Development of a Grey online modeling surface roughness monitoring system in end milling operations

The industry faces a trend of fabricating a product with high-variety and short life cycles since the change of consumer behaviors. The quality control becomes critical. In milling process, surface roughness is key quality characteristic measured by an off-line method. The off-line measurement is time consuming. Therefore, online surface roughness monitoring systems, which can eliminate the measurement time and enhance the productivity, have gained popularity for machining. However, to accurately construct the system, a large amount of data and time for off-line training and modeling are required, which make the system not suitable for a high-variety job shop production. It is necessary to develop a method without large data and off-line training and modeling. Therefore, a Grey online modeling surface roughness monitoring (GOMSRM) system is proposed in this study by utilizing the Grey theory GM(1,N) with bilateral best-fit method, which requires less data and no training time to perform the online modeling surface roughness monitoring system. The ability of an online modeling makes the system flexible to monitor different types of products. The model could be built during each changeover process of the milling, which can significantly eliminate the time of training in advance used by other off-line modeling methods. The results show the GOMSRM system can accurately predict the surface roughness. The comparison between the GOMSRM system and the system developed by a neural network shows the GOMSRM system has better accuracy and fewer samples for modeling. [GRAPHICS] .