| 引用本文: | 谢彦红,耿志成,李元.支持向量数据描述及降幅重构方法在间歇过程 故障分类中的应用研究[J].控制理论与应用,2015,32(3):380~387.[点击复制] |
| XIE Yan-hong,GENG Zhi-cheng,LI Yuan.Research on the fault classification based on support vector data description and drop reconstruction in batch process[J].Control Theory and Technology,2015,32(3):380~387.[点击复制] |
|
| 支持向量数据描述及降幅重构方法在间歇过程 故障分类中的应用研究 |
| Research on the fault classification based on support vector data description and drop reconstruction in batch process |
| 摘要点击 2813 全文点击 2435 投稿时间：2014-06-11 修订日期：2014-11-07 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/CTA.2015.40536 |
| 2015,32(3):380-387 |
| 中文关键词 间歇过程 故障分类 支持向量数据描述 降幅重构 核函数 |
| 英文关键词 batch process fault classification support vector data description (SVDD) decline reconstruction kernel function |
| 基金项目 国家自然科学基金重点项目(61034006), 国家自然科学基金项目(61174119)资助. |
|
| 中文摘要 |
| 针对间歇生产过程中的故障分类问题, 为进一步研究故障所属类型, 本文采用支持向量数据描述(support vector data description, SVDD)的方法. 在多种类型的故障数据库基础上, 应用SVDD建立对应故障种类的模型, 利用 核函数求出各个模型超球面半径; 对于新的待分类故障样本, 先考察其与各个种类模型超球面球心的距离, 再比较 此距离与半径的大小, 进而确定故障所属类型, 尤其是可能超出各个故障模型检测范围的待测故障样本, 对其进行 降幅重构迭代, 确定其所属类型. 该方法不但能够准确识别独立发生的故障, 而且对于其他方法难以识别的多种并发的故障也能够有效地实现分类, 应用于数值仿真和青霉素发酵过程实验中, 验证了其有效性和准确性. |
| 英文摘要 |
| For batch process fault classification problem and further studies of failure of a type, Support Vector Data Description (SVDD) method is adopted in this paper. On the basis of the database of many types of faults, corresponding fault types of models are built applying SVDD, then super-spherical radius of each model is obtained using kernel function; for the new fault samples to be classified, examining its distance to the center of the sphere hyper-sphere various types of models, then comparing the size of this distance with radius, and to determine the type of fault, especially for the tested fault samples may be beyond the detection range in each fault model, performing iterative reconstruction decline to determine their types. The method not only can accurately identify faults occurred independently, but also can be effectively achieved for a variety of other methods are difficult to identify concurrent fault, and it is used in penicillin fermentation process and numerical simulation experiments to verify its validity and accuracy. |
|
|
|
|
|