作　　者： ; ; ; (陈晓锋）;

机构地区： 重庆大学动力工程学院

出　　处： 《动力工程学报》 2010年第11期833-837,共5页

摘　　要： 针对超临界锅炉燃料量不易准确测量和燃料发热量难以在线检测的问题,通过分析某厂直流锅炉燃料量控制系统缺点,设计了基于径向基函数神经网络(RBFNN)-信息融合方法的燃料量优化控制方案.该优化控制综合了RBFNN的强自学习、并行计算能力和信息融合方法的互补性、冗余性等特点,重构燃料发热量系数,在燃烧内扰时用来在线精确修正实际燃料量,外扰时作为燃料量需求的前馈控制,以保证燃料量实时符合机组负荷变化或锅炉内扰下各工况的实际需求.通过磨煤机切换内扰下燃料量控制实验和负荷变换外扰下燃料量控制实验表明:在内外扰动工况下,此控制系统可保证燃料量控制的准确性、快速性和稳定性. Aiming at the difficulties in direct measurement of fuel flow for supercritical boiler and the online detection of the fuel heat value,and by analyzing deficiencies of the fuel flow control system for a certain once-through boiler,a new optimized scheme based on radial basis function neural network（RBFNN）data fusion method is proposed,which reconstructs the fuel heat coefficient by combining RBFNN＇s strong self-study and parallel-computing capability with data fusion＇s complementarity and redundancy.It can exactly online correct the actual fuel flow during internal disturbance and realize feedforward control on fuel flow demand during external disturbance,thus meet the instant fuel flow demand of varying unit load or internal boiler disturbance.Experimental results indicate that this system has high accuracy,rapidity and stability in fuel flow control under both internal and external disturbance conditions.

关 键 词： 超临界锅炉 燃料量 信息融合 径向基函数 神经网络

领　　域： [自动化与计算机技术] [自动化与计算机技术]