Setup Configuration Research Articles

Evaluating the Accuracy of 3D Point Cloud Data of Elevated Structures Using Terrestrial Laser Scanners at Various Distances

Terrestrial laser scanning (TLS) is a powerful tool for generating detailed 3D models of elevated structures such as bridges, towers, and buildings. However, the quality of the resulting models heavily depends on the setup configuration of the TLS system. This research evaluates the precision of 3D point cloud data of elevated structures acquired through TLS at different distances. The data processing was performed using Cyclone Register360 software. The study aimed to evaluate the accuracy of point cloud data obtained from various TLS setup locations and compare it with the measurements obtained from a Total Station. Four different distances were used to set the TLS to scan the three elevated structure piers. The acquired data was then processed using Cyclone Register360 software to eliminate noise, visually align, and precisely register the point clouds. The results indicated that shorter distances between TLS setups resulted in more accurate point cloud data, with reduced error rates, highlighting the need to locate the scanner effectively. The study also highlighted the capabilities of Cyclone Register360 in improving the precision of point cloud data through effective data processing techniques. The findings demonstrate the significance of precise scanning distance evaluation in TLS applications to ensure high-quality data capture. It is vital for comprehensive 3D modeling and analysis of elevated structures. These valuable insights apply to specialists in surveying, engineering, and architecture. It offers guidance on the best practices for TLS setups, which can improve the accuracy and reliability of measurements. Further studies should examine the influence of other factors, such as scanning angles and environmental conditions, on the precision of TLS data.

Geant4-based technical simulation study of plastic scintillators for Positron Annihilation Lifetime Spectroscopy (PALS)

The influence of detector setup configuration and scintillator material choice on spectrum quality in Positron Annihilation Lifetime Spectroscopy (PALS) is fundamentally acknowledged primarily by empirical observation. However, this study quantifies the effects of ultra-fast plastic scintillators (BC422Q) within a conventional collinear (180-degree) detector setup used in the laboratory using the Geant4 simulation toolkit. We examine the impact of the scintillator's dimension and geometry (truncated cone vs. cylinder) on the detection efficiency for specific gamma-ray energies (1274 keV and 511 keV) and their proportion of corrupt events such as backscattering or multiple detections. Results indicate that the selection of scintillator dimension and shape significantly enhances detection efficiency albeit with an increase in corrupt events. Furthermore, we investigate the influence on the instrument response function (IRF) of the scintillators, showing how truncated cones offer superior precision and, thus, a narrower IRF compared to cylindrical shapes. Additionally, the effect of the sample material itself, in the case of a truncated cone as the scintillator shape, on the scintillator response was studied. It is observed that with an increasing atomic number of the sample the detection efficiency substantially decreases while the proportion of corrupt events also diminishes. Despite the sample material's influence on the energy of gamma-quanta interacting with the scintillator, no measurable impact on the IRF was detected for the chosen windows of the pulse height spectra. This investigation encourages a profound examination of the influences on spectrum quality in PALS measurements using Geant4 as a simulation tool highlighting the critical balance between detection efficiency and corrupt event frequencies.

PERFORMANCE COMPARISON OF THE VAEX AND DASK LIBRARIES

The purpose of the study was to compare the performance of the Vaex and Dask libraries, designed to enhance data processing efficiency. In thıs regard, experiments involving the assessment of time consumption for various classes of operations were conducted. The research included dataset preparation, data sampling, environment configuration execution, installation and setup of the aforementioned modules, Python script development, performance testing and subsequent analysis of the results obtained. It was observed that Vaex exhibits high performance when processing large datasets comprising of million objects on a single local machine; Dask's metrics performance is inferior to the former library. This fact indicates that Vaex is a more efficient tool for processing large datasets under conditions similar to those used in this study. The results and conclusions of the study emphasize the importance of choosing the optimal library when processing large volumes of data, and also confirm the advantages of the Vaex library in this context.

Modelling the Nickel Electroforming Process of a Mechanical Vane Fit for Industry 4.0

Electrochemical forming is a chemical additive manufacturing process employed for the development of a variety of niche components, from micro-manufacturing to “heavy industry”. As the Industry 4.0 era unfolds, there is a need to develop models for electroforming which are based on electrochemically sound data. For modelling to be useful, physical and electrochemical parameters which could play a significant role in process optimisation and scaling up should be examined, followed by continuous cross-validation through appropriate experiments and measurements. That way a model can be a valuable aid, allowing predictability in tooling, piloting, and manufacturing in a reliable manner, while minimising the number of manufacturing trials and process waste [1].In this presentation we aim to discuss the experimental and modelling studies conducted as part of an attempt to predict the electroforming of a vane part used in aerospace applications. Mechanical vanes play an integral role in gas turbine engine design and, therefore, are of great industrial interest. Since their main function is to guide and optimise the air flow as the fluid moves through the engine, their design must abide by strict profile and thickness requirements, followed by low tolerances during manufacturing [2]. As near net shape parts, vanes are an interesting challenge for the electroforming process.In that effort, the effect of physical and electrochemical parameters on an electroformed vane was studied and a well-informed modelling tool, using COMSOL Multiphysics, was developed for modelling its industrial electroforming process. The model was subsequently validated through experiments.Following comprehensive scaling-up studies of nickel electroforming on a lab-scale rotating disk electrode (RDE), it was previously established that secondary current distribution could adequately simulate the forming process [3]. Subsequently, experiments were carried out in an industrial piloting tank utilising an industrial vane mandrel. Qualitative analysis of experimental results was carried out to determine process characteristics such as the deposition rate, the direction in which material deposition takes place on the surface, as well as to determine the process predictability. Moving forward, simulations of deposition at current densities up to ~ 5 A/dm2 were run and validated against experimentally achieved thicknesses. Attention was especially focused on the rate with which deposition evolves at the tip of the vane mandrel (“nose”) and the rate at its front and side faces (sides) to determine whether industry requirements for a thick “nose” against thinner sides could be met. The last part of the investigation included the micro-structural characterisation of the electroformed vanes.Simulations of the process at applied currents ≤5 ASD predicted thickness distribution at the sides in reasonably good agreement with the experimentally observed one. However, thickness at the “nose” area was under-predicted by almost 30%. On the other hand, simulations of the process at applied currents >5 ASD underpredicted both the thicknesses at the “nose” and those at the sides of the part. Nevertheless, it is proposed that the model can still be used for qualitative predictions of the growth of the deposit.Scanning electron microscopy was used to determine the microstructure of the electroform and the purity of nickel. Imaging suggested that pyramid-shaped nickel particles evolve during deposition. Another interesting observation revealed a periodicity in the growth mechanism which leads to “necklace”-like zones of ~ 100 μm in thickness at the “nose”. It is proposed that these periodic zones might coincide with a periodic regeneration of the active NiOHads intermediate.Lastly, a qualitative assessment of the COMSOL Multiphysics model’s ability to simulate the effect of “masks on the electroforming process was carried out. Even though the relevant modelling results were not validated through practical experiments, the simulations strongly indicated that current distribution and, consequently, thickness distribution of demanding geometries could be efficiently controlled by the use of “masks” which could potentially be an important aid in the efforts to decrease dendritic growth at mandrel leading edges, just through modifications of the process setup configuration. Specifically, a “shell-type mask” configuration studied in this work was found to be significantly effective, “guiding” the current towards the “nose” and away from the leading edges. That way the desired thickness profile of the final product could imitate the target thickness profiles much closer than in the case when no “mask” is used.

Enhancement of PV and concentered PV panels using evaporative cooling during summer and winter seasons: A case study in Egypt

Numerous efforts are made to comprehend the performance evolution of photovoltaic (PV) panels. The elevated surface temperature causes a deterioration in PV panels' performance, hindering the expected increase in harvested energy associated with integrating solar reflectors. In this context, the year-round efficacy of cooling PV panels supported by reflective mirrors is investigated to assess the effectiveness of solar reflector implementation. A conventional PV panel, a CPV (cooled PV panel), and a CCPV (concentrated cooled PV panel) were compared side-by-side to determine their relative performance and energy productivity. Evaporative cooling pads are utilized to chill the two modified PV systems by varying the coolant flow rate during the summer and winter seasons of 2022 and 2023, respectively, in the environmental conditions of the Tenth of Ramadan City, Egypt. The findings indicate that the average daily power increment ratio (PIR) for CPV ranges from 4.7 % up to 7.4 %, while CCPV ranges from 6.7 % to 12 % for different climate conditions year-round. Finally, changing the setup configuration from CPV to CCPV increased the (PIR) 1.5 times during the summer tests and 1.7 times during the winter tests.

Discriminative Power of the Flow through Cell Dissolution Tester in Predicting the In Vivo Performance of Pentoxifylline SR Product under Fed and Fasting Conditions

This study explored, for the first time the role of different designs of the Flow-Through-Cell (FTC, USP IV) dissolution Tester in predicting the in-vivo performance of Pentoxifylline (PTX) sustained-release (SR) market product, under fed & fasting conditions. Release studies of Trental® SR 400 mg (Sanofi, Egypt), were carried-out in the FTC under different conditions, including: different volumes / compositions of release media, variable FTC flow patterns as well as applying open / closed loop configuration setups. Pharmacokinetic (PK) data, obtained from literature, were converted to in-vivo fraction-absorbed [FA] using Wagner-Nelson (WN) method. A 1:1 IVIVC was investigated by comparing PTX fraction-dissolved [FD] under different FTC release designs versus calculated [FA]. Predicted PK parameters were evaluated, and compared with actual data, with estimation of prediction-error (PE%). The suggested FTC design; a closed-loop setup, with turbulent-flow pattern of the dissolution medium; provided the most acceptable PTX release according to USP labeled limits (USP 27). Also, results showed that PTX release was pronouncedly increased in a finite-volume of gradient-buffer system rather than water, which guarantee complete resemblance to GIT environment. This release design presented the most predictive IVIVC model with PTX in-vivo performance under fasting / fed states, with acceptable PE% values in terms of Cmax and AUCs. A suggested FTC design is proposed as an alternative dissolution model in the official USP-monograph for PTX SR products.Graphical

Effect of screen configuration on the neck angle, muscle activity, and simulator sickness symptoms in virtual reality.

There is a lack of information about the optimal setup of multiple screen configurations in virtual reality (VR) office work. The objective of this study was to evaluate the effects of different screen configurations on neck flexion, rotation, neck muscle activity, and simulator sickness symptoms during Virtual Reality (VR) office work. Twelve participants (7 males; 21 to 27 years old) performed copy-paste and drag-drop tasks in three different screen configurations (single screen, primary-secondary screen, and double screen) in a randomized order. Optical motion capture system, electromyography (EMG) device, and simulator sickness questionnaire (SSQ) were used to measure the users' responses. Neck rotation angles, muscle activities, and VR sickness were significantly affected by the screen configurations (p < 0.021). The primary-secondary screen showed the highest right rotation angle (median: -33.47°) and left sternocleidomastoid (SCM) muscle activities (median: 12.57% MVC). Both single (median: 22.42) and primary-secondary (median: 22.40) screen showed the highest value of SSQ. The screen configurations in VR could be an important design factor affecting the users' physical demands of the neck and VR sickness symptoms. Asymmetric neck rotations caused by the primary-secondary screen conditions should be avoided.

Convection induced by centrifugal and Coriolis buoyancy in a rotating Hele-Shaw reactor

The study of heat and mass transfer in a Hele-Shaw cell rotating around a perpendicular axis has various advanced technological applications. These include the design of microfluidic devices and continuous-flow chemical microreactors, to name a couple. In this setup configuration, the quasi-two-dimensional design allows for recording the density field using optical methods, and the rotation enables control of this field through spatially distributed inertial forces. As is known, in the limit of an infinitely thin layer, the Coriolis force vanishes within a standard mathematical model. However, experimental observations of fluid flow in a rotating Hele-Shaw cell indicate the opposite. In this paper, we show that the correct derivation of the equation of motion under the Hele-Shaw approximation leads to the appearance of a Boussinesq-type term for the Coriolis force. To study the effect of the Coriolis buoyancy, we consider the problem of fluid stability during the internal generation of a transfer component, which can be either the concentration of the dissolved substance or the temperature of the medium. The careful study of system dynamics involves linear stability analysis, weakly nonlinear analysis, and direct numerical simulation. The general properties of the disturbance spectrum are analyzed. The branching of solutions near the first bifurcation is studied using the technique of multiple time scales. A stationary convection is replaced by an oscillatory one under the action of the Coriolis force, as demonstrated by weakly nonlinear analysis. Finally, we investigate the nonlinear dynamics using direct numerical simulation.

Unraveling the phase transition and electrochemical application of MoSe2 material for energy conversion and storage devices

Layered transition metal selenides have garnered widespread attention as the potential non-noble metal-based materials to cater to diverse energy conversion and storage aspects. Therefore, we have implemented an integrated experimental and theoretical methodology to explore the application of molybdenum diselenide (MoSe2) for HER, OER, and supercapacitor avenues via employing nanostructural and surface engineering performance enhancement strategy. Herein, we have demonstrated the successful execution of a laboratory-scale electrolyzer for water splitting with a low cell voltage of 1.54 V at 10 mA cm−2 and 1.74 V at 50 mA cm−2 current rate for 50 h. This was accomplished using synthesized exfoliated conductive carbon enwrapped MoSe2-E/C material. Further, the DFT calculations revealed the origin of better electrochemical performance on MoSe2-E (103) plane. Moreover, MoSe2-E/C showcased a high capacitance of 734 F g−1 (@ 1 A g−1) and 472 F g−1 (@ 20 A g−1) in a three-electrode setup and symmetric cell configuration respectively along with capacity retention of 95.02 % at a current density of 10 A g−1 after 2000 cycles. Hence, the prepared material exhibited its pertinence for rendering distinctive energy-related challenges owing to its unique structural arrangement.

Bayesian reconstruction of 3D particle positions in high-seeding density flows

Abstract Measuring particles' three-dimensional (3D) positions using multi-camera images in fluid dynamics is critical for resolving spatiotemporally complex flows like turbulence and mixing. However, current methods are prone to errors due to camera noise, optical configuration and experimental setup limitations, and high seeding density, which compound to create fake measurements (ghost particles) and add noise and error to velocity estimations. We introduce a Bayesian Volumetric Reconstruction (BVR) method, addressing these challenges by using probability theory to estimate uncertainties in particle position predictions. Our method assumes a uniform distribution of particles within the reconstruction volume and employs a model mapping particle positions to observed camera images. We utilize variational inference with a modified loss function to determine the posterior distribution over particle positions. Key features include a penalty term to reduce ghost particles, provision of uncertainty bounds, and scalability through subsampling. In tests with synthetic data and four cameras, BVR achieved 95% accuracy with less than 3% ghost particles and an RMS error under 0.3 pixels at a density of 0.1 particles per pixel (ppp). This shows 57% to 97% less ghost particles and 1.3 to 2 times lower RMS error than standard MART and IPR reconstructions. In an experimental Poiseuille flow measurement, our method closely matched the theoretical solution. Additionally, in a complex cerebral aneurysm basilar tip geometry flow experiment, our reconstructions were dense and consistent with observed flow patterns. Our BVR method effectively reconstructs particle positions in complex 3D flows, particularly in situations with high particle image densities and camera distortions. It distinguishes itself by providing quantifiable uncertainty estimates and scaling efficiently for larger image dimensions, making it applicable across a range of fluid flow scenarios.

Angle-Resolved Fluorescence of a Dye Coupled to a Plasmonic Nanohole Array

Gold nanohole arrays are periodic metasurfaces that are gathering huge interest in biosensing applications. The bi-dimensional grating-like structure defines their plasmonic response, together with the corresponding mode of angular dispersion. These properties can be used to investigate the interaction processes with the fluorescence features of a properly chosen emitting molecule. By employing a custom gold nanohole array alongside a commercial organic dye, we conducted an accurate angle-resolved optical characterization resorting to fluorescence, reflectance, and transmittance spectra. The coupling between the plasmonic modes and the fluorescence features was then identified as a modification of the dye fluorescence signal in terms of both spectral redistribution and enhancement. By carefully analyzing the results, different measurement efficiencies can be identified, depending on the set-up configuration, to be properly engineered for sensitivity maximization in plasmon-enhanced fluorescence-based applications.

Techno-economic evaluation of transesterification processes for biodiesel production from low quality non-edible feedstocks: Process design and simulation

The global demand for fossil fuels has led to increased pollutant emissions and depleted fossil fuel resources. Biodiesel, a fossil fuel alternative, is widely produced via transesterification. This study assesses the techno-economic performances of three transesterification processes (alkaline, acid, and CaO catalytic) for biodiesel production from low-quality non-edible feedstocks. The study explores effects of elevated free fatty acids (FFAs) and oil/ethanol flow rates on these processes, focusing on their impact on yield, purity, economics and energy aspects. Aspen Plus® V10 software was used for simulations. Despite meeting international biodiesel standards, significant technical and economic variations exist among the processes. The acid catalytic process exhibits energy requirements surpassing those of alkaline and CaO catalytic processes by over 29.58%, leading to operational costs exceeding those of CaO catalysis by 13.11%. The study establishes CaO catalysis as the most feasible option due to its simplicity, adaptability, and substantial energy and cost reductions. By introducing a closed-loop blending setup configuration, the study reveals that CaO catalysis outperforms alkaline and acid catalysis, achieving 11.59% cost reduction and 13.31% energy decrease in closed-loop configurations. The overall results highlight the potential of non-edible feedstocks in biodiesel production for a more environmentally friendly and sustainable energy future.

YOLOv5 untuk Menghitung Sel Darah Merah dan Sel Darah Putih

Health practitioners use hemocytometer to manually counting the blood cells, and it is considered time-consuming, arduous, and expert-dependent. Automated methods are costly, require meticulous maintenance, can lead to misidentify abnormal cells. This research proposed an application that swiftly, precisely, and easily count red and white blood cells. YOLOv5 is used to detect red and white blood cells in digital images. The model is trained on BCCD dataset and BCCD+ALL-IDB1 using YOLOv5s configuration and 736x736 image input size, and achieve 89.9% mAP50 value for red blood cell counting using BCCD dataset. About 17.7% mean absolute percentage error (MAPE) is obtained using YOLO5x configuration with 416x416 image input size tested on BCCD dataset. The YOLOv5s configuration setup with 736x736 image input size gives 10.9% error rate against ALL-IDB1 dataset. The system is developed using Laravel and Flask, and it proficiently detects and counts red and white blood cells.

Multi-camera tracking of mechanically thrown objects for automated in-plant logistics by cognitive robots in Industry 4.0

Employing cognitive robots, capable of throwing and catching, is a strategy aimed at expediting the logistics process within Industry 4.0’s smart manufacturing plants, specifically for the transportation of small-sized manufacturing parts. Since the flight of mechanically thrown objects is inherently unpredictable, it is crucial for the catching robot to observe the initial trajectory with utmost precision and intelligently forecast the final catching position to ensure accurate real-time grasping. This study utilizes multi-camera tracking to monitor mechanically thrown objects. It involves the creation of a 3D simulation that facilitates controlled mechanical throwing of objects within the internal logistics environment of Industry 4.0. The developed simulation empowers users to define the attributes of the thrown object and capture its trajectory using a simulated pinhole camera, which can be positioned at any desired location and orientation within the in-plant logistics environment of flexible manufacturing systems. The simulation facilitated ample experimentation to be conducted for determining the optimal camera positions for accurately observing the 3D interception positions of a flying object based on its apparent size on the camera’s sensor plane. Subsequently, a variety of calibrated multi-camera setups were experimented while placing cameras at identified optimal positions. Based on the obtained results, the most effective multi-camera configuration setup is derived. Finally, a training dataset is prepared for 3000 simulated throwing experiments where the initial part of the trajectory consists of observed interception positions, through derived best multi-camera setup, and the final part consists of actual positions. The encoder–decoder Bi-LSTM deep neural network is proposed and trained on this dataset. The trained model outperformed the current state-of-the-art by accurately predicting the final 3D catching point, achieving a mean average error of 5 mm and a root-mean-square error of 7 mm in 200 real-world test experiments.

Asymmetrical multi‐level inverter by using space vector pulse width modulation technique for low and medium power applications

SummaryA DC to AC converter is developed in this paper. This proposes a single‐phase five‐level inverter is developed in such a way that one DC source is used. The structure of the inverter is modular and the capacitor has self‐voltage balancing capability, which eliminates the need for sensors or an additional control loop. This paper presents the schematic diagram of the proposed topology, principle of operation, and the concepts of sinusoidal pulse width modulation (SPWM) for single‐phase and space vector pulse width modulation (SVPWM) for three‐phase. To further prove the advantage, power loss, and parameter computations are calculated mathematically. Low‐voltage stress is achieved in the proposed topology because the blocking voltages of all the switches are constrained to the input voltage. A comparison table is presented between SPWM and SVPWM schemes in three‐phase inverter. The proposed asymmetrical five‐level configuration setup is constructed, and experiments are performed inside the laboratory. The controller of the converter has been implemented in the field‐programmable gate array (FPGA) platform. In order to confirm the viability of the suggested topology, experimental results are provided for single‐phase arrangement. By using the FPGA platform, real‐time implementation is done for a three‐phase inverter topology.

Optimizing Setup Configuration of a Collaborative Robot Arm-Based Bimanual Haptic Display for Enhanced Performance

Optimizing Setup Configuration of a Collaborative Robot Arm-Based Bimanual Haptic Display for Enhanced Performance

Modelling of Cryopumps for Space Electric Propulsion Usage

Electric space propulsion is a technology that is used in a continuously increasing number of spacecrafts. The qualification of these propulsion systems has to run in ground-based test facilities which requires long testing times and powerful pumping systems. In these usually large test facilities, high pumping speeds are achieved with cryopumps. Cryopump operation is very expensive with respect to electrical energy and cooling water consumption. Therefore, being able to optimize pump shape, cold plate material, and pump placement in a chamber is beneficial. Pump design and tuned operating strategies can reduce costs and increase intervals between regeneration. Testing different pump configuration setups in a large facility is mostly prohibitive due to high costs and long testing times. Optimization via modelling is a better choice for design and also, later, for operation. Therefore, having a numerical model and proven guidelines at hand for optimization is very helpful. This paper describes a new model developed at DLR for the optimization of cryopump layout and operation. Model results are compared with cryopump operational and warm-up data. This validation is the basis for further optimization actions like multi-layer insulation layouts and pump cold plate upgrades, and helps in understanding and mitigating the detrimental effect of water condensates on the cryopump cold plates.

Shortest Path Routing Performance Evaluation over SDN Environment

Static routing has a manual configuration setup system, and the scope of static routing in an SDN network is just for small networks. The solution to this problem rises up with the new technology defined as software-defined networking (SDN) based on shortest path first dynamic routing. SDN has the facility of a centralized controller that smooth the controls and routes computation over a data packet. The performance analysis of SDN networks that have SDN switches connected to the network based on the shortest path first protocol are simulated on Mininet. The POX controller with Mininet programming feature for creating smart topologies was chosen. In this research, the SDN network using Dijkstra’s algorithm, Bellman-Ford algorithm, extended Dijkstra’s algorithm and Floyd Warshall Algorithm were implemented. The quality factors of SDN created by using four algorithms are measured in terms of delay, jitter, latency, packet loss, transmit, received, throughput, and bandwidth based on experimental results and European Telecommunications Standards Institute (ETSI) data. The performance parameters of SDN network topology created using Dijkstra’s, bellman ford, extended Dijkstra’s, and Floyd Warshall algorithms were compared and the experimental results showed that Bellman-Ford algorithm is better in terms of performance parameters than the other three algorithms.

Hybrid electromagnetic field-flow by electrode configuration for the removal of phenol from contaminated water

In this study, we introduce an electrode configuration setup to study the effect of an electromagnetic field (EMF) and electric potential (EP) on phenol removal in a continuous adsorption process. Feed water contaminated with dissolved phenol was exposed to EMF and passed through charged activated carbon electrodes prepared from Phoenix dactylifera (palm tree) biowaste. This study investigated the effects of magnetic field exposure, electric potential, feed salinity, and electrode substrate type. The results revealed that the highest adsorption uptake capacity of 47.2 mg/g was achieved using an aluminum substrate (AS) with EMF at zero salinity. The phenol adsorption kinetics followed a pseudo-second-order adsorption rate and diffusion model, indicating that intraparticle diffusion was the primary rate-limiting mechanism. The increased phenol adsorption uptake capacity brought on by the EMF may result from the improved mass transfer and mobility of the adsorbate and, consequently, the increased effectiveness of the internal diffusion of the pollutants. Finally, a regeneration efficiency of 99 % was achieved using NaOH at a salinity of 5,000 ppm.

Just-in-time deep learning for real-time X-ray computed tomography

Real-time X-ray tomography pipelines, such as implemented by RECAST3D, compute and visualize tomographic reconstructions in milliseconds, and enable the observation of dynamic experiments in synchrotron beamlines and laboratory scanners. For extending real-time reconstruction by image processing and analysis components, Deep Neural Networks (DNNs) are a promising technology, due to their strong performance and much faster run-times compared to conventional algorithms. DNNs may prevent experiment repetition by simplifying real-time steering and optimization of the ongoing experiment. The main challenge of integrating DNNs into real-time tomography pipelines, however, is that they need to learn their task from representative data before the start of the experiment. In scientific environments, such training data may not exist, and other uncertain and variable factors, such as the set-up configuration, reconstruction parameters, or user interaction, cannot easily be anticipated beforehand, either. To overcome these problems, we developed just-in-time learning, an online DNN training strategy that takes advantage of the spatio-temporal continuity of consecutive reconstructions in the tomographic pipeline. This allows training and deploying comparatively small DNNs during the experiment. We provide software implementations, and study the feasibility and challenges of the approach by training the self-supervised Noise2Inverse denoising task with X-ray data replayed from real-world dynamic experiments.