial Neural Network

Based on the research of the part of some domestic and foreign mortar strength detection method,finding that detection methods for bearing structure of mortar strength analysis is relatively mature,for example:the cylinder pressure method,the method of mortar flake and mortar rebound method on the detection of sintered common brick and the sintered porous brick,but non structural components' masonry mortar strength detection is relatively insufficient and the corresponding research results is scarce.Therefore,the research of non structural components' masonry mortar strength detection is particularly important.Making the mortar models of three types of strength grades(M5.0,M7.5,M10.0)for the four-age-period(28days,60 days,90days and 150days)for the detection of the mortar rebound method,penetration resistance method and cylinder compression method.The all materials which are used for the test are from Zhangjiakou area,which are widely applied in engineering projects.Based on standard curing data of mortar block data,compared to the test data of the rebound method,penetration resistance method and cylinder compression method,finding that the cylinder pressure method's data is relative high precision.On this basis,using origin9.0 software based on the principle of least squares method to analyze the test data by means of the form of the five kinds of commonly used functions and establish the rebound method,penetration resistance method and cylinder compression method's formula about non structural components' masonry mortar strength.There are many factors that are uncertain affecting non structural components' masonry mortar strength,and use the existing formula to determine their expression is difficult.but using the artificial neural network which is based on precise mathematical model to solve nonlinear problems is well.so,research on non structural components' masonry mortar detection methods which is based on artificial neural network is raised.The artificial neural network mathematical models which are based on experimental data about the rebound method,penetration resistance method and cylinder compression method has been set up by means of the MATLAB system its own toolbox.By training the testing samples,it is proved that the artificial neural network model can be used to predict the mortar strength.Compared to mortar strength that is predicted by traditional regression algorithm,the mortar strength that is predicted by the artificial neural network models of the non structural components'the rebound method,penetration resistance method and cylinder compression method has relative high precision.