基于IPSO-SVR算法的结构钢板激光弯曲成形预测  

Laser Bending Forming Prediction of Structural Steel Plate Based on Improved PSO-SVR Algorithm

作　　者：李文翠 Li Wencui(School of Architecture,Xinxiang Vocational and Technical College,Xinxiang Henan 453000,China)

机构地区：[1]新乡职业技术学院建筑学院,河南新乡453000

出　　处：《山西冶金》2025年第2期168-169,172,共3页Shanxi Metallurgy

摘　　要：为了提高高层建筑结构钢板材的激光弯曲成形质量,通过改进粒子群优化算法(PSO)对向量回归机(SVR)进行优化,并对钢板激光弯曲成形加工参数进行组合优化,获得满足板材弯曲角的最优组合参数。研究结果表明:无论是激光扫描速度还是激光功率,采用经过训练的模型预测与样本真实值很接近,证明该模型的准确性。扫描速度预测相对误差介于0.29%~7.41%,功率预测误差为1.26%~5.13%,该模型能较好地预测出激光弯曲成形能力。该研究有助于提高钢板的成形效率,具有很好的实际价值。In order to improve the laser bending forming quality of tall building structural steel sheet,the vector regression machine(SVR)was optimized by improved particle swarm optimization(PSO)algorithm,and the laser bending processing parameters of steel sheet were combined to optimize,and the optimal combination parameters satisfying the bending angle of the sheet were obtained.The results show that whether it is laser scanning speed or laser power,the trained model is close to the real value of the sample,which proves the accuracy of the model in this paper.The relative error of scanning speed prediction is 0.29%~7.41%,and power prediction error is 1.26%~5.13%.The model can predict the laser bending ability well.The research is helpful to improve the forming efficiency of steel plate and has good practical value.

关 键 词：结构钢板 激光弯曲 IPSO-SVR算法 预测误差 

分 类 号：TG665[金属学及工艺—金属切削加工及机床]