Numerical Solutions Of Inverse Couple Stress Problems

This paper presents a finite element model facilitating to the sensitivity analysis for solving the direct/inverse couple stress problem, and direct and adjoint methods are used to conduct on the sensitivity analysis. Gauss-Newton technique, homotopy method and conjugate gradient method are employed to identify unknown constitutive parameters for the inverse problem respectively. The effects of initial guesses, number of sample points and data noise on the solutions are investigated, and satisfactory results are exhibited in the numerical verification.Via FEM (finite element method), this paper develops a direct numerical model problem for the layered rock mass formulated by Cosserat theory, deduces the formulae of the singled/combined identification for in-situ stress and Young's modulus. Ant colony algorithm based Gridding Partition and Gauss-Newton technique are employed to solve above inverse problem respectively. The effects of data noise and value ranges on results are investigated, numerical validation exhibits good results.In addition, this paper uses the Ant colony algorithm in conjunction with the inverse transient method to detect leakage locations and lumped leak coefficients for water supply system. The effects of data noise and value ranges on results are investigated, numerical validation exhibits good results.This paper provides a necessary reference for researching on couple stress theory and leak detection of water supply system, its application in practice is looked forward to.