联合循环余热锅炉汽包水位晃荡信号智能预测  

Intelligent prediction of water level sloshing information of drum in combined cycle waste heat boiler

作　　者：谭季秋[1] 鄂加强[2] 钟定清[1] 

机构地区：[1]湖南工程学院机械工程学院,湖南湘潭411101 [2]湖南大学机械与运载工程学院,湖南长沙410082

出　　处：《中南大学学报（自然科学版）》2012年第3期966-971,共6页Journal of Central South University:Science and Technology

基　　金：湖南省自然科学基金资助项目(11JJ9011)

摘　　要：为揭示联合循环余热锅炉汽包水位晃荡机理,应用混沌分形理论,选择合适的滞时τ,对联合循环余热锅炉汽包水位晃荡信号时间序列进行相空间重构,得出联合循环余热锅炉汽包水位晃荡信号具有混沌特性,同时建立输入层节点数为3、隐含层节点数为5、输出层节点数为1的神经网络预测模型对联合循环余热锅炉汽包水位晃荡时间序列进行预测。研究结果表明:混沌优化算法对联合循环余热锅炉汽包实验水位晃荡信号预报精度高,经过1000次粗搜索迭代和761次细搜索迭代,可将误差降低至101.7。In order to revel the mechanism of water level sloshing of the drum in the combined cycle waste heat boiler, using chaos theory in the water level and selecting a suitable value of time lag r, the time-series embedding space about water level sloshing information were rebuilt. In the meantime, when the neural network model including input layer is 3, the hidden layer node is 5 and the output layer node is 1, the sloshing water level time series were predicted. The results show that prediction precision is very high when chaos optimization algorithm is applied in optimization of the parameters of BP neural network, and its prediction errors can be decreased to 10^-1,7 after 1 000 coarse search iterations and 761 fine search iterations.

关 键 词：余热锅炉 汽包 水位晃荡信号 智能预测 

分 类 号：TK316[动力工程及工程热物理—热能工程]