Additive manufacturing provides a number of benefits in terms of infinite freedom to design complex parts and reduced lead-times while globally reducing the size of supply chains as it brings all production processes under one roof. However, additive manufacturing (AM) lags far behind conventional manufacturing in terms of surface quality. This proves a hindrance for many companies considering investment in AM. The aim of this work is to investigate the effect of varying process parameters on the resultant roughness of the down-facing surfaces in selective laser melting (SLM). A systematic experimental study was carried out and the effects of the interaction of the different parameters and their effect on the surface roughness (Sa) were analyzed. It was found that the interaction and interdependency between parameters were of greatest significance to the obtainable surface roughness, though their effects vary greatly depending on the applied levels. This behavior was mainly attributed to the difference in energy absorbed by the powder. Predictive process models for optimization of process parameters for minimizing the obtained Sa in 45° and 35° down-facing surface, individually, were achieved with average error percentages of 5% and 6.3%, respectively, however further investigation is still warranted.

This article reports on the investigation of the effects of process parameters and their interactions on as-built part quality and resource-efficiency of the fused filament fabrication 3D printing process. In particular, the influence of five process parameters: infill percentage, layer thickness, printing speed, printing temperature, and surface inclination angle on dimensional accuracy, surface roughness of the built part, energy consumption, and productivity of the process was examined using Taguchi orthogonal array (L50) design of experiment. The experimental results were analyzed using ANOVA and statistical analysis, and the parameters for optimal responses were identified. Regression models were developed to predict different process responses in terms of the five process parameters experimentally examined in this study. It was found that dimensional accuracy is negatively influenced by high values of layer thickness and printing speed, since thick layers of printed material tend to spread out and high printing speeds hinder accurate deposition of the printed material. In addition, the printing temperature, which regulates the viscosity of the used material, plays a significant role and helps to minimize the dimensional error caused by thick layers and high printing speeds, whereas the surface roughness depends very much on surface inclination angle and layer thickness, which together determine the influence of the staircase effect. Energy consumption and productivity are primarily affected by printing speed and layer thickness, due to their high correlation with build time.

Fused deposition modelling (FDM), one of the most commonly used additive manufacturing techniques in the industry, involves layer-by-layer deposition of melted material to create a 3D structure. The staircase and beading effect caused by the printing process and temperature variation cause delamination and poor surface finish in FDM-printed parts. This hinders the use of these specimens in various applications, which are then usually resolved using pre-processing and post-processing techniques. Higher surface finish in pre-processing is achieved by increasing the resolution, changing layer thickness and optimizing build orientation. However, this increases the processing time considerably. On the other hand, post-processing techniques involve different processes such as mechanical, chemical, thermal and hybrid methods but can affect the mechanical and structural properties of the printed components. This review paper analyses three different aspects in the area of improving the surface finish of FDM-printed parts. First, this article reviews the state-of-the-art attempts made to improve the surface finish of FDM-printed parts concentrated mainly on different vapour polishing techniques and their respective merits and demerits. Second, it focuses on the changes in mechanical properties before and after polishing. Finally, the paper explores the development in the 3D printing of thermosets and composite materials and their post-processing processes and process parameters.

According to the Industry 4.0 vision, humans in a smart factory, should be equipped with formidable and seamless communication capabilities and integrated into a cyber-physical system (CPS) that can be utilized to monitor and recognize human activity via artificial intelligence (e.g., deep learning). Recent advances in the accuracy of deep learning have contributed significantly to solving the human activity recognition issues, but it remains necessary to develop high performance deep learning models that provide greater accuracy. In this paper, three models: long short-term memory (LSTM), convolutional neural network (CNN), and combined CNN-LSTM are proposed for classification of human activities. These models are applied to a dataset collected from 36 persons engaged in 6 classes of activities – downstairs, jogging, sitting, standing, upstairs, and walking. The proposed models are trained using TensorFlow framework with a hyper-parameter tuning method to achieve high accuracy. Experimentally, confusion matrices and receiver operating characteristic (ROC) curves are used to assess the performance of the proposed models. The results illustrate that the hybrid model CNN-LSTM provides a better performance than either LSTM or CNN in the classification of human activities. The CNN-LSTM model provides the best performance, with a testing accuracy of 97.76%, followed by the LSTM with a testing accuracy of 96.61%, while the CNN shows the least testing accuracy of 94.51%. The testing loss rates for the LSTM, CNN, and CNN-LSTM are 0.236, 0.232, and 0.167, respectively, while the precision, recall, $F1$ -Measure, and the area under the ROC curves (AUC _S ) for the CNN-LSTM are 97.75%, 97.77%, 97.76%, and 100%, respectively.

High-quality images acquired from medical devices can be utilized to aid diagnosis and detection of various diseases. However, such images can be very expensive to acquire and difficult to store, and the process of diagnosis can consume significant time. Automatic diagnosis based on artificial intelligence (AI) techniques can contribute significantly to overcoming the cost and time issues. Pre-trained deep learning models can present an effective solution to medical image classification. In this paper, we propose two such models, ResNext101_32x8d and VGG19 to classify two types of brain tumor: pituitary and glioma The proposed models are applied to a dataset consisting of 1,800 MRI images comprising in two classes of diagnoses; glioma tumor and pituitary tumor. A single-image super-resolution (SISR) technique is applied to the MRI images to clarify and enhance their basic features, enabling the proposed models to enhance particular aspects of the MRI images and assist the training process of the models. These models are implemented using PyTorch and TensorFlow frameworks with hyper-parameter tuning, and data augmentation. Experimentally, receiver operating characteristic curve (ROCC), the error matrix, Precision and Recall are used to analyze the performance of the proposed model. Results obtained demonstrate that VGG19 and ResNext101_32x8d achieved testing accuracies of 99.98% and 100%, and loss rates of 0.0120 and 0.108, respectively. The F1-score, Precision, Recall, and the area under the ROC for VGG19 were 99.89%, 99.90%, 99.89%, and 100%, respectively, while for the ResNext101_32x8d they were all 100%. The proposed models when applied to MRI images to provide a quick and accurate approach to distinguishing between patients with pituitary and glioma tumors, and could aid doctors and radiologists in the screening of patients with brain tumors.