Inelastic deformation processes with random parameters––methods of analysis and design Original Research Article

Methods of analysis and design for inelastic deformation processes with random input parameters are exposed. Stochastic analysis makes use either of function evaluations for a synthetic statistical input sample (Monte Carlo technique) or of a Taylor-series expansion of the response around the mean input. The latter relates to the perturbation method, extended over the duration of the deformation process following temporal integration. Tasks concerning process design are defined and the employment of the stochastic analysis techniques is discussed for exploring the parameter space as well as for design improvement and optimization. The suitability of either analysis technique for performing the particular tasks is pointed out. The design problem raises the issue of robustness to input scatter. Robust optimization respects both the mean and the variance of the design objective via the desirability function, a weighted combination of the two quantities. The exploration potential of synthetic sampling is demonstrated by an application where sensitivity-based techniques prove inadequate because of the local nature of the approximation.