Reducing Computation Effort By Parallel Optimization For Modular Multilevel Converters

The efficiency of a converter is important for automotive applications and influences a car's range. Modular Multilevel Series Parallel Converters offer increased efficiency compared to conventional converters. However, one drawback is the need of a complex and time-consuming control structure because online optimization methods are needed. This paper presents a control method which is based on dynamically updated look-up tables. Using parallel optimization enables the calculation of the optimal switching states asynchronously to the switching frequency. Thereby, the maximum switching frequency is increased. The simulative results clarify that active balancing and efficiency optimization shows a performance similar to online optimization methods. This finding should be considered when the control platform is chosen since a smaller and cheaper logic and processing unit can be used.