Concrete DEM BPM Model Parameter Verification And Its Multi-objective Optimization Algorithm

Rock,ceramic,concrete and other brittle materials play a huge role in dealing with mechanical problems such as deformation and destruction law et al.in all walks of life.The traditional analytical methods are difficult to carry out in-depth study.With the rapid development of computing performance,using discrete element method bonded particle model(Bounded Particle Model,referred to BPM)to simulate and explore the mechanics of brittle bulk materials has become a new way.However,before this method,the pre-meso-parameter of discrete element BPM model must be checked.During the process of checking,there exists plurality of target values interacting with each other.This process need to be bugged and modified repeatedly.The checking process is very complicated.In this project,the application of discrete element method in checking mesoscopic BPM model parameters of concrete,and how the target values for multi-objective optimization were got are mainly focused.The study contents:1.Development of the study in brittle block material mechanics,and the significances of research and analysis in using discrete element method have been described.In addition,the theory of discrete element BPM model,the research status domestic and abroad have been reviewed,the necessity of checking the mesoscopic parameters of discrete element BPM model has been pointed out.2.The basic principle of discrete element software PFC and its numerical simulation application have been briefly introduced,especially in brittle material.The two different forms of binding constraint key and their parameters that need to be verified in Fishtank module have been described.The mechanics model building process of brittleness of bulk materials has been introduced.Using the uniaxial compression numerical test,the macro mechanical parameters’ simulation and calculation process of the BPM have been conducted.3.Using orthogonal test method,the impact on the three macro mechanical parameters,elastic modulus,Poisson’s ratio and uniaxial compressive strength,caused by the parameters of BPM model constructed with parallel constraint key has been analyzed,which is high coordinated with formal verification experiences and provides scientific theoretical basis for the verification of discrete element BPM model of the brittle material.4.The uniform Latin square experiment design method was selected to sample microscopic parameters of discrete element BPM model.Using the response surface agent model method,the relation between parameters of discrete element BPM model and target values of simulation has been matched.Three response surface agent models for the elastic modulus,Poisson’s ratio and uniaxial compressive strength of C20 concrete have been built.The precision measurement for the determination coefficient and adjustive determination coefficient of approximate response surface agent model has been proceeded.The results indicate that the precision of three response surface agent models we built is high,which makes it able to replace the BPM model of PFC to proceed complex simulation process,and establish the foundation of multi-objective optimization for discrete element BPM model simulation in the next step.5.The optimum design method and multi-objective optimization design method have been summarized in this paper,and the genetic algorithm-combinative method of smart computing technology and multi-objective optimization design method have been introduced.The elitism selection of non-dominated sorting genetic algorithm has been used to solve multi-objective optimization for three response surface agent models of C20 concrete.It obtains a series of Pareto front solutions.The satisfactory solutions need to be selected and to be used in discrete element BPM model for operation simulation.Error analysis between the target values we got and the macro performance index of real material need to be proceeded.The results indicates that: using this multi-objective optimization checking method,not only the optimization results we got meet the expected optimization design requirements,but also the checking has the quite great commonality and substitutability when verify parameters with different brittleness of bulk materials of discrete element BPM models.