Using Neural Network to Determine Rockfill Parameters of Duncan E-B Model

Duncan E-B model has been widely used in engineering practice. And how to determine the parameters of the model accurately and easily is the main step to ensure the reliability of the results. Based on BP neural network in MATLAB toolbox, this paper aim to establish the non-linear relation between the density γ, porosity e and the constitutive model parameter K (K_b, R_f, Q₀) which is difficult to be determined, and through the model above, forecast the parameters of E-B model for one certain project. And the results meet the project accuracy requirements, Which provides a new method to determine the parameters of constitutive model.