Tube hydroforming process in order to manufacture complex partsperfectly, for example y-shapes;
demands precise loading paths, Otherwise because of complexity of these parts, some deficiencies like
wrinkling, bursting or imperfect forming may occur. Finding an effective combination of applied forces
depends on repeating experimental experience numerously which in turn needs a lot of time and cost.So
using numerical simulations, which are remarkably faster and cheaper than practical experiments, can
resolve this problem influentially. In this study firstly a y-shape was modelled using FEA software and then
obtained results were compared with experimental ones to verify and confirm the simulation precision. In
the next step objective and constraints functions were defined using Neural Network. To do this, simulation
was repeated in certain numbers and needed Neural Networks were trained by a statistical model produced
using that results.Finally in order to optimize the objective