Zhang Huajun,Han Chongzhao
(School of Electronics and Information Engineering,Xi'an Jiaotong University,Xi'an 710049,China)
Abstract:To reduce the computational complexity of Volterra series on-line identification based on batch least square filter, and to save the data store spaces needed, an identification method based on recursive batch least square filter was proposed and applied to nonlinear fault diagnosis. By means of this method, the data store spaces can be saved via fixing the dimensions of observation matrix, and the computational complexity can be reduced by calculating the inverse of correlation matrix in a recursive way instead of calculating it directly. Meanwhile, in order to prevent the correlation matrix from becoming ill-conditioned, the concept of the effect factor was introduced to select the data such that the numerical stability of the identification was enhanced. The effectiveness of this identification method was illustrated by applying it to the fault diagnosis of certain type of helicopter's electric rudders.
Keywords:recursive batch least square filter;fault diagnosis;effect factor;electric rudder