Harmonic Cancellation for Dual-Generator Power Systems in More-Electric Aircraft: A Neural Network Approach

In recent years, dual-generator power systems have garnered much attention due to the ability to enhance engine efficiency and stability margins. They make power transfer between the two generators, increasing the system reliability. In this system, both generators supply power to the DC bus simultaneously. To mitigate the effects of harmonics on the DC bus, accurate harmonic computation is essential. A data-based harmonic computation approach using artificial neural networks (ANNs) is proposed. By training an ANN with reliable data, the harmonics can be quickly predicted. Compared to traditional computation models, the ANN approach demonstrates higher accuracy. The more precise the harmonic computation, the greater the improvement will get in subsequent harmonic suppression efforts. This approach extends the lifespan of the DC bus capacitors and reduces their weight.