摘要
随着现代技术的发展,在汽车和航空航天等领域都在追求材料的轻量化,而材料的高强度高韧性是轻量化的基础。7000系铝合金(Al-Zn-Mg-Cu系铝合金)具有高强度、高硬度、良好的耐腐蚀性能等优点,其中7055铝合金是所有铝合金中强度最高的。喷射成形工艺是7055铝合金常用的制备方法。铝合金锭坯在沉积过程中的稳定生长是喷射成形工艺制备形貌一致、沉积质量均匀的大规格锭坯的基础。受喷射成形过程中众多工艺参数变化的影响,现有的理论模型难以满足实际生产过程中质量控制要求。文中通过所采集的喷射成形过程工艺历史数据与锭坯沉积表面直径的相关性分析,结合BP神经网络与遗传算法,构建锭坯直径的GA-BP神经网络预测模型及生长速度调控模型;根据工艺参数的实时波动,计算直径变化量,作为输入层导入经过训练的速度调节神经网络模型,对沉积基底的升降速度进行优化调控,使得锭坯沉积生长形貌均匀稳定。最后采用该方法对锭坯沉积生长速度调控工艺进行实验分析。结果表明,通过该工艺参数优化方法制备的大规格锭坯直径尺寸偏差在5%以内,验证了生长速度调节的可行性。
With the development of modern technology, the automotive and aerospace fields are pursuing the lightweight of materials, and the high strength and high toughness of materials is the basis of lightweight. 7000 series aluminum alloys(Al-Zn-Mg-Cu series aluminum alloys) have the advantages of high strength, high hardness, good corrosion resistance, et al. Among all aluminum alloys, 7055 aluminum alloy has the highest strength. The common preparation method of 7055 aluminum alloy is spray forming process. Stable growth of the aluminum ingot during deposition is the basis for the preparation of large-size ingots with uniform deposition quality by the spray forming process. Due to the variation of numerous process parameters during the jet forming process, the existing theoretical model is difficult to meet the requirements of quality control in the actual production process. This paper built a GA-BP neural network prediction model for the diameter and a model for regulating the growth rate of the ingot billet based on the correlation analysis between the historical data of the injection molding process and the diameter of the deposited surface of the ingot billet, by combining BP neural network and genetic algorithm. Based on the realtime fluctuation of process parameters, the diameter variation was calculated and used as an input layer into a trained velocity regulation neural network model to optimally regulate the lifting speed of the deposition substrate,resulting in a uniform and stable deposition growth profile of the ingotst. Finally, this method was used to regulate the growth rate of ingots. The results show that the deviation of large-size ingot diameter is within 5%, which verifies the feasibility of growth rate regulation.
作者
冷晟
付有为
马万太
钱浩
虞钧鹏
蒋云泽
吴尚霖
LENG Sheng;FU Youwei;MA Wantai;QIAN Hao;YU Junpeng;JIANG Yunze;WU Shanglin(College of Mechanical&Electrical Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,Jiangsu,China;Jiangsu HaoRan Spray Forming Alloy Co.Ltd.,Zhenjiang 212200,Jiangsu,China;China Aerospace Science and Industry Nanjing Chenguang Group Co.,Ltd.,Nanjing 210006,Jiangsu,China)
出处
《华南理工大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2023年第2期27-34,共8页
Journal of South China University of Technology(Natural Science Edition)
基金
国家重点研发计划项目(2021YFB3700903)
江苏省重点研发计划项目(BE2019726)
直升机传动技术国家重点实验室基金资助项目(HTL-0-21G13)。
关键词
喷射成形
大规格锭坯
工艺参数
神经网络
速度调控
spray forming
large-size ingots
process parameter
neural network
rate regulation
