基于RBF神经网络的煤粉着火判别  

Judging Fire of Pulverized Coal Based on RBF Neural Network

作　　者：刘禾[1] 周茵[1] 石晓峰 

机构地区：[1]华北电力大学控制科学与工程学院,北京102206 [2]大唐首阳山电厂,洛阳471000

出　　处：《模式识别与人工智能》2005年第3期366-369,共4页Pattern Recognition and Artificial Intelligence

基　　金：华北电力大学博士学位教师科研基金(No.2003-008)

摘　　要：对煤粉火焰图像特征进行了分析,提取了反映燃烧状态的特征参数,在此基础上提出采用径向基函数神经网络的火焰图像煤粉着火判别方法。由于神经网络具有自学习特性,故判别方法所需调整的参数少。判别方法不仅能够着火/灭火判别,而且还可对火焰图像传感器进行故障判别,应用表明判别方法使用方便,判别正确率高,具有实际应用价值。The way to judge fire of pulverized coal is developed. The characteristics of pulverized coal flame image are analyzed and feature parameters reflecting the fire states are extracted. A method using Radial Basis Function (RBF) neural network to judge fire is presented. As the neural network has the self-learning ability, the method has few parameters to be adjusted. The method can judge fire is on/off, and diagnose failure of flame image sensor. Applications show that the method is convenient, has high right ratio, and can be put into practical use.

关 键 词：锅炉 火焰检测 燃烧 径向基函数神经网络 图像 

分 类 号：TM621[电气工程—电力系统及自动化] TP183[自动化与计算机技术—控制理论与控制工程]