项目名称: 煤粉燃烧的图像与数据融合检测及其控制方法研究
项目编号: No.60874096
项目类型: 面上项目
立项/批准年度: 2009
项目学科: 能源与动力工程
项目作者: 张小刚
作者单位: 湖南大学
项目金额: 22万元
中文摘要: 煤粉燃烧广泛存在于工业生产的各个领域,燃烧过程的实时检测和最优控制是影响产品质量和生产能耗的关键技术因素。课题以工业氧化铝回转窑烧结为实际研究对象,研究燃煤工艺过程数据的智能分析检测和现场应用。提取煤粉燃烧火焰图像特征,结合相关热工数据,有效地检测煤粉燃烧温度和燃烧工况、并且从中提取有意义的模式规则,辨识回转窑煤粉燃烧过程工况、优化控制参数、实现更加稳定高效的窑前过程控制。课题研究了关联挖掘、粗糙集、免疫克隆等软计算方法在窑前数据分析中的应用,重点研究了窑前燃煤火焰和物料的非接触测量图像软测量方法,利用HMM、SVM,贝叶斯分类器等建立窑前火焰和热工数据的工况识别模型,建立基于多源异类信息融合和面向煤耗最优的智能专家控制系统,形成了专有技术。最后将上述理论成果成功应用到了某大型氧化铝回转窑生产现场,取得较好的应用效果。
中文关键词: 煤粉燃烧;回转窑;图像处理;软计算
英文摘要: The coal combustion exists in many industry processes. The realtime detection and optimizing control for the coal burning process are the sticking point for product quality and coal consume. The sintering temperature and its process characters of coal's combustion in alumina rotary kiln are measured through burning image sequences and correlative process data. The Immune ,level wise search and other soft computing algorithms are studied to classify the burning process and optimize control parameters.The coal combustion's image feature extraction is studied by CCD image sequences. HMM,SVM and Bayesian network models are used to extract the characters of process data. The expert control rules and optimizing experience are extracted from the fusion study of process data and image to achieve a higher robust and stability for the coal's combustion process control. A final expert controller of the rotary kiln is built based on the fusion studies of image and data to optimize the coal's consumption.
英文关键词: coal combustion;rotary kiln;image process;soft computing