This project focuses on estimating the Remaining Useful Life (RUL) of batteries using variational autoencoders for Industrial Process Monitoring. The method proposed in this project provides an interpretable assessment of RUL estimation, allowing for accurate predictions based on degradation patterns in sensor data.
- Utilizes variational autoencoders for RUL estimation
- Provides interpretable assessment of RUL
- Visualizes latent representations of data learned by the encoder
- Includes regression model for numerical RUL prediction
- Supports training on diverse datasets for improved generalization
The project utilizes a battery cycling dataset from the Hawaii Natural Energy Institute, tracking voltage and current over time to model degradation and estimate RUL. The dataset consists of 14 batteries cycled over 1000 times, with a nominal capacity of 2.8 Ah.
- Clone the repository from GitHub
- Install the necessary dependencies
- Run the provided scripts to train the model and evaluate RUL estimation
- Explore the latent representations and regression results
Feel free to reach out for any questions or feedback regarding this project. Thank you for your interest in Estimating Remaining Useful Life using Variational Autoencoders!