Modeling, analysis and control of high-order switching power converters

The output voltage gain of the conventional dc-dc boost converter is often limited as operating this converter at some extremely high values of the duty ratio may lead to a reverse recovery problem of the diode and little room for control when dealing with load and line disturbances. To solve this problem, many transformer-based and transformer-less topologies have been proposed over the past few years to achieve a sufficiently high gain at some lower values of the duty ratio. However, if the industrial application does not demand for dc isolation, the use of a transformer-based dc-dc converter would only increase the overall cost and size of the system. Also, the efficiency degrades due to the losses associated with the secondary winding of the transformer. This makes the transformer-less topologies a more interesting choice and to date, a number of such topologies have been reported in the relevant literature. The objective of this thesis is to study the various modeling and control related aspects of such high-order non-isolated dc-dc boost converters. Towards this end, several control schemes are studied. As compared to the conventional dc-dc boost converter wherein there are only two state variables available for feedback purposes, the newly developed high-order topologies present certain challenges. Since they contain more inductors and capacitors to achieve the high gain, they present more state variables which can be used for feedback purposes. Therefore, it is necessary to select the most suitable state variables when designing the controllers for these converters. To address this issue, a detailed comparative study of two current-mode controllers (using the input and output inductor currents of the converter) is presented for some high-order dc-dc converters such as the positive output elementary Luo (POEL) converter and the 2-stage cascade boost converter. The objective is to find the most suitable inductor current for the controller design. The analyses of the current-controlled systems are carried out using the state-space and frequency-response approaches. Sliding-mode (SM) controller is another widely used control methodology for dc-dc converters and it offers several advantages such as excellent large-signal handling capability, guaranteed stability and robustness against load, line and other parametric variations and ease of implementation. The major concern when implementing the pulse-width-modulation (PWM)-based SM controller is the difficulty in achieving a good steady-state regulation using a single integral term acting on the output voltage error. Even though this problem can be alleviated via the use of a double-integral term in the sliding surface, this approach not only increases the order of the controller but may also require more state variables such as two currents for feedback purposes. Ideally, the controller should be of a lower-order to reduce the cost and for ease of implementation. To solve this problem, a fixed-frequency PWM-based SM controller for the quadratic boost converter using a reduced number of state variables is proposed. The proposed controller requires only one current for its implementation while enjoying the advantages offered by both fixed frequency and double-integral approaches. Two PWM-based SM controllers using the input and output inductor currents of the converter are separately designed to find the most suitable inductor current for the controller design. Moreover, a systematic approach to find the most suitable inductor current for designing the hysteresis-modulation (HM)-based SM controller of a hybrid high-order dc-dc boost converter is also provided. Lastly, a new adaptive current-mode controller is proposed to regulate the high step-up dc-dc converter. The proposed controller uses the estimate of the load resistance to compute the control signal. This estimate is calculated using the adaptive algorithm and the structure of this adaptive algorithm is such that it leads to an optimized derivative of the estimate which is bounded by a user specified constant. The approximate stability analysis of the controlled system is provided and some tuning guidelines are given to select the appropriate values of the controller gains. Finally, some experimental results are provided to demonstrate better regulation properties of the proposed adaptive current-mode controller compared to existing current-mode controllers.Doctor of Philosophy (EEE