Experimental Study of Damage and Defect Detection during Drilling of CFRP Composites.

摘要

Rejection of parts at the assembly stage due to poor quality hole with drilling induced defect and damages, high rate of drill tool wear, and the formation of gaps when drilling stacks are major problems in the manufacturing of structural components from carbon fiber reinforced plastic (CFRP) composites. Interrupting the drilling process to inspect or monitor these problems during operation increases the cost of production and is a great economic challenge for industries; therefore, there is a great need for on-line monitoring method without interrupting the drilling process. This study will address these problems through various experimental investigations.;Quality of holes and drilling induced damage and defects when drilling CFRP composites laminates were experimentally studied. The influence of drilling parameters, drilling conditions, and the type of surface plies on the resulting quality of the produced hole and on various drilling induced damage and defects were investigated. Drilling induced defects and damages such as drilled hole surface roughness, hole surface morphology, fiber pullout, and delamination were studied through qualitative measurements and SEM examination. In some cases a preliminary investigation on the use of acoustic emission and vibration signal for damage and defect detection was performed.;Analytical models to predict the critical thrust force at the onset of exit ply delamination when drilling unidirectional CFRP composites was developed and proposed using an elliptical delamination zone with clamped boundary condition, and since the load applied by the drill tip is circular, the lateral uniform load is taken over a circular region rather than over an elliptical region. In addition to the analytical model, an experimental investigation was performed through a punch test on a blind hole to characterize the critical thrust force at the onset of exit ply delamination. The critical thrust force measured from the experimental investigation was compared with the results found by the newly developed analytical model. Comparison of experimentally measured values and predicted values by the new model developed in this study with predicted values of thrust force values at the onset of delamination by other models was presented. Based on this comparison, the predicted values by the new proposed model show better correlation with the experimentally measured values than values predicted by other models.;An experimental investigation on online detection and monitoring of tool wear when drilling CFRP composite laminates was conducted utilizing different signal acquisition systems and signal analysis tools. Two new approaches namely, using the signal amplitude and using output variables from recurrence quantification analysis (RQA) were proposed and studied in this investigation. Thrust force, vibration, acoustic emission, and audio microphone signals were acquired and the variation on the signals signature were studied and correlated to the progression of the drill flank wear. The amplitude of the thrust force and torque increased when flank wear increases, whereas, the amplitudes of acoustic emission and audio microphone signals decreased when flank wear increases. six out of eight of the output variables from the RQA increases with the increase of the drill flank wear, whereas, two of the output variables decrease with the increase of the drill flank wear in both cases of drilling conditions.;In addition, a novel approach to detect the presence of a gap and to estimate the amount of the gap when drilling CFRP composite stacks through various signals was proposed in this investigation. Thrust force, acoustic emission, vibration, and audio microphone signals were acquired when drilling CFRP composite stacks. By introducing preexisting gap between the two plates before drilling, an estimation of the known preexisting gap was performed from the signal profiles. Thrust force signal shows high accuracy of estimating the preexisting gap with a maximum estimation error of 4.17 % in both 0.5 and 1 mm pre-existing gap, whereas, acoustic emission and vibration signal profiles found to be very close in both cases of gaps with an estimation error of 17 % and 5.6 % for 0.5 and 1 mm preexisting gaps respectively. Audio microphone signal delivers comparable estimation accuracy with an average estimation error of 8.63 % and 6.13 % for the preexisting gap of 0.5 mm and 1 mm respectively.