Limiter Temperature Characteristics Prediction Based on VMD Decomposition and Autoencoder

Limiter is a key component for EMP protection. Accurately predicting the temperature characteristic curve of the limiter has crucial guiding significance for selecting protective components. Machine learning-based methods can effectively predict the characteristic curves of devices. However, the temperature characteristic curve of the limiter contains non-linearity and fluctuations, which traditional machine learning methods fail to predict accurately. Therefore, this paper proposes a prediction method integrating Variational Mode Decomposition (VMD) and autoencoders. It decomposes the temperature characteristics into four components and predicts each component individually, thereby accurately predicting the temperature rise characteristic curve of the limiter.