Flexible roll forming is a new roll forming process that produces parts with variable cross sections. This forming process is proposed to meet the demand of weight reduction of automobile industry. In order to study t...Flexible roll forming is a new roll forming process that produces parts with variable cross sections. This forming process is proposed to meet the demand of weight reduction of automobile industry. In order to study the mechanisms and material flow rules in this new forming process,the finite element mothod( FEM) model of a nine-step flexible roll forming of an ultra-high-strength steel bumper is established based on deep understanding and reasonable simplification of the process.Given that the material model is an important factor that influences the simulation accuracy,three material models which consist of different yield criteria and hardening models are adopted in the FEM models. Sheet thickness and springback amount calculated with three material models are studied comparatively. According to sheet thickness reduction and springback amounts,it is found that the MKi( Mises yield criterion and kinematic hardening law) model's result is larger than MI( Mises yield criterion and isotropic hardening law) model and HI( Hill's yield criterion and isotropic hardening law) model. Therefore,it is concluded that material models do have influences on the flexible roll forming simulation and need to be determined carefully.展开更多
The rapid development of robotics technology has made people’s lives and work more convenient and efficient.Theresearch and simulation of robots combined with reinforcement learning intelligent algorithms have become...The rapid development of robotics technology has made people’s lives and work more convenient and efficient.Theresearch and simulation of robots combined with reinforcement learning intelligent algorithms have become a hotspot in variousfields of robot applications.In view of this,this study is based on deep reinforcement learning convolutional neural networks,combined with point cloud models,proximal strategy optimization algorithms,and flexible action evaluation algorithms.A sealcutting robot based on deep reinforcement learning has been proposed.The final results show that the descent speed of the sealcutting robot with the root mean square difference as the performance standard is about 1%faster than the flexible actionevaluation algorithm.About 2%is faster than the proximal strategy optimization algorithm.It is about 4%faster than the deepdeterministic strategy gradient algorithm.This indicates that the research model has certain advantages in terms of actualaccuracy after cutting.The fluctuation of this model is about 10%smaller than the evaluation of flexible actions and about 60%smaller than the gradient of deep deterministic strategies.Therefore,the research model has the highest overall stability withoutfalling into local optima.In addition,compared to the near-end strategy optimization algorithm,it falls into local optima,resultingin a low coincidence degree of about 17%.The deep deterministic strategy gradient algorithm has a large fluctuation amplitudeduring the seal cutting process,and the overall curve is relatively slow,with a final overlap of about 70%.The overlap degree offlexible action evaluation is slightly higher by about 83%.The maximum stability of the model’s overlap is best around 90%.Through experiments,it can be found that the seal cutting robot proposed in the study based on deep reinforcement learningmaintains certain advantages in performance indicators in various types of tests.展开更多
基金Supported by the National Natural Science Foundation of China(No.51205004)Beijing Natural Science Foundation(No.3164041)the National Key Technology R&D Program(No.2011BAG03B03)
文摘Flexible roll forming is a new roll forming process that produces parts with variable cross sections. This forming process is proposed to meet the demand of weight reduction of automobile industry. In order to study the mechanisms and material flow rules in this new forming process,the finite element mothod( FEM) model of a nine-step flexible roll forming of an ultra-high-strength steel bumper is established based on deep understanding and reasonable simplification of the process.Given that the material model is an important factor that influences the simulation accuracy,three material models which consist of different yield criteria and hardening models are adopted in the FEM models. Sheet thickness and springback amount calculated with three material models are studied comparatively. According to sheet thickness reduction and springback amounts,it is found that the MKi( Mises yield criterion and kinematic hardening law) model's result is larger than MI( Mises yield criterion and isotropic hardening law) model and HI( Hill's yield criterion and isotropic hardening law) model. Therefore,it is concluded that material models do have influences on the flexible roll forming simulation and need to be determined carefully.
文摘The rapid development of robotics technology has made people’s lives and work more convenient and efficient.Theresearch and simulation of robots combined with reinforcement learning intelligent algorithms have become a hotspot in variousfields of robot applications.In view of this,this study is based on deep reinforcement learning convolutional neural networks,combined with point cloud models,proximal strategy optimization algorithms,and flexible action evaluation algorithms.A sealcutting robot based on deep reinforcement learning has been proposed.The final results show that the descent speed of the sealcutting robot with the root mean square difference as the performance standard is about 1%faster than the flexible actionevaluation algorithm.About 2%is faster than the proximal strategy optimization algorithm.It is about 4%faster than the deepdeterministic strategy gradient algorithm.This indicates that the research model has certain advantages in terms of actualaccuracy after cutting.The fluctuation of this model is about 10%smaller than the evaluation of flexible actions and about 60%smaller than the gradient of deep deterministic strategies.Therefore,the research model has the highest overall stability withoutfalling into local optima.In addition,compared to the near-end strategy optimization algorithm,it falls into local optima,resultingin a low coincidence degree of about 17%.The deep deterministic strategy gradient algorithm has a large fluctuation amplitudeduring the seal cutting process,and the overall curve is relatively slow,with a final overlap of about 70%.The overlap degree offlexible action evaluation is slightly higher by about 83%.The maximum stability of the model’s overlap is best around 90%.Through experiments,it can be found that the seal cutting robot proposed in the study based on deep reinforcement learningmaintains certain advantages in performance indicators in various types of tests.
