PhD thesis
This thesis discusses the proposal of nonlinear deadbeat control for continuous conductionmode (CCM) boost converter and the experimental verification. First, the nonlinear state equa-tion is derived, and second a nonlinear current reference deadbeat control is proposed. Third,a new nonlinear controller to implement the load disturbance compensation is proposed. Afterthe simulations using PSIM software and verifications by experiments, it was confirmed thatunder the conditions of an input voltage 12 V, an output voltage of 20 V, a load resistance of 4Ω and a sampling frequency of 100 kHz, the voltage command tracking capability of a settlingtime of 280μs was achieved, and an output voltage recovery time of 1.46 ms was achieved fora sudden unknown load change. Mathematical analysis was performed and confirmed asymp-totic stability and robustness of the control method during voltage and current perturbation,disturbance occurrence and parameter variations. It was confirmed that the voltage and cur-rent errors eigen values converge towards inside of the unit circle thus maintaining asymptoticstability for each perturbation case investigated. Methods to design the controller parameterswere stipulated to be within the physical realization and can be applied to boost converterof any application in CCM. The proposed control method was compared with other literaturethat applied different digital control methods to boost converters of various applications. Itwas found that nonlinear deadbeat control proposed in this thesis was about twice as fast forreference tracking response, and could reject disturbances quickly for a load current three timesbigger than other literature. Therefore, it is concluded that these data are the best even thoughthe proposed control is based on nonlinear equations. Few differences were observed betweenexperiments and simulations. Upon investigations, those differences were found to be caused bytime delay in the switching device and other un-modeled nonlinear switching device phenomena.Future work will be focused on improving the control method to compensate for those observednonlinearities.
Kawamura Laboratory, Yokohama National University