Die-Life Analyzer Employing Independent Instance Databases of Cold Forging Die Life on the Internet

Abstract

The author developed a die-life analyzer to be used in cooperation with independent databases of cold forging die life on the Internet. These Internet instance databases (written in HTML and XML and accessed using a “www” browser) are used by forging engineers as references in their forging process design. The die-life analyzer uses these databases as a knowledge base to predict and analyze forging die life, as well as the causes of short die life. The author considers die-life analysis as risk analysis of whether the life of a target tool will be shorter than the life of instances and developed a die-life analyzer using risk-tree networks and fuzzy linguistic risk calculation. The author subsequently investigated the causes of short life in order to create a risk analysis tree network. A certain cause, or a “parent node” in other word, is divided into more detailed causes called “daughter nodes.” The daughter nodes are connected to their parent nodes with weights, and the risk of the parent node is calculated as the average of daughter nodes' risks using the extension principle. Through this process, the risks, which cause die life to be shorter than for instances (that is, the risk rate of the top of all nodes), are calculated. Finally, the system predicts die life with interpolation using die life instances and the corresponding risks that shorten die life.