Abstract:
Inductive Power Transfer (IPT) systems provide the ability to transfer power across a relatively large air gap without contact. Most traditional IPT systems are similar in that they consist of a primary ac current owing through a primary coil, allowing power to cross the air gap. This often has known or controlled coupling with limited variation, so that the system design can be carefully managed to ensure highly efficient transfer to a secondary coil. Controlled outputs are most commonly dc but when an ac output is required another conversion stage is usually added. More recently however some applications have evolved so that their now exists unavoidable mis-tuning due to movement or control - these include the applications of stage lighting and electric vehicle (EV) charging. This thesis specifically focuses on these two applications since both generate unwanted and unavoidable reactive power which must be provided by the primary supply and usually can only be compensated on the primary side. For stage lighting applications a new class of controllers are evaluated which generate inherent reactive power back to the primary when supplying power across its output control range. Features such as sinewave output (unique in IPT), reduced harmonic content and distortion, and higher operating frequencies compared to existing controllers provide significant advantages, however when all lights are turned on simultaneously an accumulated effect places strain on the input VA to the primary power supply. The second focus of the thesis is modelling of stationary battery charging systems for electrical vehicles through a lumped IPT system. A battery charging system is most often designed for a specific secondary compensation but does have some tolerance built-in (10-15%) for positioning variations affecting coupling. However in future one primary power supply must be capable of delivering power to different compensated secondaries, and these differences amplify the mis-tuning which must be tolerated. To understand the impact of Volt-Ampere Reactive (VAR) generation in an IPT primary the proposed lighting circuits and typical EV charging system are carefully analysed to determine the likely reflected impedance and the additional input VA requirements. This is achieved through the use of mathematical modelling and analysis, simulations and experimental results. The parallel ac lighting circuit uses two uniquely related resonant tanks to create an almost harmonic free sine wave output. This is a very different approach from traditional IPT controllers, however under varying load it reflects a variable impedance on to the primary. The reactive part of this reflection appears as a capacitive load to the primary power supply, with the only exceptions near zero or at rated output power when the primary power supply loading becomes inductive. The proposed series ac lighting circuit used two resonant tanks to control its output and its reflected impedance characteristics load the primary capacitively over the complete operating range. Primary side control is the only option for these controllers due to the importance of the circuits changing impedance for control. Steady state lumped models are developed reducing these 4th order circuits down to 2nd order models. GSSA modelling is used to develop dynamic models for future control designs. Operational variations resulting from stationary EV charging misalignment, and secondary magnetic systems which are no longer matched using series or parallel tuning, provided criteria to investigate VAR mitigation on the primary side. Selected approaches to fixed primary tuning are compared in terms of input VA with mis-alignment. General models were developed that include tuning terms that allow for the adjustment of frequency to tune the system, along with changes to the primary inductance. A term for tuning the secondary inductance is included for completeness, however secondary side control is not considered in this thesis due to operating restrictions in the lighting circuits and commercial restrictions associated with EV charging and therefore was henceforth ignored. The developed models were used to provide general insights into mis-tuning effects on both the primary and secondary and includes potential impact of direct compensation or frequency based tuning. The approach used keeps the output power constant and monitors changes in the input VA to enable an easy comparison with changes in charging position, secondary compensation or circuit operation. The primary tuning strategies also investigated adjustment to both the primary compensation and frequency variation. It was shown that if frequency could be changed to bring the secondary closer to its ideal tuning, this would improve the input VA and require less additional compensation to obtain a tuned primary. While absolute tuning might not be the aim, these results show that this type of strategy could work and gives a strong foundation for further work.