Date and Time: 15 Mar 2024 11:40-13:00
Organisers: Tomi Pitkäaho, Nestor Garcia, Enrico Villagrossi, Maria Ruottinen
Main questions to be answered: How feasible are AI and robotics for the disassembly of EV batteries ?
Additional questions: How XR technology and Digital twins enable physical hardware interaction in the training of a robotic operator?
Additional questions 2: Is more sustainable mobility achieved through efficient repurposing and recycling of EV batteries?
Workshop Description: The workshop explores AI-driven autonomous disassembly of used EV batteries to overcome safety hazards and inefficiencies inherent in manual labor. We present solutions utilizing commercial robots and novel software, with machine learning and digital twins optimizing the dismantling process dynamically. Task and motion planning are discussed as an essential part of the dynamic disassembly of battery packs. Sensor fusion technologies enhance safety and efficiency, offering preemptive hazard detection, while innovative training environments leverage XR technology to foster skill development in operators, facilitating adaptive and optimized dismantling strategies. We underscore the economic and environmental dividends of this shift, highlighting enhanced safety, reduced labor costs, and a sustainable recycling process, aiming to foster a greener future. https://net.centria.fi/en/rdi/
Intended Outcome: Workshop on recent developments in technology or applications, Workshop discussion topics of common interest, success stories, use cases, etc
Approach: Due to challenges related to the autonomous disassembly of used EV batteries, recycling is typically a work of manual labor. However, the disas- sembly phase of end-of-life or damaged EV batteries exposes human work- ers to potential hazards such as toxic chemicals, fire, electric arcs, and high voltages. Robotized disassembly is an option for manual labor; however, au- tonomous disassembly is challenging due to variations in the EV battery’s mechanical condition, physical shape, and methods utilized during the as- sembly. We introduce concepts for the robotized dismantling of challenging EV batteries. Mechanically, the solutions are based on commercially available industrial robots, while the partners of this workshop develop the software solutions. Artificial intelligence is at the heart of each presented solution, connecting seamlessly to the robot controller. Digital twins of the disman- tling cells optimize the battery dismantling trajectories in the background while cells work on physical batteries. We present machine learning algorithms to enhance the capabilities of robotic dismantling systems. Presented interactive teaching techniques, ad- vice, and real-life examples of simulating the dismantling process and train- ing the AI to dismantle new battery types and optimize the trajectories for known battery types. The robots can learn to anticipate potential issues, adjust their movements accordingly, and improve overall performance by training the AI models on a large dataset of battery disassembly scenarios. The use of advanced Task and Motion planning algorithms is mandatory to cope with the variability of battery packs and the presence of a team of robots cooperating on the disassembly tasks. Even in this case, machine learning techniques drive the research activities for releasing flexible and effective solutions that are reliable for industrial adoption. Sensor fusion is an essential part of creating efficient and safe disman- tling cells. Various sensors in the dismantling cells enable early detection of hazards by signaling the cell operator about hot or leaking battery cells inside the battery assembly being dismantled. By detecting the temperature and voltage of the battery and various gases inside the cell enables efficient detection of potential hazards. We discuss the importance of data collec- tion and analysis during the robotic dismantling process. By gathering data on the disassembly steps, robot movements, and battery conditions, we can continuously improve the efficiency and safety of the dismantling operation. This data-driven approach allows for a better understanding of the chal- lenges and potential optimizations in the disassembly process. We introduce an interactive teaching factory approach to dismantling, including digital twins of the dismantling cells, XR training environments, and autonomous robot programming systems. The AI algorithms continu- ously learn from the operators’ actions and feedback, allowing for adaptive and optimized dismantling strategies. By leveraging virtual and augmented reality, operators can train in a safe and controlled environment, increasing their knowledge and skills in dismantling different battery types. Finally, we address the economic and environmental benefits of imple- menting robotic dismantling systems in the recycling industry. By automat- ing the disassembly process, we can increase safety and significantly reduce the labor costs associated with dismantling and recovering valuable materi- als from used EV batteries more efficiently, contributing to the sustainability of the recycling process. By embracing robotic technologies, we can pave the way for a greener and more sustainable future Centria, Tero Kaarlela / Intro and Digital twin of battery dismantling cell. / 10 min. ———————————— Eurecat, Nestor Garcia / Enabling Sustainable Battery Recycling: Harnessing Human-Robot Synergy. / 10 min. ———————————– CNR-STIIMA, Enrico Villagrossi /Robotic autonomous battery pack disassembly: from planning to control. / 10 min. ———————————————— Probot OY, Maria Ruottinen / Benefits and challenges of robotized recycling process. / 10 min. ——————————————— Panel discussion 40 min Ozak Durmus: Ford Otosan, Product Sustainability Leader (Agreed) Matti Tikanmäki: Probot OY, CEO (Agreed) Alireza Rastegarpanah: The University of Birmingham, Senior Robotic Scientist (Agreed) Rana Pant: European Commission, Policy Officer (Agreed)
Contributors: Dr. Tero Kaarlela, Researcher, Centria University, Confirmed, Speaker Dr. Nestor Garcia, Researcher, Eurecat, Confirmed, Speaker Dr. Enrico Villagrossi, Researcher, CNR-STIIMA, Confirmed, Speaker Mrs. Maria Ruottinen, Specialist, Probot OY, Confirmed, Speaker Ozak Durmus: Ford Otosan, Product Sustainability Leader (Panelist, Agreed) Matti Tikanmäki: Probot OY, CEO (Panelist, Agreed) Alireza Rastegarpanah: The University of Birmingham, Senior Robotic Scientist (Panelist, Agreed) Cesar Santos: EC Commission (Panelist, Invited)
Further Information: https://net.centria.fi/en/project/recirculate/ https://doi.org/10.3390/act12060219 https://doi.org/10.3390/machines11010013 https://doi.org/10.1016/j.mechatronics.2018.01.016 https://doi.org/10.1109/TSMC.2017.2756856