集员辨识与T-S模型相结合的非线性系统建模及其故障检测算法

Nonlinear System Modeling and Fault Detection Algorithm Using Set Membership Identification and T-S Model

针对带有未知但有界(Unknown But Bounded-UBB)噪声的非线性系统的建模及其故障检测问题,提出了一种集员辨识与T-S模糊模型相结合的非线性系统建模及其故障检测算法。在建立非线性系统模型时,利用系统正常状态下的运行数据,选用T-S模型对其进行离线建模。首先采用模糊聚类的方法对输入空间进行模糊划分,然后利用T-S模型为参数线性模型的特点,使用参数线性集员辨识算法辨识T-S模型的结论参数。由于集员辨识算法所得到的是参数的集合估计,在系统运行过程中,可以很方便地利用所建模型预测实际系统的输出范围,如果测量所得实际系统的输出不在预测输出范围之内,则可判断系统发生了故障。通过与其他算法相比,验证了本方法的性能。

A modeling method is proposed and applied in fault detection for nonlinear systems with unknown but bounded noises. This method used a set membership identification algorithm to build Takagi-Sugeno （T-S） fuzzy models of nonlinear systems. After some input and output data of a system were obtained when it run without a fault, the input space was partitioned using a fuzzy clustering algorithm, and then the consequence parameters of the T-S fuzzy model of this system were estimated using a linear-in-parameter set membership identification algorithm. Since the result of the estimation was a set of parameters, it could be easily used to predict the interval of the actual system output. If the measured output was out of the predicted interval, it could be determined that a fault had occurred. Simulation experiments are performed, showing the performance of our method as compared to the other method.