基于人工神经网络的大型电厂锅炉飞灰含碳量建模  

ARTIFICIAL NEURAL NETWORK MODELLING ON THE UNBURNED CARBON IN FLY ASH FROM UTILITY BOILERS

作　　者：周昊[1] 朱洪波[2] 曾庭华[2] 廖宏楷[2] 岑可法[1] 

机构地区：[1]能源清洁利用和环境工程教育部重点实验室,浙江大学热能工程研究所,浙江杭州310027 [2]广东省电力集团公司,广东广州510600

出　　处：《中国电机工程学报》2002年第6期96-100,共5页Proceedings of the CSEE

基　　金：国家重点基础研究专项经费项目(G1999022204)

摘　　要：飞灰含碳量是影响锅炉热效率的一个重要因素,但飞灰含碳量受煤种、锅炉设计结构、操作参数等多种因素影响,关系复杂。煤种变化往往导致燃烧工况偏离试验调整获得的最优值。在对某台300MW四角切圆燃煤电厂锅炉飞灰含碳量特性进行多工况热态测试的基础上,应用人工神经网络的非线性动力学特性及自学习特性,建立了大 型四角切圆燃烧锅炉飞灰含碳量特性的神经网络模型,并对此模型进行了校验。With the developing demand for high efficiency of the utility boilers, more attention is paid to the unburned carbon content in the fly ash from the high capacity tangential firing boiler, but the unburned carbon content in the fly ash is complicated and it is affected by many factors, such as coal character, boiler's load, air distribution, boiler style, burner style, furnace temperature, excess air ratio, pulverized coal fin-eness and the uniformity of the air and coal distribution, etc. In this paper, the unburned carbon content in the fly ash of a 300MW utility tangentially firing coal burned boiler is experi-mental investigated, and taking advantage of the nonlinear dy-namics characteristics and self-learning characteristics of artifi-cial neural network, an artificial neural network model on un-burned property of the high capacity boiler is developed and verified.

关 键 词：人工神经网络 大型电厂 锅炉 飞灰含碳量 建模 

分 类 号：TM621.2[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]