摘要
提出了一种从输入输出数据中在线辨识锅炉主蒸汽压力系统模型的新方法,它采用T-S模型结构,利用一种新颖的在线模糊聚类算法和改进的卡尔曼滤波算法,将有监督和无监督学习方法相结合,实现了模型的在线自适应辨识,它可在运行中自动学习,适应很大工况范围及锅炉特性的时变性.仿真结果验证了提出的建模方法的有效性。
A new method for the online identification of the main steam pressure system of a boiler from input-output data is proposed by adopting a T-S(Takagi-Sugeno) model structure.Through the use of a kind of innovative online fuzzy-clustering algorithm and an improved Kalman filter algorithm and by combining a supervised learning method with an unsupervised one online self-adaptation identification by the model has been realized.The model can conduct self-learning during operation and adapt to a very large range of operating conditions as well as the timevariation character of boiler characteristics.The results of simulation have verified the effectiveness of the model-building method put forward by the authors.
出处
《热能动力工程》
EI
CAS
CSCD
北大核心
2006年第2期197-200,共4页
Journal of Engineering for Thermal Energy and Power