Abstract: Surface waviness is one of the important evaluation indicators for surface quality of micro abrasive air jet machining technology.Based on the experiment of micro abrasive air jet machining of silicon,the influences of process parameters and their interaction effects on surface waviness are studied.Surface waviness generalized regression neural network model is also developed.The results indicate that the nozzle traverse speed has the most significant effect on the surface waviness.The effects of standoff distance,working pressure of abrasive jet machine,the interaction of nozzle traverse speed and standoff distance,as well as the interaction of working pressure of abrasive jet machine and standoff distance are followed.While the interaction of working pressure of abrasive jet machine and nozzle traverse speed has insignificant effect on the surface waviness.The surface waviness increases with an increase in working pressure of abrasive jet machine,while it increases firstly and then decreases with an increase in standoff distance.It decreases with an increase in nozzle traverse speed.The combination of low working pressure of abrasive jet machine and large standoff distance,as well as the combination of high nozzle traverse speed either with medium-low or large standoff distance results in low surface waviness.The neural network model is developed according to the theory of generalized regression neural network.It is found that the model can give an adequate prediction of surface waviness after verification.

Key words: micro abrasive air jet, surface waviness, process parameter, interaction effect, generalized regression neural network