矿物燃料锅炉受热面污染模式的分级自组织识别  

Hierarchical Self-Organizing Recognition of Mineral Fuel Boiler's Heating Surfaces Pollution Pattern

作　　者：沈曙光[1] 王广军[1] 陈红[1] 

机构地区：[1]重庆大学动力工程学院,重庆400044

出　　处：《计算机仿真》2006年第5期134-137,共4页Computer Simulation

摘　　要：在充分考虑锅炉具体运行工况的前提下构造广义的污染模式特征向量,有效地避免了运行工况变动对污染模式识别的影响。广义特征向量表达了锅炉系统运行参数与运行工况数间的数值对应关系,因而决定了系统的一种特定污染模式。利用SOM网络可以对这种隐式的对应关系通过自组织进行有效地类聚并加以识别。采用多个SOM网络逐级模式识别方案,有效地提高了锅炉污染模式识别效果,并可以确定发生污染的具体部位。Based on the sufficient consideration of work conditions, the eigenvector for analyzing pollution pattern is constructed. Since that, the effect of work condition variation on the pollution mode identification can be avoided. The generalized eigenvector shows the relation between operation parameters and the values of work condition for a boiler, at the same time, one kind of particular pollution pattern is presented. SOM network is applied to cluster and identify the implicit correspondence relationship effectively by self - organizing. By using hierarchical pattern recognition with much SOM network, the identification effect of boiler pollution mode is improved, and the accurate locations can be acquired.

关 键 词：锅炉 受热面污染 模式识别 自组织映射 

分 类 号：TP271[自动化与计算机技术—检测技术与自动化装置]