Research On Aid Design Of Tube Spinning Process Parameters Based On Knowledge And Neural Network

Power spinning is a metal forming process, which combine with the characteristic of forging, extrusion, bending, deep drawing, ring roller, rolling and is frequently used for manufacturing hollow rotational symmetric shapes. Power spinning is widely used in the fields of aeronautics, astronautics, manufacture of weapons, etc.It is a complicated plastic forming process, because of the complication of its deformation mechanism and the multiple forming factors, the design techniques of its process parameters is immature. As a result, engineers have to establish process parameters by some experience formulae, trial-and-error approach. This means is usually of great subjectivity and uncertainty. Highly qualified operators are needed towork out corresponding parameter, which leads to high engineering costs and low efficiency.Knowledge-based Engineering is introduced into tube spinning process planning.The design experiences and knowledge of tube spinning process were deal with systematically. Knowledgebase of products and process p-lanning of tube spinning was created, including knowledgebase of mat-erial, knowledgebase of case, knowledgebase of rule. Methods for case representation and case retrieval applicable to tube spinning parts were designed.A mod-el to calculate the similarity of two cases are constructed.A neural network based response surface method was proposed in order to overcome the limitation of quadratic response surface in solving non-line problems. Relationship between maximum of spinning force variety and material parameters, tube spinning process parameters was established and optimum was achieved by using Particle Swarm Optimization algorithm, then optimization of tube power spinning process parameters was realized.Based on the analysis of finite element method and process characteristics of tube spinning, the three dimension plastic-elastic finite element model is built through FEM software-MSC.MARC. The feasibility of a sample was verified through finite element simulation.