神经网络在超临界锅炉热偏差优化调整中的应用  

Neural Network in Supercritical Thermal Deviation of Boiler Optimization and Adjustment Application

作　　者：陈端雨[1] 崇培安[1] 陶丽[1] 丁士发[1] 

机构地区：[1]上海发电设备成套设计研究院,上海200240

出　　处：《锅炉技术》2013年第4期4-8,共5页Boiler Technology

摘　　要：将神经网络理论运用到超临界锅炉热偏差燃烧调整中,用神经网络模型预测锅炉末式过热器屏问热偏差,通过神经网络的学习,模拟各种风门挡板开度、锅炉运行参数等,并通过单个参数的连续调整找出影响锅炉屏间热偏差的主要因素,为降低主蒸汽热偏差提供指导依据和方向。这种方法不仅大大减小了热态调整的时间提高了效率,同时通过数值试验模拟减少燃烧调整对锅炉运行的影响,通过输入参数连续变化对锅炉燃烧的影响,为锅炉的热偏差调整提供便捷、准确全面的试验信息。This paper introduces the neural network theory to the supercritical boiler thermal deviation of combustion adjustment,predict the thermal deviation of the supercritical boiler Final superheater with neural network model,through the network to simulate running conditions such as air door,working conditions of coal grinding machine,and using a single parameter of continuous adjustment to find the main factors which effects the thermal deviation,providing guidance and direction for reducing the main steam heat deviation.At the same time,this method not only greatly reduced thermal adjustment time and improving efficiency,but also reducing the impaction on boiler running through the numerical simulation of combustion adjustment,it provides a convenient,accurate and comprehensive information by Observation of the input parameters of continuous variation on boiler combustion effect for the boiler combustion adjustment.

关 键 词：超临界锅炉 燃烧调整 热偏差 神经网络 

分 类 号：TK229.2[动力工程及工程热物理—动力机械及工程]