Design and Control of Capacitive Power Transfer Systems

Abstract

Wireless power transfer (WPT) has enjoyed growing popularity in the past few decades as it provides a flexible and effective way of delivering power to loads without any physical contact. Among all the forms of conducting WPT, inductive power transfer (IPT) technology is the most extensively used. IPT has made its way into many industrial and domestic applications. However, the magnetic field that IPT technology relies on brings diffculties when delivering power through metallic barriers. In addition, the widespread magnetic flux that creates electromagnetic interference (EMI) can become another concern in IPT. Capacitive power transfer (CPT) technology, as a substitute transfer medium, has received considerable attention in recent years. By employing an electric field, CPT is able to deliver power through metallic obstacles. Some other features CPT enjoys include its compact volume, low EMI and simple design. CPT technology has been studied widely but more attention is needed in relation to the control area of it. This thesis is mainly focused on the study and development of various control techniques for CPT systems. The CPT systems examined in this thesis have adopted a Class-E inverter approach for high-frequency AC signal generation due to the remarkably high effciency it can achieve. To begin with, the basic operations and theoretical constraints of the Class-E inverter are studied. The theoretical constraints relate to the minimum allowed shunt capacitance, which is crucial as it limits the selection of switching devices.The operation of a Class-E inverter is very sensitive to variations in circuit parameters. Motivated by this drawback, two auto-frequency tuning mechanisms based on PI control approach were developed. The first one aimed to maintain a high power transfer efficiency by keeping the zero voltage switching (ZVS) condition achieved continuously. The objective of the second auto-frequency tuning mechanism was to regulate the output voltage and minimise switching losses. These two designs address the above mentioned problem by making self-adjusting Class-E inverters, so the desired system performance can be guaranteed irrespective of variations in the circuit parameters. Next, optimal control techniques were applied to CPT systems. The first harmonic model of a CPT system consisting of a Class-E inverter and a full-bridge rectifer was developed by employing first harmonic approximation (FHA). The first harmonic model was then linearised at its operation point to yield a linearised state space model. Based on this linearised model, two optimal control theories were studied and applied to CPT systems. Firstly, a Linear Quadric Gauss (LQG) optimal control approach was explained and an LQG controller was designed. After that, a robust H1 output feedback controller was developed. In this design, the variations in the coupling capacitors and the load are described as polytopic uncertainties in the linearised state space model. A novel robust H1 output feedback controller design technique for the linear state space model with polytopic uncertainties was then proposed. The simulation and experimental results suggested that both of the LQG controller and the H1 controller tolerated the circuit uncertainties with guaranteed output voltage and minimised switching losses.