Parameter estimation in nonlinear dynamic systems can be used for system identification, adaptive control, fault detection and identification. A new model-based parameter estimation technique in nonlinear dynamic systems is proposed which is based on the representation of system dynamics in terms of probability of transition between sets of user specified intervals of the state variables within user specified time intervals. This representation scheme does not require the linearization of the system equations and can be generated from almost any type of mathematical model which describes the system dynamics. The algorithms for the implementation of the proposed method lead to compact codes with low computational time and memory storage. The technique is illustrated using example non-linear nuclear and mechanical dynamic systems. The results show that the method: (a) is suitable for parameter estimation in conditions of noisy data, incomplete information on the system state, random variations of the parameters, and uncertainty in the mathematical model, and, (b) has potential for on-line use.
展开▼
