摘要
在钢铁材料产品质量检验中,硬度与抗拉强度是两个最常用的力学性能指标,它们之间存在一定的正相关关系.基于最小二乘支持向量机(LS-SVM)原理,结合遗传优化算法(GA),建立材料洛氏硬度作为输入值和抗拉强度为输出值的模型,对低碳钢的洛氏硬度与抗力强度之间的关系建立模型并分析.结果显示,应用GA-LSSVM建立的数学模型,可通过硬度预测抗拉强度,实验值与模型值的最大相对误差为0.237 2,均方误差为0.008 4,从而证明此模型的精确性和适用性.
In the quality inspection of steel materials, hardness and tensile strength were two main me- chanical properties, and there was the positive correlation between them. Based on least squares support vector machine （LS-SVM） principle coupled with genetic algorithm （GA） optimization strategy, with Rockwell hardness as input parameter and tensile strength as output parameter, the predictive model of the relationship between them was developed and analyzed. The results indicated that the GA-LSSVM model can be capable of capturing the relationship between them, the tensile strength can be predicted by Rockwell hardness. The relative maximum error between experimental and model value was 0. 237 2, and the mean square error was 0. 008 4, which verified the accuracy and validity of the model.
出处
《中北大学学报（自然科学版）》
CAS
北大核心
2016年第3期258-261,278,共5页
Journal of North University of China(Natural Science Edition)
基金
部级预研基金项目
