Optimization and Sustainability Assessment of the Recovery of Carbon Fibers from end-of-life CFRPs via Electrohydraulic Fragmentation

Carbon fiber recycling is a form of downcycling, that diminishes fiber quality. Research is being conducted on multiple technologies to enhance the quality and quantity of secondary carbon fibers available in the market. This thesis focuses on optimizing the recovery of carbon fibers from carbon fiber-reinforced polymers using electrohydraulic fragmentation. For this, a study of parameters was conducted using the design of experiments approach. The fiber’s quality and quantity were determined by analyzing the length and weight of the samples extracted from electrohydraulic fragmentation. The sustainability of the process was monitored by evaluating energy consumption and wastewater constituents. The selected varied parameters significantly influence the quality and quantity of the fiber output. Deionized water and smaller composite parts have a positive impact on the output, while a lightning rod negatively affects it. Clear correlations have been identified and reported. However, the distribution of fiber length and coverage of the matrix were suboptimal. The electrohydraulic fragmentation process displays relatively low energy consumption, but, on the other hand, the quality of the resulting carbon fibers is also comparatively low. To facilitate comparison of projects and parameters, key performance indicators got established to objectively evaluate also future projects. The assessment of the wastewater constituents shows elevated levels of benzene and derivatives. It can be concluded that a pyrolysis reaction occurs inside the vessel during the process of electrohydraulic fragmentation of carbon fiber reinforced polymers. A continuous process may improve fiber length distribution and decrease the presence of fiber bundles, thereby reducing energy consumption. Just one experiment was conducted for each combination of parameters, potentially diminishing accuracy. Additionally, fiber length analysis was conducted on reduced sample sizes, but sometimes the sample size was still too large to provide an accurate result.