Design of high power density, capacitor-based multilevel inverters

Christopher Barth

Ph.D. student with advisor R. Pilawa

Recent advances in semiconductor and microcontroller technology are opening new opportunities in power electronics design. Traditionally, inverters have employed a minimal number of active switching components and used large passive inductors to filter two or three voltage levels into sinusoidal currents. This results in a relatively simple system with acceptable performance characteristics. Newer inverter designs seek to minimize system volume and weight while maintaining or improving efficiency and performance. Because capacitor energy has been shown to be around two orders of magnitude higher than inductor energy, one method to increase power density is by using capacitors instead of inductors as the filtering components in inverter designs.


To this end, our research is investigating both modular multilevel inverters, as shown in Fig. 1, and flying capacitor inverters. These inverters utilize a higher number of active switches and microprocessor-based control than standard two or three voltage-level inverters. In exchange, the volume of required inductors as well as the overall inverter can be reduced while system control and reliability are potentially improved.


In addition to reducing inverter size, using capacitor-based multi-level inverters provides significant design benefits in electric motors with high power density. It is possible to build lightweight machines out of non-metallic materials; however removing steel results in lower machine reactance and reduces the motor’s ability to filter non-sinusoidal voltages. This necessitates using a near-sinusoidal voltage from a multilevel inverter. Working in conjunction with other members of the Power and Energy research group, research is also being done on the optimized design of combined motor and inverter systems. This research is supported by NASA grant NASA NNX14AL79A.