The folding defect is one of the most common flow-induced defects in the die forging, and the prediction of folding defects is of great significance in product and processing design. To this end, a versatile approach for the prediction of three typical types of folding defects was developed in this work. Firstly, the folding defects in die forging are classified into three types: confluence-type, bending-type, and local-loading-type, according to the material flow characteristics during their development. Then, three simple Eigen experiments and the corresponding FE models are developed to capture the formation processes of these three typical of folding types. On these bases, a versatile prediction approach is established and verified for all three typical folding defect types. This approach is realized by combining a folding index, judging criteria, and a mathematical model relating the folding index and formation parameters. A folding index based on the integration of the strain rate over the free surface of workpiece is proposed, which can evaluate the risk of all three typical types of folding. The corresponding judging criteria for folding defects based on folding index are determined by multiple uniform experiments for each type of folding defects. The mathematical model relating the folding index and forging parameters can be developed by common metamodeling techniques based on FE simulation results; the response surface methodology is employed as an example in this work. Compared to informed folding prediction methods, the present approach presents the following main advantages: (1) wider application scope to all typical types of folding in die forging, (2) no geometrical parameter dependence, (3) the ability to apply the folding index not only to predict folding defects, but also to serve as an objective function to optimize forging parameters to minimize the risk of folding. The results will provide an important tool for the workpiece and processing design to improve the forming quality during die forging.
Read full abstract