Showing 121–121 of 121 results

  • Placeholder

    Low Data AI for Energy Orchestration in Autonomy

    This work presents a set of hierarchical machine learning (ML) models for precise cycle life prediction in lithium-ion batteries (LIB), LIB packs, and system-level LIB packs. To address data limitations, synthetic data generation is employed across different scales, enhancing prediction accuracy. The presentation concludes by demonstrating the practical deployment of these ML models for accelerated degradation prediction—useful in battery cell development and manufacturing feedback—and the onboard implementation of low-data AI for energy management during operation. Discussions include key topics like battery aging, data-driven health assessment, and the model’s capacity to handle unexpected effects during use.

    This webinar will focus on the following key topics:

    • Accelerated degradation based on low data AI for battery development for targeted applications
    • Data-driven insights: machine learning for battery state of health assessment
    • Prediction of rejection thresholds during cell manufacturing for application oriented cell development
    • Prediction of targeted C-Rates for specific device applications
    • Real-world impact: practical deployment of low data ML during real time device operation

    Presenter
    Dr. Vikas Tomar – Professor at Purdue University

    Prof. Tomar’s interests lie in directed cell development using low-data AI and vertical integration of targeted cells in c-rate and energy density-specific devices. His research group has published extensively on topics related to developing data-driven models for agnostic BMS in UxVs and EVs, predicting degradation of COTS Li-ion batteries. The technology is now part of a startup, Primordis Inc., focused on launching small language models for autonomous systems within the framework of autonomous energy intelligence.

    Buy Now