Prediction of Material Removal and Surface Roughness in Powder Blasting using Neural Network and Response Surface Analysis

신경회로망 및 반응표면분석법을 이용한 파우더 블라스팅시의 표면거칠기 및 재료제거량 예측

Powder blasting technique has been considered one of the most appropriate micro machining methods for hard and brittle materials, since the productivity is high and the heat layers caused by material removal are very thin. Recent development of special purposed parts, such as the parts for semiconductor processing, the parts for LCD, sensors for micro machine fabrication and so on, has been expanded. Thus, it is essential to develop powder blasting technologies for micromachining of hard and brittle materials such as glass, ceramics and so on. In this paper, the characteristics of powder blasted glass surface were tested under various blasting parameters. Finally, we proposed a predictive model for powder blasting process using the neural network and the response surface method. Detail analysis of the simulation results is carried out and the performance of two predictive models is compared.