I worked with Haoran Li under Minjie Chen to create a dataset of a
large amount of voltage and current data of different magnetic
components measured in the real world. With the collected data,
we experimented with combinations of existing techniques and deep
learning algorithms to predict magnetic core loss.
- Develop neural networks with PyTorch to predict magnetic core loss under arbitrary excitations
- Perform hyperparameter optimization with Optuna to find optimal neural network architecture
Python
PyTorch
PyTorch Lightning
Optuna
Weights & Biases