| Summary: | The conventional methods of observer poles placement in sensor fault detection usually adopt
the trial-and-error methods. These methods cannot achieve global optimal performance because of their fixed
poles placement and it leads to an observer with constant parameters, which could be reducing the system
performance. Therefore, this paper proposes a fuzzy-based observer tuning method to optimize and adapt the
selection of poles locations to determine the optimal gains of the observer, and it is experimentally applied to
a composite sensor fault detection. Fuzzy logic is a promising method that could overcome the trial-and-error
method challenges by introducing better adaptation and system robustness. The proposed observer structure
includes adaptive tuning corresponding to an unknown input. Utilizing self-tuning for the observer correction
stage, the gain is going to be updated online using the proposed fuzzy adaptive poles placement (FAPP)
system. This paper validated the system simulation by implementing fault detection algorithms by using a
real-time embedded observer-based system. The experimental results demonstrate the effectiveness of the
proposed fuzzy-based observer schemes at detecting sensor faults in the Brushless DC (BLDC) motors, with
significantly better performance than conventional counterparts’ methodologies. The experiments indicate
that the average estimation error is 0.146, which less by 43.8% than was obtained for high levels of noise
and disturbances compared with the traditional Luenberger observer approach.
|
|---|