To solve the contradiction between high quality and high efficiency of CFRP machining surface, the quadratic regression models of 3D surface roughness Sa and surface damage layer depth Dd were established by using the response surface method, and the genetic algorithm was used for multi-objective optimization to obtain small Sa, Dd and maximum material removal rate VMRR. The results show that the regression models of Sa and Dd are explicit and reliable, and the feed speed vf has the most significant influence on Sa and Dd, followed by the grinding depth ap, the spindle speed n and the ultrasonic amplitude A. The results of response surface analysis show that the interactions of n and A, vf and ap, vf and A have significant effects on Sa. The interactions of n and A, vf and A, vf and ap, ap and A have significant effects on Dd. When the weight ratios of Sa, Dd and VMRR are 1/5, 1/5 and 3/5 respectively, compared with the central point results, the optimized Sa decreases by 11.01%, Dd decreases by 10.08%, and VMRR increases by 62.02%. The absolute values of the relative errors between the experimental and the predicted values of Sa and Dd under the optimized process parameters are 8.25% and 9.41% respectively, indicating that the prediction model has high accuracy and can be used for the optimization and prediction of the process parameters of CRFP ultrasonic vibration grinding.

Low frequency vibration-assisted drilling (LFVAD) of CFRP/Ti stacks is a promising method of one-shot drilling to increase efficiency and extend tool life while adaptive approaches are applied to adjust the cutting parameters in each layer. Thus, the interfacial recognition method is significant to automatically change the cutting parameters. In this paper, two recognition methods are proposed based on the analysis of the features of cutting forces under the LFVAD process in both time and frequency domains. With the recorded thrust force signals at different wear stages, both the proposed methods identify the transition point when the drill bit starts to contact the Ti layer within allowable time delay. Compared with the traditional threshold method, the time domain method and the frequency domain method respectively increase the identifying speed by 19.8% and 46.7%, besides the reduction of implementation cost. In contrast, the time domain method reduces the programming and calculation time, while the frequency domain method improves the average recognition speed. Furthermore, an adaptive drilling system embedded with the established time-domain method is designed and the accuracy of the method is proved of 100% in a drilling test of all 20 CFRP/Ti stack holes. Moreover, the effect of the adaptive LFVAD process in improving tool wear and increasing machining efficiency is verified by reducing the force growth rate by 11.7% and time decrease of 37% in a hole-making cycle compared with the traditional LFVAD process.

Hole-making of stacks composed of carbon-fiber-reinforced plastic (CFRP) and titanium alloy under Low-frequency vibration-assisted drilling (LFVAD) is beneficial to both machining quality and tool wear. However, complicated deformation of the bottom Ti layer and even severe interfacial damage appears under the thrust forces with periodic fluctuations. Owing to lack of theoretical guidance in suppressing the deformation, this study established an analytical model to explain the influence of the fluctuating thrust force on the deformation of the workpiece. Data from validation trials performed on the CFRP/Ti stacks show that the deformation and thrust forces exhibit the same fluctuation behavior. The prediction errors of the average value and vibration amplitude of the Ti deformation were within 0.9 % and 10.7 %, respectively. Finally, a case study was conducted to optimize the drilling parameters in LFVAD of CFRP/Ti stacks based on the proposed model. As a result, the CFRP delamination damage was reduced by 17.0 %, and the stack drilling efficiency increased by 40.8 %.