中文摘要：

锅炉汽温对象具有明显的非线性和不确定性,在许多情况下很难用精确的数学模型来表达,或者所建立的非线性数学模型难以应用于汽温控制系统设计.该文以T-S模糊规则模型为基础,通过熵方法和竞争学习算法对输入空间进行聚类,利用递推最小二乘辨识算法（RLS）确定模型的结论参数,实现了汽温对象的在线模糊辨识.通过两台锅炉汽温对象在线辨识实例,验证了在线模糊辨识方法对于两类典型汽温对象的有效性,不仅具有较高的辨识精度,同时还具有较为理想的泛化性能和跟踪能力.

英文摘要：

With its obvious nonlinearity and uncertainty, it is difficult to describe the steam temperature of boiler by accurate mathematical model or the mathematical which is established can＇t be applied to design steam temperature control system. Based on T-S fuzzy rule model, the input data is clustered by the entropy method and competitive learning algorithm, the ultimateness parameter is ascertained by the recursive least-square （RLS） and the on-line fuzzy identification of steam temperature is achieved. Through the examples of on-line identification of two boilers steam temperature object, the effective of on-line fuzzy identification method to the two kinds typical steam temperature object is proved, it not only has upper identification precision, but also has quite perfect generalizable performance and traceable ability.