Development of Design Processes for Multi-Spindle Drilling using the Neural Network and Expert System

This work presents the integration of multilayer neural network with an expert system for the automatic choice of the design process of multi-spindle drilling housing. An intelligent design system approach is developed to integrate various phases of the mechanical process including neural network subclasses and MLANN. This solution reduces the time during the design preparation process and to improve production. The automatic choice is carried out in three steps: Firstly, we started with the formulation of training base experiments of the experts of the field as well as the necessary knowledge of expertise, and which is among the general criteria for the choice of the design process. The second step was devoted to the creation of many multi layers NN1…NNm for the choice of the design mode. The final step is related to the application of the outputs as results and an input for chaining by the expert system. This chaining is based on several models based chaining (input data collection) from the neural network results and processing (output results). The results are the kinematics schema of the multi spindle drilling housing.

Keywords: Computer Aided Design, Design Methodology, Expert System, Mechanical Design, Neural Networks, Object-Oriented Language.