The optimal clearance design of micro - punching die

AbstrAct Purpose: The purpose of this research paper is focused on the optimal clearance design of micro-punching die by abductive network and SA method. Design/methodology/approach: The punching data (input) and wear size (output) were collected for a training database. In order to select proper clearance to evaluate the wear of die, the abductive network was used to establish an efficient relationship between input parameters and output result. This can help to predict wear size under any degree of clearance and hence to replace worn punches and dies at the right time. A simulated annealing (SA) optimization algorithm with a performance index is then applied to the neural network for searching the optimal clearance parameters, and obtains rather satisfactory result as compared with the corresponding experiment verification. Findings: This study aims to identify the relationship between clearance and service life of micro punches using the Neural Network, and to find relational data involving the service life of punches and punching parameters in non-metal blanking processes. The result can be used to estimate optimal clearance between punch and die for industrial applications. Research limitations/implications: In this study, the practical punching processes with different punching conditions were carried out for a set of training data. A trained model exhibited a relationship between service life and clearance of micro punch and die through an abductive network system. The predicted value of wear by abductive network is very close to the actual experimental value, with an error of less than 8%. This result satisfies the required standard for IC factory production. Originality/value: A good clearance design not only increases the quality of product manufactured, but also reduces product's burr. As a result, the wear of punches and dies can be greatly reduced and the life expectancy of punching dies increased.