Novel algorithm to protect converter from impulsive overvoltages by using neural networks

Conference: PCIM Europe digital days 2020 - International Exhibition and Conference for Power Electronics, Intelligent Motion, Renewable Energy and Energy Management
07/07/2020 - 07/08/2020 at Deutschland

Proceedings: PCIM Europe digital days 2020

Pages: 6Language: englishTyp: PDF

Skorokhod, Yury; Antushev, Nikolay; Volskiy, Nikolay (Transconverter LLC, Russia)
Volskiy, Sergey (Moscow Aviation Institute, Technical University, Russia)
Gainanov, Artem; Frolov, Nikita (DataData LCC, Russia)

The paper presents the algorithm to protect converters from DC contact railway system impulsive overvoltage. The operational principle, primary concepts, mathematical expressions and simulation results of the protection unit were described. The controlling algorithm is based on the neural network concept for the contact voltage processing. The neural network concept allows to predict with high probability moments of contact system overvoltage impulses. The algorithm was successfully studied in the protection unit of the locomotive auxiliary converter of 235 kVA. The paper presents interest for overvoltage protection unit experts in the field of power engineering and power converters.