Study On Properties And Proportion Of Geomechanical Model Test Materials

In the geomechanical model test,massive tests are often required,and the constant adjustment of the ratio leads to the similar model after molding to meet the test requirements to obtain a rapid proportioning model of similar materials,which not only costs a lot of money,but also has a long experimental period.The similar material was prepared by the orthogonal experience design,the density,cohesion,internal friction angle,elastic modulus,compressive strength and tensile strength of the similar materials were obtained by weighing,direct shear test,uniaxial compression deformation test and split test.Through the application of range analysis and analysis of variance,the sensitivity and significance of each factor in the above physical and mechanical properties are obtained.It is necessary to conduct a more in-depth study of the performance law of similar model materials to maximize the actual engineering value.It is better not only to obtain the importance of the raw material for each physical and mechanical parameter,but also to obtain an empirical formula that is more accurate and widely used.A mathematical model was constructed by using multiple linear regression analysis of the experimental data by using test datas.In the multivariate regression analysis,factors A、B and C were selected as X₁、X₂、X₃density,elastic modulus respectively,and compressive strength were selected as Y₁、Y₂andY₃.Through multiple regression analysis of the data obtained from the test,three empirical formulas for X and Y were obtained,and the required amounts of various raw materials can be calculated based on arbitrary density,elastic modulus,and compressive strength.In order to improve the accuracy and rationality of the model,the BP neural network optimized by genetic algorithm is used to establish the prediction model.By comparing the prediction effects of the two models with the linear regression mathematical model and the BP neural network model,the accuracy of the BP neural network prediction model optimized by genetic algorithm is more accurate.In the future,this method can be used preferentially in the processing of data problems of this kind.At the same time,great attention should be paid to the standardization and procedural treatment of methods in the process of experimental research,which can avoid the deviation of test results caused by different research units,provide a unified standard for similar tests in the future,and enable the test success to give full play to the actual engineering value.By analyzing the different curing conditions and test methods of similar materials,it is better to set the curing time to about 48h,and it is more reasonable and accurate to set the curing temperature at normal temperature.The compressive strength and elastic modulus of similar materials increase with the increase of loading speed,and the tensile strength decreases with the increase of loading speed.The loading speed should be selected strictly according to the relevant specifications in different tests,making the test conclusion scientific and feasible.