Adaptivity in isogeometric analysis of structures using error estimation methods based on stress recovery

In this research, for the first time, the net of control points in the isogeometric analysis has been improved by employing an error estimator based on a stress recovery method. First, an error estimation algorithm based on stress recovery is used to obtain the energy norm for each element. Then, artificial rods are defined between control points and the estimated values of errors in the control points located at the vicinity of a typical point is assigned to each rod as a thermal gradient. Now, by analyzing this hypothetical truss problem under temperature changes a new arrangement of control points and consequently the knot vectors can be obtained. Repeating this process in isogeometric analysis will lead to a better distribution of errors in the domain of the problem and results in an optimal net of control points to calculate the integrals. To evaluate the efficiency of this method, the results of modeling and analysis of two elasticity problem is presented. The obtained results show that this innovative approach has a good performance and can be employed for reducing the analysis errors.