运用计算机图像处理和神经网络技术对炉膛火焰进行诊断  

FIRING DIAGNOSIS OF FLAME IN THE FURNACE OF UTILUTY BOILERS BY MEANS OF COMPUTETER-BASED IMAGE PROCESSING AND NEURAL NETWOEK TECHNOLOGY

作　　者：王飞[1] 马增益[1] 严建华[1] 王新军[1] 赵敬德[1] 丁经纬[1] 倪明江[1] 岑可法[1] 

机构地区：[1]浙江大学,浙江杭州310027

出　　处：《热力发电》2003年第2期24-28,共5页Thermal Power Generation

摘　　要：研究利用火焰图像处理技术进行炉膛火焰燃烧诊断的方法。根据火焰图像处理系统获取燃烧图像 ,并提取用于燃烧诊断的火焰图像特征参数 ,构造用于燃烧诊断的BP神经网络 ,将提取的火焰图像特征参数作为神经网络的输入量 ,通过训练和测试 ,进行火焰燃烧状态的预测。开发的火焰图像处理和燃烧诊断系统已在 30A firing diagnosis method of flame in the furnace carried out by means of flame image processing technology is studied. Based on the firing image obtained from the flame image processing system, and property parameters of flame image collected for firing diagnosis, a BP neural network used for firing diagnosis was constructed. Then, the collected property parameters of flame image are used as input values of the said neural network, via training and testing, to carry out prediction of the flame firing state. The developed flame image proccessing and firing diagnosis system has already been used on a 300MW utility boiler.

关 键 词：计算机 图像处理 神经网络 炉膛火焰 诊断 锅炉 燃烧诊断 

分 类 号：TK227.1[动力工程及工程热物理—动力机械及工程] TP391.41[自动化与计算机技术—计算机应用技术]