Noninvasive Online Capacitor Monitoring Method for Three-Level Converter Based on Active Neutral-Point Current Adjustment
Bo Guan, Xiaoliang Zhen
Abstract
Capacitor monitoring is one of the effective measures to assess the health status of power converters. Although some conventional methods have been proposed, they require adding auxiliary current sensors, changing operation conditions, or deteriorating the output performance. The lack of noninvasive online capacitor monitoring methods is a great problem in an ac/dc/ac system. To address this urgent need, in this article, we propose a truly noninvasive online capacitor monitoring method specifically for three-level neutral-point-clamped converters. The word “noninvasive” means no additional sensors, no changes to operation conditions, and little influence on the output performance. It monitors capacitors by actively adjusting the neutral-point (NP) current to induce a ripple in the NP voltage. Then, it reconstructs the injected NP current based on the switching states and three-phase currents, and obtains the response ripple in the NP voltage through a recursive least square (RLS) filter. Finally, according to the relationship between the injected NP current and the response ripple in the NP voltage, it estimates the capacitance simply via an RLS observer. Both numerical simulation and experimental results are provided as well to verify the correctness and feasibility of the proposed method.