Development of an Intelligent Knowledge Based System (IKBS) for forging die design

• Main Author: Bakhshi-Jooybari, Mohammad
• Published: University of Birmingham 1995
• Subjects: 671.3; TJ Mechanical engineering and machinery
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.693070

The work in this thesis is concerned with further development of an Intelligent Knowledge-Based System (IKBS) for forging die design. It follows on from initial work carried out at the School of Manufacturing and Mechanical Engineering. The main parts of the original design for the system are a sequence design program (SDP) for two and three dimensional parts, an interface program which can be connected to a finite-element program for metal forming simulation and a Control Module which supervises these two parts and co-ordinates their activities. Of these three modules, only the SDP and the Control Module existed when the current work was started. The purpose of the work reported here is to develop, improve and validate the original system. Among the five different families of components within the original IKBS, Stub Axles have been selected for the current research work. An interface program has been written which can generate a datafile for the available finite-element program (EPFEP3). This interface program inputs one preform stage as the geometry for mesh generation and the corresponding product stage in order to determine the boundary conditions. It also inputs the data within the SDP database for completing the other parts of the datafile. This program is efficient, rapid and user friendly and can easily be extended for the other families of components in the SDP. In the IKBS, when a new component is input to the system, each forming stage of the component should be compared with the same stage of the same family of all the components stored in the database. To do so, the significant processing and geometrical parameters and also their weighting effects should be input to the system. A new experimentally-based approach has been developed to obtain the weighting effects of the significant parameters. The weighting factors obtained are saved in the knowledge-base and have been shown to lead to the correct predictions when data for real forgings was used. The method for obtaining the weighting effects of the significant parameters can be extended to the other families of components within the IKBS. Programs have been written to perform computer-aided reasoning in the IKBS. In particular, recognising and extracting the values of the significant parameters of the operational sequence of a component, creating the IKBS database based on real data and performing the comparison procedure for a new component stage with those stored in the IKBS database.