Burr Prediction And Process Optimization Of Aeronautical Assembly Hole Drilling Based On Deep Neural Network

With the advancement and development of aerospace manufacturing technology,the structural performance requirements of a new generation of aircraft are gradually increasing.Aircraft assembly,as the main production link affecting aircraft structural performance and production cycle in the whole aircraft manufacturing process,is naturally the research focus of aviation manufacturing development.The quality of the connecting hole is the key technical index that affects the assembly quality of the aircraft components.However,the traditional hole making process lacks the targeted optimization of the quality performance index of the hole.Therefore,the process optimization method that can improve the quality performance index of the hole is guaranteed key technologies for aircraft assembly quality and fatigue life.In this paper,following the above research orientation,in order to improve the quality of the hole making,the research on the aeronautical assembly hole burr prediction and process optimization method is carried out.Firstly,based on the formation mechanism of hole drilling burr,the advantages and disadvantages of previous process optimization methods are analyzed.The research ideas and specific research schemes of aeronautical material hole making process optimization method based on deep neural network are designed.Then,by fully understanding the formation process of the burr and the experimental results,a prediction model of the burr formation based on the deep neural network is constructed.Then,the optimization and training of the prediction model are completed under the condition of sufficient data samples.Based on this,a hole making process optimization method based on the convolutional neural network glitch prediction model is proposed,and the performance optimization of the hole burr performance index is carried out effectively.Finally,based on the 125μm export burr height standard used by Boeing,the verification experiments of multiple sets of different burr height optimization targets are designed.The results prove the effectiveness of the proposed prediction model and process optimization method.