Development Of Adaptive Optimization Platform For Improving Forming Uniformity Of Rotary Forgings

Rotating(or axisymmetric)structural parts are widely used in aerospace,gas turbine,hydropower and other large equipment.For example,as the heart of aeroengine,there are many rotating components such as an organic case,a turbine disk,a fan disk,a pneumatic disk,etc.This kind of component has been in the complex working condition of high temperature,high pressure and alternating load for a long time,so it has a high demand on the homogeneity of structure and mechanical properties.Integral die forging is the main forming method of this kind of component at present,and the reasonable design of the shape and size of blank and preformed blank is an important guarantee to obtain high-quality forging.However,the "trial and error" method adopted in the current production practice is difficult to obtain the optimized forming size parameters efficiently and reliably.Therefore,it is of great theoretical and engineering value to establish a fast optimization method for the geometric size of the billet in view of the deformation uniformity of the rotary forging.This paper first summarizes the advantages and disadvantages of the existing die forging process optimization methods.According to the requirements of the connection between the mechanical property test of materials and the application of finite element,the development of the generator of material card is studied.Aiming at the problem that the current commercial CAD/CAE software can not automatically realize the joint feedback optimization,resulting in repeated and inefficient trial and error operation in the process design stage,a black box optimization model with genetic algorithm as the core is constructed based on deform and UG background mode and python language A self-adaptive joint optimization platform is developed for the deformation uniformity of rotary forging.The main work and achievements include:(1)In view of the problem that most of the testing data of mechanical properties of materials to the application of finite element analysis rely on manual operation at present,a material card generator is developed based on Numpy/Pandas,which is directly connected with the application of finite element,which improves the efficiency and accuracy of numerical simulation.(2)Taking genetic algorithm as the kernel control and the proportion of the expected strain range units to the total units as the fitness function,the implicit model relationship between the deformation amount and the parameters to be optimized is constructed,and the optimal parameters combination of blank is obtained by self-adaptive iteration of selection,crossover,variation and other mechanisms;the deformation is developed by using Python language,UG background mode and deform text mode The software platform of adaptive optimization of die forging blank size for uniform problem.To some extent,the platform can solve the problems of many forming parameters,large change space,high calculation cost of exhaustive trial and error method and random optimization direction,while the approximate models such as orthogonal experiment,response surface and artificial neural network need a large number of samples,and the prediction results are easy to fall into local optimum and the approximate model has no universality.(3)Taking a TC6 low pressure plate forging as an example,the size parameters of blank and preform are optimized adaptively.The trial production results of the optimized scheme of the low pressure disk show that the mechanical properties,physical and chemical inspection and ultrasonic detection results of the parts meet the requirements of the standard,which proves that the optimization is effective.