Search results

In recent years, the convolutional neural network (CNN) based deep learning approach has succeeded in data-mining the relationship between microstructures and macroscopic properties of materials. However, such CNN models usually rely heavily on a large set of labeled images to ensure the accuracy and generalization ability of the predictive models. Unfortunately, in many fields, acquiring image data is expensive and inconvenient. This study aims to propose a data augmentation technique to enhance the performance of the CNN models for linking microstructural images to the macroscopic properties of composites.

Microstructures of composites are synthesized using discrete element simulations and Potts kinetic Monte Carlo simulations. Macroscopic properties such as the elastic modulus, Poisson's ratio, shear modulus, coefficient of thermal expansion, and triple-phase boundary length density are extracted on representative volume elements. The CNN model is trained using the 3D microstructural images as inputs and corresponding macroscopic properties as the labels. The comparison of the predictive performance of the CNN models with and without data augmentation treatment are compared.

The comparison between the prediction performance of CNN models with and without data augmentation showed that the former reduced the weighted mean absolute percentage error (WMAPE) for the prediction from 5.1627% to 1.7014%. This significant reduction signifies that the proposed data augmentation method can effectively enhance the generalization ability and robustness of CNN models.

This study demonstrates that data augmentation is beneficial for solving the problems of model overfitting, data scarcity, and sample imbalance for CNN-based deep learning tasks at a low cost. By developing more and advanced data augmentation techniques, deep learning accelerated homogenization will boost the multi-scale computational mechanics and materials.

This paper aims to control the deformation of a thin wall CrZrCu cylinder components (wall thickness 5 mm, diameter 400 mm) during thermal spray alumina-titania (AT13) coating by adjusting the spray parameters without deteriorating its quality evidently.

The deformation was controlled by lowering the temperature of the component in the way of adjusting the spray parameters. The main parameters adjust included extending the spraying distance, from normally 120 mm to 140 mm, decreasing plasma power from 50to 42 kW. An alumina-titanium (AT13) ceramic coating was chosen for protecting the substrate from corrosion. Microscopic morphology and phase analysis, insulation resistance testing, neutral salt test and electrochemical method were used to analyze the anti-corrosion and insulation performances of the coating.

The results indicate that, after adjusting the spraying parameters, the coating has a relatively high porosity, with an average value of 8.96 ± 0.77%. The bonding strength of the coating is relatively low, with an average value of 17.69 ± 0.85 MPa. However, after sealing, the polarization resistance of the coating in seawater can be maintained above 6.25 × 10⁶ Ω.cm² for an extended period. The coating has a high resistance (=1.1 M Ω), and there is no apparent galvanic corrosion when contacted with TC4 alloy. Additionally, analysis of corrosion products on the sample surface reveals that the samples with sprayed alumina-titanium ceramic show no copper corrosion products on the surface, and the coating remains intact, effectively isolating the corrosive medium.

By adjusting the spraying parameters, the deformation of the cylinder thin-walled component can be effectively controlled, making the φ 400 × 392 mm (thickness 5 mm) CrZrCu cylinder com-ponent with a maximum diameter deformation of only 0.14 mm. The satisfactory corrosion performances can be achieved under adjusting spraying parameters, which can guarantee the application of ceramic coating for weapon launching system of naval ships.

This study aims to investigate the microtopography transformation at a low-speed heavy-load interface with the lubrication of powder particles and its nonlinear friction effect on the sliding pair in contact.

Based on the universal mechanical tester (UMT) tribometer and VK shape-measuring laser microscope, comparative friction experiments were conducted with graphite powder lubrication. The friction coefficient with nonlinear fluctuations and the three-dimensional morphology of the boundary layer at the interface were observed and analyzed under different operating conditions. The effects on lubrication mechanisms and frictional nonlinearity at the sliding pair were focused on under different surface roughness and powder layer thickness conditions.

At a certain external load and sliding speed, the initial specimen surface with an appropriate initial roughness and powder thickness can store and bond the powder lubricant to form a boundary film readily. The relatively flat and firm boundary layer of powder at the microscopic interface can reduce the coefficient of friction and suppress its nonlinear fluctuation effectively. Therefore, proper surface roughness and powder layer thickness are beneficial to the graphite lubrication and stability maintenance of a friction pair.

This research is conducive to developing a deep understanding of the microtopography transformation with frictional nonlinearity at a low-speed heavy-load interface with graphite powder lubrication.