A data-driven optimisation approach for reusing electric vehicles batteries
Why is recycling EV batteries challenging?
Reusing electric vehicle (EV) batteries that reach the end of their useful first life is a sustainable and cost-competitive option. However, effective processes for recycling and remanufacturing EV batteries are not yet adequately developed. This is largely due to uncertainty about the quality of these batteries due to complex chemical reactions and physical conditions. To overcome these challenges, our project developed advanced data-driven models to help remanufacturers make informed and robust decisions. These theoretical models are proposed to reduce the uncertainty of battery quality and improve the accuracy of quality estimations for recycled EV batteries — ultimately increasing profits by reducing remanufacturing costs.
How does this research change practice?
Our research offers a practical method for remanufacturers to make better acquisition decisions using multiple sources of battery information. Our data-driven models will potentially help remanufacturers find ways to improve recycling and remanufacturing processes and save costs. Specifically, remanufacturers can benefit from refining valuable quality information without needing to take apart or inspect the battery cores they acquire.
The models provide insights into using limited information effectively, especially when it comes to the quality of cores from different sources. This knowledge is crucial for reducing carbon emissions from used EV batteries and avoiding landfill disposal. In addition, our research introduces a smart remanufacturing system that helps remanufacturers better track battery usage. This system promotes innovation and design improvements in battery remanufacturing. By implementing our models, remanufacturers can use advanced information processing techniques to optimise efficiency.
What does this mean for the industry?
- Improved decision-making: our model can accurately determine the quality probability intervals based only on the multisource data collected. This helps decision-makers narrow down uncertainty and prevents aimless decision-making.
- Reduced uncertainty: Our models effectively navigate uncertainty, minimising the chances of making decisions that deviate from the best outcomes and correcting biases.
- Profit maximisation: our data-driven models can use multisource quality information. This not only maximises the profit for remanufacturers but also provides valuable insights.
Monash researchers
Want to know more?
Research output
Yang, C. H., Su, X. L., Ma, X., & Talluri, S. (2024). A data-driven distributionally robust optimization approach for the core acquisition problem. European Journal of Operational Research, vol. 318(1), pp. 253-268.