中速直吹磨组径向基函数神经网络信息融合-模糊启动控制  

Radial Basis Function Neural Network-data Fusion and Fuzzy Control During Direct-firing Medium-speed Mill's Start-up

作　　者：周洪煜[1] 张振华[1] 陈晓锋 

机构地区：[1]重庆大学动力工程学院,重庆市沙坪坝区400030 [2]广东省电力设计院,广东省广州市510663

出　　处：《中国电机工程学报》2011年第11期119-125,共7页Proceedings of the CSEE

摘　　要：中速直吹磨煤机组的大延迟及惯性、燃料量的直接测量和发热量的在线检测是保持超临界燃煤发电机组燃烧控制稳定性和安全性的难点。通过讨论某厂直流锅炉燃料量控制及其中速直吹磨煤机组启动控制系统,引入基于径向基函数神经网络的信息融合方法重构燃料发热系数,以在线精确修正实际燃料量,并以此为基础利用模糊控制的鲁棒性和克服非线性能力,优化了对磨组启动过程中咬合时间、电流等不确定参数的控制。实验结果表明,此方案中燃料量控制可兼顾快速性和准确性,在磨组启动中模拟量控制系统和燃烧器管理系统可更好地协调控制,并保证了燃烧工况的平稳过渡。Direct-firing medium-speed mills＇ long tune lag and large inertia, besides fuel flow＇s direct measure and its heat calories＇ online detection are key issues to remain the stability and security while controlling combustion, especially on supercritical coal fired power units. By discussing a certain once through boiler＇s fuel flow and medium-speed mill＇s start-up control system, a new control scheme about it was proposed. Using a data fusion method based on radial basis function neural network（RBFNN）, it reconstructed heat coefficient so as to exactly correct actual fuel flow online. And based on this, through introducing fuzzy control and using its robusticity and capability of conquering nonlinearity, it optimized the control of uncertain parameters such as coal-bited time and current during mill＇s start-up. Results indicate this scheme＇s celerity and precise on fuel flow control, advantages on mill＇s start-up coordinated control between modulating control system and burner management system as well as stable transition of the combustion situation.

关 键 词：燃料量 信息融合 径向基函数 模糊 中速磨 超临界机组 

分 类 号：TK32[动力工程及工程热物理—热能工程]