Pattern recognition-based strategy to evaluate the stress field from dynamic photoelasticity experiments

摘要

For avoiding fails in loaded structures, adjust their geometry, removing material, or quantify residual stresses, photoelasticity studies often is limited by complex experiments, excessive computational procedures, expert supervision, narrow applications, and static focus. This paper proposes a pattern recognition-based strategy for evaluating the stress field from simplex dynamic experiments. Here, temporal color variations are processed to extract, select and classify stress magnitudes, isotropic points, and inconsistent information. This approach used synthetic photoelasticity videos from analytical stress models about disk and ring under diametric compression. Additional to improve limitations in conventional photoelasticity approaches, this strategy identifies isotropic and inconsistent points.