Estimation of parameter drift rate for detection and prediction of faults

System faults may be detected and predicted via estimation of trajectories of the system dynamic parameters in the parameter space. During normal operation, the parameters exhibit random motion around their nominal values. A drift of one (or more) parameters may indicate a fault. Two algorithms are developed here for the estimation of the parameter drift rate. The performance of the developed algorithms is also evaluated.