基于遗传算法的超超临界机组过热器动态模型参数辨识

作者: 胡文凯, 方彦军, 李鑫, 熊扬恒
成果类型: 期刊论文
发表日期: 2011
年卷期页: 2011,v.40;No.299,(11 ),34-38
期刊名: 《热力发电》
中图法分类号: TK223.32;TP18
关键词: 超超临界机组;过热器;主蒸汽温度;遗传算法;模型参数辨识
英文关键词: ultra-supercritical unit;superheater;main steam temperature;genetic algorithm;model parameter identification
摘要: 针对超超临界机组过热器MISO动态模型参数辨识问题,提出了一种基于遗传算法的模型参数辨识方法。将遗传算子改进为排序保优选择、非均匀线性交叉和高斯变异,引入均匀设计方法构造初始种群,采用Sigmoid函数适应性调整交叉和变异概率,以提高遗传算法的收敛精度和运算速度。对海门电厂超超临界机组运行数据进行参数辨识仿真的结果表明,该改进遗传算法具有较好的运算性能,预测的输出变化与主蒸汽温度实际变化基本一致,能够反映高温过热器的动态特性。
英文摘要: Directing against the MISO dynamic model of superheater for ultra-supercritical units,a model parameter identification method based on genetic algorithm(GA) has been put forward.The genetic operators have been improved as displacement sequence with ensuring quality selection,non-uniform linear crossover,and Gaussian mutation,the uniform design method is used to construct the initial sort group.The adaptability of Sigmoid function is used to adjust the probabilities of crossover and mutation,so as to enhance the convergence precision and operational speed of GA.The result of parameter identification emulation for operation data of ultra-supercritical unit in Haimen Power Plant shows that the improved GA method boasts better operational performance,the predicted output variation tallies basically with the actual variation of main steam temperature,can reflect the dynamic properties of high temperature superheater.
语言: chi
相关文献
作者其它文献
My JSP 'reqreturndiv.jsp' starting page