Drill-Path Optimization with Time Limit and Thermal Protection
The drilling of the printed circuit board (PCB) is a critical process due to occupying a third of the total PCB production time. This study integrates the K-nearest neighbor (K-NN) algorithm and the bi-objective multi-population genetic algorithm (BMGA) for deriving a near-optimal drilling path to reduce the drilling time, to protect product from overheating, and to balance the loading of the drill axes. Firstly, the holing coordinates are parsed from a CNC drilling program, clustered with the K-NN algorithm, which divides the global holes into several areas with local holes. Next, the local holes are classified and reconnected using 1-NN algorithm and BMGA, respectively. The BMGA fitness function is adjusted to eliminate overheating and unbalancing through drill jumping to ensure product quality. Finally, the near-optimal drilling paths can be derived within the time limit by checking a stable factor. The practical values of BMGA have already been demonstrated in 20 PCB samples. By including the thermal protection and the load balancing constraints and resolving the optimal path within an acceptable time limit, i.e. 30 min, this method can reduce the drilling length by an average of 15.5% compared to the original drilling length if the number of holes to be drilled is less than 5k. For cases with the number of holes between 5k and 20k, the average length of the drilling paths can be reduced to 25.6% of the original length. Moreover, the thermal protection maintains the samples with mean yield 99.8%.
Zone-Ching Lin, You-Min Huang, Chao-Chang Arthur Chen and Liang-Kuang Chen
H. C. Yang et al., "Drill-Path Optimization with Time Limit and Thermal Protection", Advanced Materials Research, Vol. 579, pp. 153-159, 2012