AbstractPrompted by a desire to minimise losses between power sources and loads, the aim of this Thesis is to develop novel maximum power point tracking (MPPT) algorithms to allow for efficient power conversion within low carbon technologies. Such technologies include: thermoelectric generators (TEG), photovoltaic (PV) systems, fuel cells (FC) systems, wind turbines etc. MPPT can be efficiently achieved using extremum seeking control (ESC) also known as perturbation based extremum seeking control. The basic idea of an ESC is to search for an extrema in a closed loop fashion requiring only a minimum of a priori knowledge of the plant or system or a cost function.
In recognition of problems that accompany ESC, such as limit cycles, convergence speed, and inability to search for global maximum in the presence local maxima this Thesis proposes novel schemes based on extensions of ESC. The first proposed scheme is a variance based switching extremum seeking control (VBS-ESC), which reduces the amplitude of the limit cycle oscillations. The second scheme proposed is a state dependent parameter extremum seeking control (SDP-ESC), which allows the exponential decay of the perturbation signal. Both the VBS-ESC and the SDP-ESC are universal adaptive control schemes that can be applied in the aforementioned systems. Both are suitable for local maxima search. The global maximum search scheme proposed in this Thesis is based on extensions of the SDP-ESC. Convergence to the global maximum is achieved by the use of a searching window mechanism which is capable of scanning all available maxima within operating range. The ability of the proposed scheme to converge to the global maximum is demonstrated through various examples. Through both simulation and experimental studies the benefit of the SDP-ESC has been consistently demonstrated.
|Date of Award||2015|
|Sponsors||European Thermodynamics Limited|
|Supervisor||Malgorzata Sumislawska (Supervisor), Keith Burnham (Supervisor) & Joseph Mahtani (Supervisor)|