From its introduction to the present day, Cascaded H-Bridge multilevel converters were employed on numerous applications. However, their floating capacitor, while advantageous for some applications (such as photovoltaic) requires the usage of balancing methods by design. Over the years, several such methods were proposed and polished. Some of these methods use optimization techniques or inject a zero-sequence voltage to take advantage of the converter redundancies. This paper describes an optimization-based capacitor balancing method with additional features. It can drive each module DC-Link to a different voltage for independent maximum power point tracking in photovoltaic applications. Moreover, the user can specify the independent active power set points to modules connected to batteries or any other energy storage systems. Finally, DC current ripple can be reduced on some modules, which can extend the lifespan of any connected ultra-capacitors. The method as a whole is tested on real hardware and compared with the state-of-the-art. In its simplest configuration, the presented method shows greater speed, robustness, and current wave quality than the state-of-the-art alternative in spite of producing about 1/3 fewer commutations. Its other characteristics provide additional functionalities and improve the adaptability of the converter to other applications.
Safe Learning and Optimization Techniques: Towards a Survey of the State of the Art
