A Fault Parameter Estimation Procedure Based on User Defined Scilab Functions

5 A Fault Parameter Estimation Procedure Based on User Defined Scilab Functions 

For illustration of parametric fault isolation by means of least squares ARR residual minimisation, the simple hybrid network in Fig. 4.1 shall be considered once again. The all-mode FSM (Table4.1) indicates that resistors Ri and R2 have the same component signature when the switch is on so that a parametric fault in one of the two resistors cannot be isolated by inspection of the coherence vector. 

As a fault scenario it is assumed that the value of resistance R abruptly doubles at f = 2 s and remains at that value for f 2 s (Fig. 6.1). As soon as this parametric fault is detected, the values of the two parameters R and R2 are estimated over a time window from 2 to 8 s. That is, the changed parameter value remains constant over the observation window. 

If mode changes are not caused by externally contfolled signals as in the case of the simple network but happen autonomously then their time instant is not known in 

6 Isolation of Multiple Parametric Faults from a Hybrid Model 

---