Abstract:In order to solve the problem that there is prone to edge cracking of engineering ceramics with common processing, this paper studies the boundary damage prediction system of ultrasonic internal grinding. The experiment was carried out independently under the conditions of 35 kHz axial ultrasonic grinding and ordinary grinding. The support vector machine was used to study the influence of the process parameters to boundary damage. The improved particle swarm optimization algorithm was used to optimize the support vector machine and the hybrid kernel function AHPSO-SVM predict model was established . The results show that the reduction of boundary damage of the specimen under ultrasonic is 10.05%~21.23%, the MSE of the AHPSO-SVM prediction model is 0.3784, the average relative error is 1.3690%, and the standard deviation of the 30-time fitness value is 0.0202. Compared with ordinary grinding, ultrasonic grinding can significantly reduce the boundary damage value of ZTA ceramics.The established AHPSO-SVM model has better learning ability, generalization performance and good stability.