Determination of earth stresses using inverse analysis based on coupled numerical modelling and soft computing

摘要

An inverse analysis method based on coupled numerical modelling and soft computing has been investigated for estimating the magnitude of horizontal earth stresses based on borehole displacements. The method integrates genetic algorithm (GA), artificial neural network (ANN) and numerical computation. In the displacement-based inverse analysis method, a hybrid ANN-GA model is used to identify horizontal earth stresses. An ANN is used to represent the non-linear relationship between the maximum and minimum horizontal earth stresses (σ_H, σ_h) and borehole displacements. GA is used to search the set of unknown horizontal earth stresses based on the objective function. Numerical computation is used to create the necessary training and testing samples for the hybrid ANN-GA model. Results of numerical experiments demonstrate that a hybrid ANN-GA model for displacement-based inverse analysis can effectively identify horizontal earth stresses based on wellbore movements during drilling.