Preform cross-section shape optimization design for tube hydroforming with low pressure

Abstract

Preform design refers to the intermediate geometric design between the initial tube blank and the desired product. A proper preform design can improve the formability of the tubular structure with small fillets. It is difficult to obtain a reasonable preform design based on engineering experience or trial-and-error methods. In this work, a Non-Uniform Rational B-Splines (NURBS) curve control point inversion algorithm-based approach is introduced to define and parameterize the preform cross-section profiles. The corresponding mathematical optimization model of preform cross-section shape design for tube hydroforming with low pressure is established, and an optimization procedure was accordingly proposed. In this procedure, the relationships between the design variables and responses are formulated by combining the DoE techniques with multiple surrogate models. The k-fold cross-validation strategy is adopted to quantify the performance of the candidate surrogate models on fresh datasets. The selected and verified surrogate models are integrated with the heuristic algorithm to determine the optimal preform shape. Compared to the direct hydroforming process, the optimized preform designs decreased the forming pressure of square, trapezoidal, and irregular polygonal tubes by 89.29%, 70.29%, and 32.67%, respectively. The validity of the proposed method is thus verified.