Optimization Of Technologic Parameters Of Mechanical Expanding Based On Neuro Network And Robust Design

Mechanical expanding is an important plastically forming process which is used to manufacture large welding pipe. This technological method is a popular final formation of UOE, JCOE, thick-wall cylinder part, thin-wall spinning cylinder part, Welding pipe with large diameter, spiral welding pipe with large diameter and other large and medium cylinder parts, because it improve precision of product, enhance yield strength and control residual stress. At the present time, many forming tests have to be done to look after a group of optimization parameters to meet the request of user under the condition of ether change of material, dimension and shape of stock pipe or change of demand of forming precision of product pipeAn Optimization method for technological parameters of mechanical expanding was gained based on genetic arithmetic and BP net with robust analysis for Pareto results through study on the optimization problem considering efficiency, precision and robust. In the mean time, dimension and shape error were objective functions, while design variables were deformation parameter and die parameters.Affection weight of parameters on dimension error and shape error of pipe were analyzed quantitatively by orthogonal design and numerical simulation. The main parameters effected on mechanical expanding and quality parameters were obtained by using numerical simulation of mechanical expanding which confirmed the design variables.Optimization of mechanical expanding by genetic algorithm based on BP was conducted. Predictive model of mechanical expanding was got through the function of BP neuro network. Sufficiency function was achieved by this model. Comparing this way with above method efficiency of optimization was improved.To make sure that optimization result is robust, the best robust optimum parameters were chosen by comparing robustness of those Pareto results acquired by genetic algorithm.Man-machine interfaces of prediction system and optimization system of mechanical expanding were got by GUI development toolSeamless pipes with specification of 109.2×5×120 mm ,which are 20 steel, were used in experiment. The optimization method for mechanical expanding is practical and effective through comparing the optimization results with experiment results.