摘要
基于我国某铜厂的实际生产数据,提出了一种机器学习中XGBoost算法的微合金化铜杆性能预测方法,以实现对微合金化铜杆性能的优化设计,结果表明,XGBoost算法的预测精度较高,抗拉强度的最大误差为15 MPa,延伸率的最大误差为0.3%。通过对抗拉强度和延伸率模型的计算,在生产过程中,将稀土元素La含量控制在2×10^(-6)~3×10^(-6)范围内时,可将铜合金的抗拉强度和延伸率分别控制在395~405 MPa和3.85%~3.90%范围内,满足产品的性能和质量要求。经过力学性能预测并优化后的微合金化铜杆力学性能波动幅度减小,提高了铜杆产品性能的稳定性。
In order to realize the optimal design of the performance of microalloying copper rods,this paper proposes a method of predicting the performance of microalloying copper rods by XGBoost algorithm in machine learning based on the actual production data of a copper plant in China.The results show that the prediction accuracy of the XGBoost algorithm is high,and the maximum error of tensile strength is 15 MPa and the maximum error of elongation is 0.3%.Through the calculation of the tensile strength and elongation model,when the La content was controlled in the range of 2×10^(–6)–3×10^(–6) in the production process,the tensile strength and elongation of copper alloy can be controlled in the ranges of 395–405 MPa and 3.85%–3.90%,respectively,to meet the performance and quality requirements of the products.After optimizing the mechanical performance prediction process,the fluctuation amplitude of microalloying copper rods is reduced,which improves the stability of the product properties of copper rods.
作者
汤卫东
陈国平
张炎涛
张达
张李娜
王英华
TANG Weidong;CHEN Guoping;ZHANG Yantao;ZHANG Da;ZHANG Lina;WANG Yinghua(Zhongtian Alloy Technology Co.,Ltd.,Nantong 226000,China;School of Metallurgy,Northeastern University,Shenyang 110819,China)
出处
《江西冶金》
2024年第4期257-262,共6页
Jiangxi Metallurgy
基金
国家自然科学基金项目(52174319)。
