VaTreBat

The dismantling and unloading of large battery systems from the automotive and industrial sectors on an industrial scale will be a key competence in an economic recycling process for such systems in the future. The aim must be to dismantle large quantities of such systems in stable and repeatable processes, taking into account occupational and process safety (incl. fire protection, exhaust air). Due to the lack of standards, battery systems show a high variance in the design of the system elements and the joining technologies used. This applies in particular to the cells used, which range from wound round cells to prismatic flat windings in hardcase housings to stacked electrode-separator composites in pouch foil. As a result, high demands are placed on the disassembly processes in terms of flexibility. Existing disassembly processes, such as the system disassembly of the project partner Redux, are therefore mostly characterised by laborious manual work, whose productivity and safety are limited by the handling of mass-intensive high-voltage components.
The project "Variant-flexible and automated separation of connections in the dismantling process of battery systems" (project keyword VaTreBat) starts at this point and develops a procedure for the automated dismantling of battery systems down to the cell level with a focus on variant flexibility, safety and productivity.
To realise an automated and variant-flexible disassembly process, the consortium first analyses the specific challenges in the disassembly of battery systems on the basis of the fully manual reference process of the project partner Redux. Based on this analysis, concrete requirements for the automation of the sub-processes "recognition and processing of information on joints", "separation of joints" and "handling of system components" are derived. Based on these requirements, the interdependencies and influencing variables of these three sub-processes are investigated experimentally in three parallel work packages.
For the localisation and classification of joints and system components, the project consortium is developing a combination of fixed and robot-guided sensors consisting of camera and laser systems. Using machine learning, the data from the sensors are evaluated in order to localise and classify joints and components to be handled in the coordinate system of the disassembly system. This provides the subsequent automated separation and handling processes with the necessary information about the joints to be separated.
For universal, contactless and minimally invasive separation, the separation process relies on the use of laser and waterjet cutting processes. By means of an influence study, the influencing variables of the various joining joints on the separation process are analysed and suitable process variants for the separation are developed.
In the investigation and development of the subsequent handling process, the various battery system components are analysed with regard to the handling properties and a variant-flexible handling system is developed on this basis. The handling system also uses the information provided by the localisation and classification to plan and implement a component-specific optimal handling strategy.
The developed solution variants of the three sub-processes are combined into a coupled disassembly concept. On this basis, a demonstrator process will be built and experimentally compared and evaluated with the existing reference process of the project partner Redux. In a final work package, the consortium will analyse the necessary interfaces of the sub-processes to each other and to upstream and downstream processes and propose a concept for scaling the developed demonstrator process to industrial scale.

Contact:

Philip Gümbel
p.guembel(at)tu-braunschweig.de

Dr. rer. nat. Maja Kandula
m.kandula(at)tu-braunschweig.de