| Manufacturing industry is the basic guarantee of a country’s economic and social activities,and intelligent manufacturing is the key to realize the rapid development of manufacturing industry.As a national strategic industry,aviation manufacturing industry needs to combine modern information technology to enhance the intelligence and advanced nature of manufacturing process.In the aviation manufacturing workshop,the production control system and its core scheduling module play the role of "commander".The perception and calculation of key parameters play a very key role in the production control and scheduling process.However,with the development of information technology and the rise of artificial intelligence methods,some traditional methods have some problems,such as low efficiency and insufficient accuracy.Therefore,this paper takes the aviation manufacturing process as the background,faces the production scheduling process,and studies the intelligent perception and calculation of key parameters.The main research of this paper is as follows:1.There are some problems in aviation manufacturing workshop,such as low efficiency of production schedule acquisition and lack of perceptual acquisition ability.Therefore,this paper proposes a production schedule acquisition method based on image data processing.Firstly,the video data in the aviation manufacturing process is preprocessed to obtain the image data,and the image data are classified and marked in combination with the process flow and site conditions in the manufacturing process.Each category represents the key progress nodes in the manufacturing process.Then the labeled data is used to train the convolutional neural networks(CNN)model,and the actual production data is used to verify the accuracy of the model.The results show that the method has higher accuracy than other traditional methods.The trained CNN model can perceive the production status in real time and collect the production progress information.2.At present,the production scheduling in the aviation manufacturing process needs more accurate processing time as the basis for production scheduling decision,and the production process is complex and affected by many related factors.The prediction of it needs to fully consider all kinds of influencing factors on the premise of ensuring the accuracy.Therefore,this paper proposes a processing time prediction method based on graph neural network.According to the actual situation and characteristics of aviation manufacturing process,the graph data model is constructed,and then the processing time prediction model is established through graph neural network model and aviation manufacturing process data.Finally,by comparing the error of the predicted value between this method and other methods,the effectiveness of the proposed prediction method is proved.The model can predict the processing time in advance in the aviation manufacturing process,provide more scientific decision-making basis for production scheduling process,and improve the rationality of production control.3.Based on the studied production schedule acquisition method and processing time prediction method,a prototype system for perception and calculation of key parameters in aviation manufacturing process is designed and developed.The architecture and function of the system are described in detail,and the application of the proposed research results in the actual manufacturing process is visually displayed,which makes the perception and calculation of key parameters more convenient and efficient. |
