Search Results (13)

The present research is related to the application of artificial intelligence (AI) approaches for classifying surface textures, specifically regular reliefs patterns formed by ball burnishing operations. A two-stage methodology is employed, starting with the creation of regular reliefs (RRs) on test parts by ball burnishing, followed by 3D topography scanning with Alicona device and data preprocessing with Gwyddion, and Blender software, where the acquired 3D topographies are converted into a set of 2D images, using various virtual camera movements and lighting to simulate the symmetrical fluctuations around the tool-path of the real camera. Four pre-trained convolutional neural networks (DenseNet121, EfficientNetB0, MobileNetV2, and VGG16) are used as a base for transfer learning and tested for their generalization performance on different combinations of synthetic and real image datasets. The models were evaluated by using confusion matrices and four additional metrics. The results show that the pretrained VGG16 model generalizes the best regular reliefs textures (96%), in comparison with the other models, if it is subjected to transfer learning via feature extraction, using mixed dataset, which consist of 34,037 images in following proportions: non-textured synthetic (87%), textured synthetic (8%), and real captured (5%) images of such a regular relief. Full article

The development of robust and scalable tablet manufacturing methods remains a key objective in pharmaceutical technology, especially when dealing with active pharmaceutical ingredients (APIs) and excipients that exhibit suboptimal processing properties. This study evaluated two alternative manufacturing strategies for tablets containing sodium naproxen (20%, API), dolomite (65%, sustainable mineral filler), cellulose (7%), polyvinylpyrrolidone (5%, binder), and magnesium stearate (3%, lubricant). The direct compression method used a vibrating ball mill (SPEX SamplePrep 8000M), while the indirect method employed wet granulation using a pan granulator at different inclination angles. Physical properties of raw materials and granules were assessed, and final tablets were evaluated for mass, thickness, mechanical resistance, abrasiveness, and API content uniformity. Direct compression using vibratory mixing for 5–10 min (DT2, DT3) resulted in average tablet masses close to the target (0.260 g) and improved reproducibility compared to a reference V-type blender. Wet granulation produced tablets with the lowest abrasiveness (<1.0%) and minimal variability in dimensions and API content. The best uniformity (SD < 0.5%) was observed in batch IT2. Overall, vibratory mixing proved capable of achieving tablet quality comparable to that of wet granulation, while requiring fewer processing steps. This highlights its potential as an efficient and scalable alternative in solid dosage manufacturing. Full article

Background/Objectives: This ex vivo study aimed to evaluate the diagnostic performance of ultra-low-dose (ULD) cone-beam computed tomography (CBCT) protocols in detecting vestibular bone defects for immediate implant planning, using intraoral scan (IOS) data as a reference. Methods: Four CBCT protocols (ENDO, A, B, C) were applied to four dried human skulls using a standardized setup and a single CBCT unit (Planmeca ProMax^® 3D Mid, Planmeca Oy, Helsinki, Finland). All scans were taken at 90 kV, with varying parameters: (1) ENDO (40 × 50 mm, 75 µm, 12 mA, 80–120 µSv, 15 s), (2) A (50 × 50 mm, 75 µm, 9 mA, 20–40 µSv, 5 s), (3) B (100 × 60 mm, 150 µm, 7.1 mA, 22–32 µSv, 5 s), and (4) C (100 × 100 mm, 200 µm, 7.1 mA, 44 µSv, 4 s). Vestibular root surfaces of single-rooted teeth (FDI regions 15–25 and 35–45) were digitized via IOS and exported as STL files. CBCT datasets were superimposed using 3D software (Blender 2.79), and surface defects were measured and compared using one-sample t-tests and Bland–Altman analysis. The level of significance was set at p < 0.05. Results: A total of 330 vestibular surfaces from 66 teeth were analyzed. Compared to the IOS reference, protocols ENDO and A showed minimal differences (p > 0.05). In contrast, protocols B and C exhibited statistically significant deviations (p < 0.05). Protocol B demonstrated a mean difference of −0.477 mm² with limits of agreement (LoA) from −2.04 to 1.09 mm² and significant intra-rater variability (p < 0.05). Protocol C revealed a similar mean deviation (−0.455 mm²) but a wider LoA (−2.72 to 1.81 mm²), indicating greater measurement variability. Overall, larger voxel sizes were associated with increased random error, although deviations remained within clinically acceptable limits. Conclusions: Despite statistical significance, deviations for protocols B and C remained within clinically acceptable limits. ULD CBCT protocols are, thus, suitable for evaluating vestibular bone defects with reduced radiation exposure. Full article

The need for scalable, immersive training systems is universal and recently has been included in fields that rely on complex, hands-on processes, such as surgery operations, assembly operations, construction processes training, etc. This paper examines the potential to support immersive training via digital tool manipulation in the domain of traditional handicrafts. The proposed methodology employs Finite Element Method simulations to compute material transformations and apply them to interactive virtual environments. The challenge is to accurately simulate human–tool interactions, which are critical to the acquisition of manual skills. Using Simulia Abaqus (v.2023HF2), crafting simulations are authored, executed, and exported as animation sequences. These are further refined in Blender (v3.6) and integrated into Unity to create reusable training components called Action Animators. Two software applications—Craft Studio (v1.0) and Apprentice Studio (v1.0)—are designed and implemented to enable instructors to create training lessons and students to practice and get evaluated in virtual environments. The methodology has wide-ranging applications beyond crafts, offering a solution for immersive training in skill-based activities. The validation and evaluation of the proposed approach suggest that it can significantly improve training effectiveness, scalability, and accessibility across various industries. Full article

A key component of the new stochastic approach for discontinuity shear strength (referred to as StADSS) is characterising the roughness of natural rock discontinuities at full scale to mitigate well-known scale effects on shear strength predictions. An investigation was conducted using the software Blender (v4.0) to determine the influence of camera orientation and position on the estimation of the standard deviation of gradients, a parameter used to quantify roughness and to make predictions of discontinuity shear strength. The existing literature has investigated the various distortions of images due to camera position and orientation; however, a comprehensive understanding of their unique influence on error in shear strength prediction is still missing. The investigation revealed that a ground sampling distance of less than 1.36 mm/pixel allows the standard deviation of gradients to be quantified within approximately ±10% relative error to the control data set. Based on investigations into camera orientation relative to the planar control data set, error on the roughness parameter due to perspective distortion was quantified. Recommendations were made to reduce perspective distortions and improve seed trace digitisation errors, including capturing images perpendicular to the seed trace with no rotation and getting as close as possible to the trace during capture to improve ground sampling distance. Lastly, the influence of these different trace digitisation errors on predictions of shear strength obtained using StADSS was investigated. Digitisation errors were found to have a disproportionate influence on shear strength prediction error, especially for rough discontinuities at low normal stress. The investigation highlighted the importance of accurately digitising the standard deviation of gradients to predict discontinuity shear strength. Full article

Background: Head and neck reconstruction following ablative surgery results in alterations to maxillofacial anatomy and function. These postoperative changes complicate dental rehabilitation. Methods: An innovative modular, stackable guide system for immediate dental rehabilitation during mandibular reconstruction is presented. The virtual surgical planning was performed in Materialise Innovation Suite v26 and Blender 3.6 with the Blenderfordental add-on. The surgical guides and models were designed and manufactured at the point of care. Results: The duration of the surgery was 9 h and 35 min. Good implant stability (>35 Ncm) and a stable occlusion were achieved. After 9 months of follow-up, the occlusion remained stable, and a mouth opening of 25 mm was registered. The dental implants showed no signs of peri-implant bone loss. Superposition of the preoperative planning and postoperative position of the fibula parts resulted in an average difference of 0.70 mm (range: −1.9 mm; 5.4 mm). Conclusions: The in-house developed stackable guide system resulted in a predictive workflow and accurate results. The preoperative virtual surgical planning was time-consuming and required extensive CAD/CAM and surgical expertise. The addition of fully guided implant placement to this stackable guide system would be beneficial. More research with longer follow-ups is necessary to validate these results. Full article

Laser-based powder bed fusion of metals (PBF-LB/M) is the most used additive manufacturing (AM) technology for metal parts. Nevertheless, challenges persist in effectively managing metal powder, particularly in blending methodologies in the choice of blenders as well as in the verification of blend results. In this study, a bespoke laboratory-scale AM blender is developed, tailored to address these challenges, prioritizing low-impact blending to mitigate powder degradation. As a blending type, a V-shape tumbling geometry meeting the requirements for laboratory AM usage is chosen based on a literature assessment. The implementation of thermal oxidation as a powder marking technique enables particle tracing. Blending validation is achieved using light microscopy for area measurement based on binary image processing. The powder size and shape remain unaffected after marking and blending. Only a small narrowing of the particle size distribution is detected after 180 min of blending. The V-shape tumbling blender efficiently yields a completely random state in under 10 min for rotational speeds of 20, 40, and 60 rounds per minute. In conclusion, this research underscores the critical role of blender selection in AM and advocates for continued exploration to refine powder blending practices, with the aim of advancing the capabilities and competitiveness of AM technologies. Full article

Industrial-quality inspections, particularly those leveraging AI, require significant amounts of training data. In fields like injection molding, producing a multitude of defective parts for such data poses environmental and financial challenges. Synthetic training data emerge as a potential solution to address these concerns. Although the creation of realistic synthetic 2D images from 3D models of injection-molded parts involves numerous rendering parameters, the current literature on the generation and application of synthetic data in industrial-quality inspection scarcely addresses the impact of these parameters on AI efficacy. In this study, we delve into some of these key parameters, such as camera position, lighting, and computational noise, to gauge their effect on AI performance. By utilizing Blender software, we procedurally introduced the “flash” defect on a 3D model sourced from a CAD file of an injection-molded part. Subsequently, with Blender’s Cycles rendering engine, we produced datasets for each parameter variation. These datasets were then used to train a pre-trained EfficientNet-V2 for the binary classification of the “flash” defect. Our results indicate that while noise is less critical, using a range of noise levels in training can benefit model adaptability and efficiency. Variability in camera positioning and lighting conditions was found to be more significant, enhancing model performance even when real-world conditions mirror the controlled synthetic environment. These findings suggest that incorporating diverse lighting and camera dynamics is beneficial for AI applications, regardless of the consistency in real-world operational settings. Full article

Virtual remote laboratories have already been successfully implemented in educational centers for practical learning of mechatronics and robotic systems. This article presents the development of a virtual articulated UR-type robot, designed as an educational tool that is suitable for programming and evaluating both the inverse kinematics control of the robot and the independent control of the robot joints. The 3D model of the virtual robot was developed in the Blender V2.79 software and uses the Modbus TCP industrial communication protocol for the communication to an external controller implemented in CoDeSys V3.5 software. The developed system allows the students to generate and test their own control algorithm for the robot joints with the visualization of the achieved performance in 3D and real time. Tailored control systems can be compared on the virtual robot. In this study, a novel technique for the joint position control based on an FSM is proposed and verified with the virtual UR5 robots to prove that the developed system is a suitable platform to teach and learn the inverse kinematics control and independent joint control of the UR5 robotic arm. Full article

Over the last decade, the conical mill has emerged as a potential piece of equipment to use for continuous dry coating pharmaceutical powders. In this work, silicon dioxide was used as a guest particle on two excipients, fast flow lactose (FFL) and grade PH200 microcrystalline cellulose (MCC), for dry coating by a conical mill with a modified screen that permitted batch and continuous mode operation. Samples were pre-processed in a V-blender. SEM images, particle size distribution, and EDS mapping were used to characterise the treated powders. Pre-processed samples showed some discrete coating of the host particle. After batch processing, the samples were covered with a complete coating. When processed at high impeller speed, coating of FFL was a mix of A200P and FFL fines generated by attrition. Continuous mode processed samples, which had a lower processing time, were coated discretely and showed a better coating than the pre-processed samples. Increasing guest/host mass ratio with FFL host particle had a positive impact on the quality of the coating. These results help to build the case that the processing time of the conical mill is a key parameter to the success of the conical mill as dry coating equipment in the pharmaceutical industry. Full article

In this research, we investigate the influence of a load-balancing strategy and parametrization on the speed-up of discrete element method simulations using Lethe-DEM. Lethe-DEM is an open-source DEM code which uses a cell-based load-balancing strategy. We compare the computational performance of different cell-weighing strategies based on the number of particles per cell (linear and quadratic). We observe two minimums for particle to cell weights (at 3, 40 for quadratic, and 15, 50 for linear) in both linear and quadratic strategies. The first and second minimums are attributed to the suitable distribution of cell-based and particle-based functions, respectively. We use four benchmark simulations (packing, rotating drum, silo, and V blender) to investigate the computational performances of different load-balancing schemes (namely, single-step, frequent and dynamic). These benchmarks are chosen to demonstrate different scenarios that may occur in a DEM simulation. In a large-scale rotating drum simulation, which shows the systems in which particles occupy a constant region after reaching steady-state, single-step load-balancing shows the best performance. In a silo and V blender, where particles move in one direction or have a reciprocating motion, frequent and dynamic schemes are preferred. We propose an automatic load-balancing scheme (dynamic) that finds the best load-balancing steps according to the imbalance of computational load between the processes. Furthermore, we show the high computational performance of Lethe-DEM in the simulation of the packing of ${10}^8$ particles on 4800 processes. We show that simulations with optimum load-balancing need ≈40% less time compared to the simulations with no load-balancing. Full article

Red Delicious apple pomace was produced at laboratory scale with a domestic blender and different non-conventional extraction techniques were performed to isolate phenolic compounds, such as ultrasound-assisted extraction (UAE), ultraturrax extraction (UTE), accelerated solvent extraction (ASE) and pulsed electric field (PEF) extraction pre-treatment. Total phenolic content (TPC) was determined by Folin–Ciocalteu assay. Phloridzin, the main phenolic compound in apples, was determined by chromatographic analysis Q-TOF-LC/MS. The results obtained with these techniques were compared in order to identify the most efficient method to recover polyphenols. The highest value of TPC (1062.92 ± 59.80 µg GAE/g fresh apple pomace) was obtained when UAE was performed with EtOH:H₂O (50:50, v/v), while ASE with EtOH:H₂O (30:70, v/v) at 40 °C and 50% of flush was the most efficient technique in the recovery of phloridzin. The concentration of the main phenolic compounds ranged from 385.84 to 650.56 µg/g fresh apple pomace. The obtained results confirm that apple pomace represents an interesti-ng by-product, due to the presence of phenolic compounds. In particular, phloridzin could be considered a biomarker to determine the quality of numerous apple products. Therefore, this research could be a good starting point to develop a value-added product such as a functional food or nutraceutical. Full article

An approach combining quality by design (QbD) and the discrete element method (DEM) is proposed to establish an effective scale-up strategy for the blending process of an amlodipine formulation prepared by the direct compression method. Critical process parameters (CPPs) for intermediate critical quality attributes (IQAs) were identified using risk assessment (RA) in the QbD approach. A Box–Behnken design was applied to obtain the operating space for a laboratory-scale. A DEM model was developed by the input parameters for the amlodipine formulation; blending was simulated on a laboratory-scale V-blender (3 L) at optimal settings. The efficacy and reliability of the DEM model was validated through a comparison of simulation and experimental results. Change of operating space was evaluated using the validated DEM model when scaled-up to pilot-scale (10 L). Pilot-scale blending was simulated on a V-blender and double-cone blender at the optimal settings derived from the laboratory-scale operating space. Both pilot-scale simulation results suggest that blending time should be lower than the laboratory-scale optimized blending time to meet target values. These results confirm the change of operating space during the scale-up process. Therefore, this study suggests that a QbD-integrated DEM simulation can be a desirable approach for an effective scale-up strategy. Full article