基于分布信息融合的直流锅炉燃料量信号重构  

RECONSTRUCTION FOR FUEL SIGNAL OF ONCE-THROUGH BOILER BASED ON DISTRIBUTED DATA FUSION

作　　者：于达仁[1] 范轶[1] 徐志强[1] 

机构地区：[1]哈尔滨工业大学能源科学与工程学院,黑龙江哈尔滨150001

出　　处：《中国电机工程学报》2004年第2期191-195,共5页Proceedings of the CSEE

基　　金：国家自然科学基金项目(59976008)~~

摘　　要：通过对超临界直流锅炉燃料量间接测量信号频谱特性的研究,指出了不同信号具有互补的频谱特性。基于广义观测器结构,利用分布信息的互补特性设计了分布信息融合算法。该文利用沿流程各汽温信号稳态精度高和给粉信号响应速度快的互补特性,经过信息融合,解决了给粉量测量的困难,提高了测量的快速性和准确性,从而有效地提高了直流锅炉机组燃水比调节系统的控制品质。通过对融合算法鲁棒性研究。The frequency spectrums of some indirect fuel measures in supercritical once-through boiler are analyzed. It is concluded that frequency spectrums of different measures are complementary. An algorithm for distributed data fusion is proposed based on the complement of distributed data and architecture of generalized observer. The paper synthesizes the complementary information of high steady precision of steam temperature measures along the fluid and high response speed of fuel measures. The problem measuring transient fuel is solved with data fusion and the response speed is improved while accuracy is still maintained. This method could considerably improves the control quality of supercritical once-through boiler fuel/water. The robustness of fusion algorithm is analysized and it is concluded this method could depress the influence of one-measure invalidation.

关 键 词：超临界直流锅炉 燃料量 信号重构 信息融合 控制系统 分布参数特性 

分 类 号：TK229.54[动力工程及工程热物理—动力机械及工程]