摘要
为了解决传统汽轮机及调节系统参数辨识方法周期长、适应性差等问题,基于系统参数对输出响应的影响程度(参数轨迹敏感度),采用粒子群算法,提出了一种"粗细结合"的参数辨识策略。该辨识策略根据模型中待辨识参数对评价指标参数的敏感度大小,提出参数"粗调节"和"细调节"区域,有效定位各参数的变化范围,以提高辨识算法的寻优效率。分别以理论参数下的响应数据和某300MW机组实测数据进行参数辨识,相较传统多参数辨识方法的结果表明,提出的"粗细结合"参数辨识策略在寻优速度和辨识精度上均具有明显的优势,既保证了整体模型的精确性,又确保了中间环节参数的合理性。该辨识策略为汽轮机及调节系统的参数辨识提供了一种新的高效辨识手段。
Since most of the traditional parameter identification methods used in the steam turbine speed governor system have the shortages of poor fitness and long period,a novel identification scheme,ranking the importance of parameters according to parameter sensitivity,then locating the range of the most important parameter in the first stage and positioning less important parameters in the second stage using particle swarm optimization,is proposed in this paper. The performance of this new identification scheme was verified through the comparisons of the steam turbine speed governor system identification results with traditional identification method using theory test data and measured data from a 300 MW thermal power unit. The results show that during identification process the scheme has higher convergence speed and is more precise comparing with the traditional parameter identification method,and it brings a new scheme for steam turbine speed governor system identification.
作者
钟晶亮
甘飞
邓彤天
ZHONG Jing-liang;GAN Fei;DENG Tong-tian(Guizhou Electric Power Test ＆ Research Institute, Guiyang 550002, China;School of Power Engineering, Chongqing University, Chongqing 400044, China)
出处
《汽轮机技术》
北大核心
2018年第2期111-115,共5页
Turbine Technology
关键词
汽轮机
调节系统
粒子群算法
参数敏感度
辨识策略
steam turbine
speed governor system
particle swarm optimization
parameter sensitivity
identification scheme
