Experimental study and modeling the tensile strength of 3D‐printed aluminium polylactic acid (PLA) parts using artificial neural networks

Background: High quality 3D printed products are in high demand, resulting in an increase in the production of 3D printed parts with precise tolerances, improved surface roughness, and overall durability. The processing parameters of 3D printers have a significant impact on the quality of 3D printed parts. Three-dimensionally printed parts must be durable, especially in terms of tensile strength, and its impact on the printer's process parameters must be investigated. Methods: Tensile test specimens were printed in the Makerbot 3D printer with aluminium polylactic acid (PLA) material. The three controllable input parameters taken into consideration were layer thickness, infill density and number of shells. The three levels for each of the respective parameters were 0.1mm, 0.2mm and 0.3mm for layer thickness; 2,3 and 4 for number of shells; 20% 40% and 60% for Infill density. Tensile testing was carried out on the specimens and data was tabulated. Using these data, an artificial neural network model was created using Matlab R2021b software’s neural network toolbox (alternatively Scilab can be used). Results: A high layer thickness (0.3mm) and a 40% infill density were found to be the most effective among all other parameters. The specimen with the lowest layer thickness of 0.1mm, four shells, and a 20% infill density had the highest tensile strength. With the tensile test data, a Matlab ANN model was developed. Validation was done by comparing the values obtained from the model with the experimental data by using random layer thickness, infill density, and number of shells. Conclusions: In conclusion, higher layer thickness has lower tensile strengths. However, as the number of shells and infill density increases, the tensile strength increases. In summary an ANN model was successfully developed and validated to predict 3D printed aluminium parts.

Automated Metrology on the Verticality of Cross-Sectioned Channel Hole at V-NAND with Over 200 Layers by Transmission Electron Microscope

In this paper, we present a nanoscale verticality measurement method for V-NAND with 200 or more layers of high layers using an automated transmission electron microscope, which has been developed a lot in the analysis field. Nanoscale measurements in cross-sectional images in 3D-NAND with such a high layer do not include both the top and bottom areas in one image of FOV. Therefore, it is very difficult for a person to objectively measure the etching angle or verticality of the channel hole. We experimented the verticality measurement of a channel hole in the two images in different areas using an automated transmission electron microscope imaging and measurement. In this paper, we present the results and analysis of the experiment and detailed metrology methods.

A Novel Sleep Staging Network Based on Data Adaptation and Multimodal Fusion

Sleep staging is one of the important methods to diagnosis and treatment of sleep diseases. However, it is laborious and time-consuming, therefore, computer assisted sleep staging is necessary. Most of the existing sleep staging researches using hand-engineered features rely on prior knowledges of sleep analysis, and usually single channel electroencephalogram (EEG) is used for sleep staging task. Prior knowledge is not always available, and single channel EEG signal cannot fully represent the patient’s sleeping physiological states. To tackle the above two problems, we propose an automatic sleep staging network model based on data adaptation and multimodal feature fusion using EEG and electrooculogram (EOG) signals. 3D-CNN is used to extract the time-frequency features of EEG at different time scales, and LSTM is used to learn the frequency evolution of EOG. The nonlinear relationship between the High-layer features of EEG and EOG is fitted by deep probabilistic network. Experiments on SLEEP-EDF and a private dataset show that the proposed model achieves state-of-the-art performance. Moreover, the prediction result is in accordance with that from the expert diagnosis.

Patch-Wise Semantic Segmentation for Hyperspectral Images via a Cubic Capsule Network with EMAP Features

In order to overcome the disadvantages of convolution neural network (CNN) in the current hyperspectral image (HSI) classification/segmentation methods, such as the inability to recognize the rotation of spatial objects, the difficulty to capture the fine spatial features and the problem that principal component analysis (PCA) ignores some important information when it retains few components, in this paper, an HSI segmentation model based on extended multi-morphological attribute profile (EMAP) features and cubic capsule network (EMAP–Cubic-Caps) was proposed. EMAP features can effectively extract various attributes profile features of entities in HSI, and the cubic capsule neural network can effectively capture complex spatial features with more details. Firstly, EMAP algorithm is introduced to extract the morphological attribute profile features of the principal components extracted by PCA, and the EMAP feature map is used as the input of the network. Then, the spectral and spatial low-layer information of the HSI is extracted by a cubic convolution network, and the high-layer information of HSI is extracted by the capsule module, which consists of an initial capsule layer and a digital capsule layer. Through the experimental comparison on three well-known HSI datasets, the superiority of the proposed algorithm in semantic segmentation is validated.

Applying self-supervised learning for semantic cloud segmentation of all-sky images

Semantic segmentation of ground-based all-sky images (ASIs) can provide high-resolution cloud coverage information of distinct cloud types, applicable for meteorology, climatology and solar energy-related applications. Since the shape and appearance of clouds is variable and there is high similarity between cloud types, a clear classification is difficult. Therefore, most state-of-the-art methods focus on the distinction between cloudy- and cloudfree-pixels, without taking into account the cloud type. On the other hand, cloud classification is typically determined separately on image-level, neglecting the cloud's position and only considering the prevailing cloud type. Deep neural networks have proven to be very effective and robust for segmentation tasks, however they require large training datasets to learn complex visual features. In this work, we present a self-supervised learning approach to exploit much more data than in purely supervised training and thus increase the model's performance. In the first step, we use about 300,000 ASIs in two different pretext tasks for pretraining. One of them pursues an image reconstruction approach. The other one is based on the DeepCluster model, an iterative procedure of clustering and classifying the neural network output. In the second step, our model is fine-tuned on a small labeled dataset of 770 ASIs, of which 616 are used for training and 154 for validation. For each of them, a ground truth mask was created that classifies each pixel into clear sky, low-layer, mid-layer or high-layer cloud. To analyze the effectiveness of self-supervised pretraining, we compare our approach to randomly initialized and pretrained ImageNet weights, using the same training and validation sets. Achieving 85.8 % pixel-accuracy on average, our best self-supervised model outperforms the conventional approaches of random (78.3 %) and pretrained ImageNet initialization (82.1 %). The benefits become even more evident when regarding precision, recall and intersection over union (IoU) on the respective cloud classes, where the improvement is between 5 and 20 % points. Furthermore, we compare the performance of our best model on binary segmentation with a clear-sky library (CSL) from the literature. Our model outperforms the CSL by over 7 % points, reaching a pixel-accuracy of 95 %.

A Comparative Analysis of Laser Additive Manufacturing of High Layer Thickness Pure Ti and Inconel 718 Alloy Materials Using Finite Element Method

Investigation of the selective laser melting (SLM) process, using finite element method, to understand the influences of laser power and scanning speed on the heat flow and melt-pool dimensions is a challenging task. Most of the existing studies are focused on the study of thin layer thickness and comparative study of same materials under different manufacturing conditions. The present work is focused on comparative analysis of thermal cycles and complex melt-pool behavior of a high layer thickness multi-layer laser additive manufacturing (LAM) of pure Titanium (Ti) and Inconel 718. A transient 3D finite-element model is developed to perform a quantitative comparative study on two materials to examine the temperature distribution and disparities in melt-pool behaviours under similar processing conditions. It is observed that the layers are properly melted and sintered for the considered process parameters. The temperature and melt-pool increases as laser power move in the same layer and when new layers are added. The same is observed when the laser power increases, and opposite is observed for increasing scanning speed while keeping other parameters constant. It is also found that Inconel 718 alloy has a higher maximum temperature than Ti material for the same process parameter and hence higher melt-pool dimensions.

Estudo QM/MM da proteína tirosina Bcr-Abl mutada T315I

Bcr-Abl tyrosine kinase protein activates the substrate starting kinase signaling cascade that results in cell division. When in excess this activation can lead to chronic myeloid leukemia. Currently, the treatment of this disease is achieved through drugs developed by rational drug design, e.g., imatinib. In this work we study descriptors of the following molecules: imatinib, nilotinib and ponatinib, together with a mutated protein. To characterize interactions between the residues of the active site against those inhibitors, a QM/MM study was carried out, using the hybrid method ONIOM, through the AMBER Force field (low layer) and the semi-empirical method PM6 (high layer).