Data Driven Optimization Of Process Parameters For Extrusion Dies

Hot extrusion die is a key part of the production process of aluminum profiles.It has a great influence on product quality,production efficiency and cost.According to statistical data,molds account for a large percentage of failures due to wear.The traditional process parameters design and formulation are mainly based on artificial experience.The arbitrariness and uncertainty are relatively large.This can lead to sometimes unreasonable process parameters,serious wear in production,and frequent mold repairs and mold changes.Big waste is not conducive to reducing production costs.In this paper,in order to solve these problems,hot extrusion dies are used as the research object,and the product is a round bar material.In order to reduce the die wear value as a concern,DEFORM-3D finite element software simulation is used to obtain the influencing factors and wear data that affect die wear.Based on these data,the least squares support vector machine was used to train,and the wear prediction model was obtained.The wear analysis and prediction of the mold were realized.Finally,through the genetic algorithm,according to the obtained model,with the minimum wear as the optimization goal to find the optimal process parameters.The specific research work in this paper is as follows:(1)According to the aluminum extrusion production process,the effects of mold temperature,extrusion speed,and mold hardness on wear are systematically analyzed,and the wear characteristics and wear influencing factors of the extrusion mold are revealed.(2)In view of the fact that the mold environment and parameters of the actual production of the company are varied,seven sets of experimental programs are designed.The simulation is performed using DEFORM-3D finite element software.The wear information is acquired by setting the collection points of the mold to obtain a set of wear information.Wear data,and for the simulation results,studied the impact of extrusion temperature,speed,mold hardness on mold wear and wear changes in the parameters change.(3)Use the finite element analysis software DEFORM-3D simulation analysis data as a learning sample,use the least squares support vector machine(LS-SVM)to fit and contrast the sample data to establish a hot extrusion wear model,and use This model predicts mold wear.The results show that the model can effectively learn the complex nonlinear relationship in die wear and the prediction accuracy is relatively high.Therefore,the model can be used as an effective predictor of mold wear in actual machining and provide basis for parameter selection in extrusion process.(4)Compile the genetic algorithm program,use the prediction model established in Chapter 4,optimize with the minimum value of wear,and find the optimal extrusion process parameters.Under the designed experimental conditions,it was found through calculation results that under the conditions of extrusion speed V=10 mm/s,mold temperature T=430° C.,and mold hardness H=45HRC,the value of mold wear is the smallest and it is an optimized process.parameter.The above research,provides a method for optimizing the extrusion process parameters for the actual production of the company.It provides ideas for the health management of the extrusion die,which is conducive to the safety of the enterprise and continuous and trouble-free production.Extend the service life of the mold to improve the market competitiveness of the company.