Genetic algorithm optimization in discrete element simulation of electric parts separation from printed circuit board

Demand for printed circuit board (PCB) recycling is increasing. To recover minor metals from PCB, the non-destructive detachment of electric parts is important, but has only recently begun to be investigated. To determine the optimal mechanical method for achieving such non-destructive detachment, a series of combined genetic algorithm (GA) and discrete element method (DEM) simulations were conducted, in the following manner: a breakable 2D PCB model was created by DEM; an impact ball with initial conditions and physical characteristics defined by a GA gene was driven into the PCB; the achievement of non-destructive detachment of an IC chip was assessed, and the fitness of the given gene was determined; genes were promoted for a hundred generations by the GA; and finally, the optimal non-destructive detachment conditions were determined. A notable feature of the combined GA-DEM program here employed was that various types of mechanical actions (e.g., blade stripping, hammer impact, and grinding) were able to be simultaneously investigated. The four different initial simulation conditions resulted in four different results; and among these, two considered suitable for actualization were investigated: stripping away the solder balls using a water jet and deforming the PCB through mechanical impact. With the water jet method, the detached IC chips were not destroyed, but universal parts detachment has not been achieved. With the mechanical impact method, IC chip detachment, probably caused by PCB deformation, was observed by a high-speed camera, which supported the simulation results.