پيش بيني حد شكل دهي در فرايند كشش عميق هيدرومكانيكي ورق فلزي با استفاده از شبكه عصبي مصنوعي و تحليل اجزاي محدود

چكيده انگليسي :

PREDICTION OF FORMING LIMIT IN SHEET METAL HYDRO MECHANICAL DEEP DRAWING PROCESS USING ARTIFICIAL NEURAL NETWORK AND FINITE ELEMENT ANALYSIS Sobhan Jamshidifard s jamshidifard@me iut ac ir Date of Submission 2011 2 28 Department of Mechanical Engineering Isfahan University of Technology Isfahan 84156 83111 Iran Degree M Sc Language Farsi Supervisors Hasan Khademi zadeh and Hasan Moosavi khademyza@cc iut ac ir moosavi@cc iut ac ir Abstract Hydromechanical deep drawing process is one of the forming processes which is devised for solving the existing problems in conventional processes such as traditional deep drawing process Hydromechanical deep drawing process has many applications in different industries such as aerospace military and automobile industries There is a good control in the wall of the cup due to applied hydraulic counter pressure in hydromechanical deep darwing process for this reason the wrinkling phenomenon has been controlled Higher economic efficiency and flexibility can be achieved by using hydromechanical deep drawing process In comparasion with covnventional deep drawing process the part quality and thickness distribution are better in hydromechanical deep drawing process Some advantages of hydromechanical deep drawing are improving the material formability reduction of friction force the accuracy of forming part and the reduction of forming stages due to improvement of limiting drawing ratio Hydromechanical deep drawing process is an efficient method for producing of complicated parts Conventional deep drawing process is applied for mass production parts but hydromechanical deep drawing process more is used for single and batch production type Though hydromechanical deep drawing process has some advantages but this method is faced with some difficulties The finding of appropriate counter pressure and drawing ratio for different materials and process parameters is one of the problems which this process confronts with them The try and error method can be used for solving these problems but this method is very expensive and requires too large time Also there is some difficulties in hydromechanical deep drawing process analysis such as variable pressure area in punch stroke and geometric complication of part shape Considering these difficulties finding an analytical solution for hydromechanical deep drawing process is a complicated task Therefore numerical method and simulation can be suitable way for solving this problem but this method is so expensive too and a separate solving must be executed for each study case A simple and quick method for solving this problem can be obtained by using an artificial neural network and implementation of a database in neural network In this thesis simple and fast solving way for hydromechanical deep drawing process has been proposed by implementation of finite element simulation results in an artificial neural network Process simulation has been done by using ABAQUS software An innovative method has been used for acting the liquid pressure load proportion to punch stroke Forming limit diagram damage criteria has been used for prediction of failure in blank In this thesis low carbon steel and aluminium materials have been investigated Overally the finite element method is a good way for simulation of hydromechanical deep drawing process also the artificial neural networks is a fast and simple way for predicting of forming limit in hydromechanical deep drawing process A good agreement between finite element analysis and artificial neural network results were found By using hydromechanical deep drawing process limiting drawing ratio 2 5 and 3 1 can be achieved for aluminium and low carbon steel respectively Keywords Hydromechanical deep drawing Drawing ratio Artificial neural network Finite element method

جمشيدي فر، سبحان

پيش بيني حد شكل دهي در فرايند كشش عميق هيدرومكانيكي ورق فلزي با استفاده از شبكه عصبي مصنوعي و تحليل اجزاي محدود

حد كششي , روش اجزاء محدود