基于数据驱动的氢粉碎工艺氢含量的预报  

Forecasting of Hydrogen Content in Hydrogen Decrepitation Process Based on Data Driven

作　　者：朱林[1] 吕品[1,2] 

机构地区：[1]内蒙古科技大学信息工程学院,内蒙古包头014010 [2]中国石油集团工程设计有限公司辽阳分公司,辽宁辽阳111003

出　　处：《工矿自动化》2010年第10期70-72,共3页Journal Of Mine Automation

摘　　要：针对目前钕铁硼氢粉碎工艺过程中氢含量无法在线检测、动力学模型难以建立等问题,提出利用数据驱动方法实现钕铁硼氢粉碎过程中氢含量的预报,详细介绍了基于LS-SVM的氢粉碎工艺氢含量预报模型的建模过程。预测结果表明,基于LS-SVM的氢含量预报模型对氢粉碎工艺中氢含量的预测误差较小,可满足生产工艺的要求。In order to solve problems that hydrogen content cannot be detected on-line and kinetic model is difficult to be established in NdFeB hydrogen decrepitation process at present,the paper proposed a scheme which uses data driven method to implement forecasting of hydrogen content,and introduced modeling process of forecasting model of hydrogen content based on LS-SVM in details.The forecasting result showed that the model can forecast hydrogen content with small errors which would meet requirement of NdFeB hydrogen decrepitation process.

关 键 词：钕铁硼材料 氢粉碎工艺 氢含量预报 在线检测 数据驱动 LS-SVM模型 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]