Robust topology optimization based on finite strain considering uncertain loading conditions

The present study proposes a method of robust topology optimization assuming uncertainties in the magnitude and direction of loads applied to a geometrically nonlinear structure. The objective function is the sum of the expected value and the standard deviation of end-compliance of a structure with a compressible Neo-Hookean hyperelasticity. In this study, quadratic approximation of the end-compliance with respect to random variables is employed to reduce computational cost. To ensure its accuracy, a complete analytical formulation is derived, and the performance and limitation of the proposed method are deeply discussed. The performance of the proposed method is verified and its numerical examples emphasize the importance of considering geometrical nonlinearity to obtain robust structures with respect to uncertainties of loading conditions. Finally, we have obtained a finding that the optimized network system consisting of thin members plays a significant role in the improvement of robustness of structures.