中文摘要：

由于热工过程往往具有非线性和不确定性,传统的线性建模方法难以精确表达其复杂特性。因此提出一种改进的基于满意模糊聚类的多模型建模方法。该方法不需要预先指定局部模型的个数即聚类数,它基于样本协方差矩阵的奇异值分解来确定初始聚类中心和新增聚类中心,并利用聚类有效性指标结合建模精确度要求来确定最佳聚类数。根据聚类结果可快速确定出局部模型网络的模型结构参数,进而采用基于加权性能指标的多模型辨识算法可得到各局部模型参数。对两个典型非线性系统和Bell-Astrom锅炉-汽轮机系统的建模结果表明,这种多模型建模方法具有辨识精确度高、子模型数少等优点。

英文摘要：

Due to the genuine nonlinearity and uncertainty, it is challenging to precisely present the com- plex dynamics of thermal process with traditional linear modeling strategies. To solve this problem, a multi- model modeling method based on an improved satisfactory fuzzy clustering technique is proposed. Without knowing the number of local models （i. e. , cluster number） as a priori ,the initial cluster centers and the incoming new cluster centers were first created by using singular value decomposition of the covariance matrix of samples, where the optimal clustering number was determined according to the combination of model accuracy requirement and cluster validity index. Then, the structure of local model network was de- termined directly from the clustering results, and the local model parameters were estimated with weighted performance function based identification algorithm. To outperform the proposed method well, some simu- lations were conducted on two popular nonlinear systems and Bell-Astrom boiler-turbine system. The re- suits suggest that the proposed method achieves high identification accuracy with less number of local models.