A computational model which tracked the trajectories of cane and leaf particles
subjected to aerodynamic, gravitational and impact forces during the cane cleaning
process was used to design a horizontal blower-type system. A cane cleaning test
facility was constructed to test this concept, which included injecting cane at a shallow
angle (downwards) into a horizontal air stream. The horizontal chamber was tested with
untopped, uncleaned cane over a range of operating parameters and configurations such as air velocity, cane injection speed, pour rate, initial levels of leaf, and chamber depth. Results are presented which show that the chamber was able to clean the untopped cane (at 100 t/h pour rate) from an initial level of 12.5% leaf to 5% leaf, at 1 % cane loss. The performance of the horizontal chamber in one pass is shown to be similar to that achieved by the conventional harvester using its topper and two extractor fans. The cleaning performance of the horizontal chamber was less than that predicted by the computational cane cleaning model, because the model had been validated on topped cane, while the cane used in the test facility was untopped. The extra leaf material resulted in a level of leaf-billet interaction which was not captured by the model. The leaf-billet interaction in thick mats of cane significantly limits the pneumatic cane cleaning process.