Classification of Non-ferrous Scrap Metal using Two Component Magnetic Induction Spectroscopy

Magnetic induction spectroscopy is the measurement of how a conductive object reflects and scatters a magnetic field over different frequencies in response to some excitation magnetic field. In recent work, we proposed using this technique to classify different non-ferrous metals for the recycling and resource recovery sector - specifically, to identify fragments of scrap aluminium, copper and brass in shredded waste streams for separation and recovery. We proposed a simple algorithm that used only two components of the spectra that gave strong purity and recovery-rates when tested on a manufactured control set cut from stock metals.In this paper, we re-examined this method using real scrap metal samples drawn from a commercial sorting line. We found moderate purity and recovery-rates of brass and copper of between around 70% and 90%. However, the classification of aluminium was poor with ≈55% and ≈80% purity and recovery rates respectively. Magnetic induction sensors are a natural fit for the specifications of the industry. They are capable of high-throughputs, are unaffected by dirt or contaminants and are mechanically and physically robust. Although our results are modest, they are not insignificant given the simplicity of the algorithm and the relatively low-cost of instrumentation. Our work suggests the MIS as a technique may have a significant role to play in the extraction and recovery of non-ferrous resources.