中文摘要：

建立热力系统逆动力学过程模型是热力系统逆动力学研究及应用的关键。该文报告了一种锅炉汽温对象逆动力学过程模型的模糊在线辨识方法与结果。通过聚类和竞争学习算法，对逆动力学模型输入数据空间进行分区，在每个局部的数据子空间上，利用递推最小二乘辨识算法建立逆动力学过程模糊规则，并通过自适应模糊推理实现系统输入过程的反演。仿真结果表明，所建立的逆动力学过程模型对时变汽温对象具有良好的自适应能力和在线跟踪能力；通过汽温对象逆动力学过程在线辨识，能够获得恰当的控制过程，保证系统输出温度按照预定的轨迹达到设定值。

英文摘要：

It is crucial to establish the inverse dynamic process model of thermal system on research and application of the inverse dynamic process of thermal system. A method and result of on-line fuzzy identification on the inverse dynamic process model of steam temperature of boiler was reported, The method partitions the input data space of the inverse dynamic model into some local regions by the clustering and competitive learning algorithms, then fuzzy rule based the inverse dynamic process is built for each local regions by the recursive least-square identification algorithm（RLS）, and the inversion of system input process is actualized by the adaptive fuzzy inference. The simulation results show that the established inverse dynamic model has good self-adaptation ability and on-line traceable ability to time variant steam temperature object. Through the on-line identification of inverse dynamic process of steam ＂temperature object, it is possible to acquire appropriate controlling process and ensure the system＇s temperature output to arrive at the set value according to reference path.