AI-Driven Digital Twins for Battery Life Cycle Assessment

Aim

Workshop Structure

📅 Day 1: Data Preparation & Setup

• Introduction to Data Preprocessing: Begin by cleaning and preparing real-world battery cycling data, focusing on time-series voltage and current data.
• Mapping to Dynamic LCA Parameters: Hands-on work mapping the processed data to dynamic Life Cycle Assessment (LCA) parameters, establishing a foundation for predictive analysis.
• Environment Setup: Participants will set up their Python environment and tools to streamline data manipulation and model building.

📅 Day 2: AI Model Training & Evaluation

• Building Predictive Models: Dive into training AI models using XGBoost and Random Forest algorithms to forecast battery degradation and predict remaining useful life (RUL).
• Model Performance Evaluation: Hands-on exercises in assessing model accuracy, tuning hyperparameters, and evaluating predictive performance. Participants will interpret the model outcomes to assess the environmental impact (e.g., carbon footprint).
• Practical Application: Use the trained models to forecast battery RUL and carbon impact, providing real-world insights into sustainability metrics.

📅 Day 3: Interactive Dashboard Deployment & LCA Visualization

• Dashboard Development: Participants will deploy an interactive dashboard using Streamlit, integrating the predictive models to visualize battery degradation and the LCA impact trajectories.
• Dynamic Visualization: Hands-on experience in building dynamic visualizations where users can adjust battery usage parameters in real-time, and instantly generate downloadable impact trajectory charts.
• Research Application: Attendees will finalize their projects by generating actionable LCA impact reports and preparing them for academic research or grant proposal submissions.

Tools and Technologies Covered: Python, XGBoost, Streamlit, Plotly.

Who Should Enrol?

Important Dates

Workshop Outcomes