Data-driven Material And Process Design Of Aluminum Alloy Cabin Structural Components

Large complex thin-walled cabin castings are usually used as components to provide structural support and load transfer in aerospace,rail transportation,weapons,and automotive industries.To meet the requirements of"long range（or deep dive）,high precision and ultra-light",aluminum alloy with high specific strength,high specific stiffness and good corrosion resistance has become the preferred material."Trial and error”has been the main designing method for the material and process of traditional aluminum alloy cabin parts which leads to high time cost and production cost of the whole R&D cycle,and is not conducive to the rapid development and application of aluminum alloy and its cabin structural parts.The computer numerical simulation technology can describe the casting process of flow field,temperature field,solidification field intuitively,so as to solve some hidden problems in casting,and ensure the quality of castings.But,this does not play much of a role in improving the efficiency of aluminum alloy cabin structural parts research and development,and the design of casting systems and casting parameters still requires the intervention of"manual experience".In summary,the use of new,efficient R&D mode has become an urgent need to promote the upgrading of China’s foundry industry.With the dramatic increase in modern computer computing power,the booming development of artificial intelligence technology and the proposed materials genome project to inject vitality into the field of casting-related research,the traditional casting process research and development began to experience trial and error mode to knowledge-driven,data-driven big data analysis-design-prediction model innovation.Thus,this paper takes large complex Al-Si-Mg alloy cabin parts as the research object,analyzes the cabin casting structure and performance requirements and the applications of complex aluminum alloy cabin in domestic and foreign aerospace,rail transportation and automotive industries,etc.,uses image data（describing the internal and external structure of the casting and casting system）,casting process data,performance data,composition and process data of aluminum alloy as the machine The study was conducted on the intelligent design of the casting system,the intelligent design of low-pressure casting process parameters and the intelligent design of Al-Si-Mg alloy composition and process used in large complex aluminum alloy pods using machine learning algorithms such as convolutional neural network,random forest,linear regression and multilayer perceptron.Finally,the performance-oriented"composition-process-property"design strategy for Al-Si-Mg alloys proposed in this paper based on machine learning method has shown excellent results,and the errors between experimental and predicted values of tensile strength of alloys predicted or guided by this strategy are kept within±5%.The tensile strength of Al-6.8-0.6Mg-0.05Sr and 540 C×10h+170 C×10h are excellent,and the quality index Q_DJR reaches 517.3,which is higher than the level of similar alloys below 500 Q_DJR,and is enough to meet the requirements of the later casting into large cabin structural parts.The above results show that the data-driven machine learning model can help improve the long cycle time,high cost and low efficiency of the traditional method of aluminum alloy material development and its low-pressure casting process design for cabin structural parts,and also provide an important reference for other new material development and smart casting process design.