Pulse system dynamics forecasting using on-line time series data
Résumé
An approach to dynamics forecasting is proposed by an example of DC-DC PWM converter dynamics simulation. The main algorithm of this approach allows the on-line identification of sub-harmonic motion types and corresponding parameter vectors. The consistency of the identification results
forms the tendency of the parameter variation. Then forecasting of dynamics evolution with the nearest bifurcation boundaries, constructed preliminary, is suggested.