Metal additive manufacturing (AM) is gaining increasing attention from academia and industry due to its unique advantages compared to the traditional manufacturing process. Parts quality inspection is playing a crucial role in the AM industry, which can be adopted for product improvement. However, the traditional inspection process has relied on manual recognition, which could suffer from low efficiency and potential bias. This study presented a convolutional neural network (CNN) approach toward robust AM quality inspection, such as good quality, crack, gas porosity, and lack of fusion. To obtain the appropriate model, experiments were performed on a series of architectures. Moreover, data augmentation was adopted to deal with data scarcity. L2 regularization (weight decay) and dropout were applied to avoid overfitting. The impact of each strategy was evaluated. The final CNN model achieved an accuracy of 92.1%, and it took 8.01 milliseconds to recognize one image. The CNN model presented here can help in automatic defect recognition in the AM industry.

Directed energy deposition (DED) has been widely used for component repair. In the repair process, the surface defects are machined to a groove or slot and then refilled. The sidewall inclination angle of the groove geometry has been recognized to have a considerable impact on the mechanical properties of repaired parts. The objective of this work was to investigate the feasibility of repairing various V-shaped defects with both experiments and modeling. At first, the repair volume was defined by scanning the defective zone. Then, the repair volume was sliced to generate the repair toolpath. After that, the DED process was used to deposit Ti6Al4V powder on the damaged plates with two different slot geometries. Mechanical properties of the repaired parts were evaluated by microstructure analysis and tensile test. Testing of the repaired parts showed excellent bonding between the deposits and base materials with the triangular slot repair. 3D finite element analysis (FEA) models based on sequentially coupled thermo-mechanical field analysis were developed to simulate the corresponding repair process. Thermal histories of the substrate on the repair sample were measured to calibrate the 3D coupled thermo-mechanical model. The temperature measurements showed very good verification with the predicted temperature results. After that, the validated model was used to predict the residual stresses and distortions in the parts. Predicted deformation and stress results can guide the evaluation of the repair quality.

In this study, thermo-mechanical simulation was conducted to predict thermal and stress behavior in Selective Laser Melting (SLM). Temperature-dependent material properties for processed material 304L stainless steel were incorporated into the model in order to capture the change from powder to fully dense solid stainless steel. Temperature and thermal stress history were tracked under conditions of different parameter sets which were designed to reduce defect formation. The thermal model predicted the temperature history for multi-track scans under different process parameters, such as laser power, effective scanning speed and hatch spacing. Subsequently, the corresponding melt-pool size, solidification rate and temperature gradients could be calculated from simulated temperature data. These three parameters from the simulation were compared with experimental melt pool size, grain structure and cell spacing data obtained from a Renishaw AM250. The experimental data were also used to determine unknown simulation parameters required by the continuum model, e.g., the optical penetration depth and thermal conductivity multiplier for the molten region. This allowed the simulation model to accurately predict melt pool size and solidification structure of SLM 304L stainless steel. Simulated stress showed that the subsequent thermal cyclic melting in successive scanned tracks resulted in alternating compressive and tensile thermal stresses. This work will provide insight for studying microstructure morphology, residual stress and deformations in the SLM process of 304L stainless steel.