20505 research outputs found
Sort by
A data-driven approach for automatic aircraft engine borescope inspection defect detection using computer vision and deep learning
Get PDFRegular aircraft engine inspections play a crucial role in aviation safety. However, traditional inspections are often performed manually, relying heavily on the judgment and experience of operators. This paper presents a data-driven deep learning framework capable of automatically detecting defects on reactor blades. Specifically, this study develops Deep Neural Network models to detect defects in borescope images using various datasets, based on Computer Vision and YOLOv8n object detection techniques. Firstly, reactor blade images are collected from public resources and then annotated and preprocessed into different groups based on Computer Vision techniques. In addition, synthetic images are generated using Deep Convolutional Generative Adversarial Networks and a manual data augmentation approach by randomly pasting defects onto reactor blade images. YOLOv8n-based deep learning models are subsequently fine-tuned and trained on these dataset groups. The results indicate that the model trained on wide-shot blade images performs better overall at detecting defects on blades compared to the model trained on zoomed-in images. The comparison of multiple models’ results reveals inherent uncertainties in model performance that while some models trained on data enhanced by Computer Vision techniques may appear more reliable in some types of defect detection, the relationship between these techniques and subsequent results cannot be generalized. The impact of epochs and optimizers on the model’s performance indicates that incorporating rotated images and selecting an appropriate optimizer are key factors for effective model training. Furthermore, models trained solely on artificially generated images from collages perform poorly at detecting defects in real images. A potential solution is to train the model on both synthetic and real images. Future work will focus on improving the framework’s performance and conducting a more comprehensive uncertainty analysis by utilizing larger and more diverse datasets, supported by enhanced computational power.Journal of Experimental and Theoretical Analyse
Anisotropy visualisation from X-ray diffraction of biological apatite in mixed phase calcified tissue samples
Get PDFX-ray diffraction is widely used to characterise the mineral component of calcified tissue. Broadening of the diffraction peaks yields valuable information on the size of coherently diffracting domains, sometimes loosely described as crystallite size or crystallinity. These domains are markedly anisotropic, hence a single number describing their size is misleading. We present a novel variation on a method for visualising crystallographic anisotropy in X-ray diffraction data. This provides an intuitively interpretable depiction of crystalline domain size and anisotropy. The new method involves creating a polar plot of calculated domain thickness for peaks in a diffractogram versus crystallographic direction. Points with the least error are emphasised. Anisotropic domain dimensions are calculated by refining an ellipsoidal model in a whole pattern fit. These dimensions are then used to overlay an ellipse on the peak broadening plot. This is illustrated by application of the method to calcifications in breast tissue with suspected cancer, which frequently contain whitlockite as well as nanocrystalline apatite. Like most biogenic apatite, this exhibits markedly anisotropic peak broadening. The nature of this anisotropy offers potentially useful information on normal function and pathology of calcified tissue and is a frequently neglected crystallographic feature of these materials.This work was supported by a Medical Research Council research grant MR/T000406/1 that funded the conduct of the research and preparation of the article. We thank Diamond Light Source for access to beamline I18 under proposal number sp30215-1.Scientific Report
Characterising and modelling plasma transferred arc for additive manufacturing
Get PDFThe thermal characteristics of a plasma transferred arc (PTA) and its mathematical representation are primary considerations when designing and modelling PTA-based wire arc additive manufacturing (WAAM). However, most of the currently used PTA thermal characteristics are derived from welding processes, which are not directly applicable to WAAM. In this study, the power density distribution, arc diameter and arc efficiency of PTA in the WAAM process were measured using the split anode calorimetry (SAC) method. The effects of key process parameters, including current intensity, plasma gas composition, plasma gas flow rate, and arc length, on the PTA power profile were systematically examined. The results show that for a typical PTA used in WAAM, the arc diameter ranged from 9.6 mm to 10.8 mm, with an arc efficiency of approximately 60 % within the tested parameter range. The PTA power becomes more concentrated as power density increases with higher current intensity and plasma gas flow rates. Additionally, a softer plasma was achieved by increasing helium content in the plasma gas or by using a longer nozzle-to-workpiece standoff distance, both of which are beneficial for avoiding keyhole defects. To accurately represent PTA power distribution, a binomial Gaussian heat source model was proposed, which captures the details of the arc power profile with a high accuracy of over 99.94 %, outperforming the conventional monomial Gaussian heat source model. The PTA calorimetry characterisation and the proposed binomial Gaussian model can be useful in establishing a better understanding of the PTA power profile and enhancing process control for high-precision WAAM.Engineering and Physical Sciences Research Council (EPSRC)The authors would like to thank New Wire Additive Manufacturing (NEWAM), United Kingdom (EP/R027218/1) programme for financial support.China Scholarship CouncilGuangyu Chen would like to express his gratitude to Cranfield University, United Kingdom and China Scholarship Council, China (No. 201706680064) for funding his research studies.International Journal of Heat and Mass Transfe
Aerodynamic design of a high-speed intake test rig for propulsion integration research
Get PDFInlet flow distortion is a critical factor in supersonic air induction systems, primarily due to its complex spatial variations and significant temporal unsteadiness. These characteristics can have a substantial impact on the performance and reliability of propulsion systems. Conducting experiments at an early stage of technology readiness can help mitigate risks associated with industrial testing and certification processes. This paper is part of a broader programme aimed at advancing experimental capabilities for high-speed propulsion system integration. The research focuses on incorporating advanced, non-intrusive measurement techniques alongside current industry-standard methods to better characterize engine compatibility. The paper details the optimization process for designing a fixed-throat supersonic nozzle that delivers high-quality flow at Mach 1.8, the structural and functional design of the working section. The integration of an intake test article is described, together with instrumentation for measuring synchronous total pressure and swirl distortion at the aerodynamic interface plane of the intake. Preliminary computational work is also presented, evaluating the flow quality within the working section and investigating the distribution of bypass flow in the radial exhaust.The authors would like to thank the Defence Science and Technology Laboratory (Dstl) for supporting this research programme.AIAA SCITECH 2025 Foru
A combined approach using structural equation modeling and analytic hierarchy process in hub airports selection
Get PDFThis study presents a method to assist airlines in selecting a hub airport using Structural Equation Modeling (SEM) combined with Analytic Hierarchy Process (AHP). This approach aims to determine the most essential elements in this selection process using observed variables collected from 300 major airports worldwide. Research on key aspects for hub airport selection is still scarce in the literature. Furthermore, most studies rely on multicriteria analyses with weights obtained through expert interviews, potentially introducing subjectivity. Therefore, this paper presents a unique approach: criteria and sub-criteria values in the AHP are determined by the results achieved in the SEM, providing a reliable scale of priorities among variables in the process of choosing a hub. The SEM tested two hypotheses that were supported by comparing two latent variables related to airport and region characteristics with another latent variable referring to aspects present in the activity of the dominant airline within each facility. Subsequently, an AHP was implemented, with criteria and sub-criteria weights based on the standardized loading factors and regression coefficients from the SEM, using sets of airports for each established world region in the study. Results indicated that airport characteristics, particularly those related to the passenger terminals, have a greater influence on the main carrier activity than region characteristics. A regional analysis revealed that in the Americas, Europe, the Middle East, and Africa, this predominance is even more pronounced, whereas in Asia-Pacific region, there is a contrary trend in which the socioeconomic factors of the city appear to be more important than airport infrastructure. The weights assigned in the AHP, based on SEM values, confirmed cohesion between the two stages of the proposed model.Journal of Air Transport Managemen
Role of acid hydrocarbon chain length on the cure kinetics and thermal degradation of epoxy- dicarboxylic acid vitrimers
Get PDFThis study investigates the cure kinetics and thermal degradation of epoxy-dicarboxylic acid vitrimers, focusing on the effect of methylene chain length. A diffusion-controlled, modified autocatalytic kinetics model was applied, based on Differential Scanning Calorimetry (DSC) data, whilst Thermogravimetric Analysis (TGA) was used to assess degradation. Increasing the methylene chain length enhanced thermal stability, with decomposition temperatures ranging from 430 °C for the hexanedioic acid formulation to 500 °C for the tetradecanedioic acid formulation. The curing process transitioned through three distinct kinetics regimes: an initial non-catalysed phase, followed by an autocatalytic stage, and finally, a diffusion-limited phase at high crosslink density. This shift leads to a 30 %-70 % reduction in apparent activation energy during the early stages. The activation energy displays a complex behaviour, initially decreasing with longer methylene sequences before rising due to competing effects of chain flexibility and reduced reactivity. Kissinger and isoconversional analyses confirmed reliable activation energy values. Despite some discrepancies in the dodecanedioic acid formulation due to secondary reactions, the model exhibits a good approximation, with an average goodness-of-fit of 84.4 %. This analysis improves understanding of vitrimer cure kinetics and thermal behaviour, providing insights for optimising industrial applications.European Polymer Journa
Key principles for assessing and implementing remote inspection with telexistence capability
Get PDFNoel, Frederic - Associate Supervisor (Grenoble University)This thesis investigates assessing and implementing telexistence capabilities for
enhancing remote inspection and control in hazardous environments.
Telexistence, allowing virtual presence via robotic systems or avatars, is crucial
for performing high-risk tasks in fields like marine engineering.
The research integrates virtual reality, industrial robotics, and sensors to improve
human abilities in maintenance, repair, and overhaul tasks, aiming for safer and
more efficient methods in hazardous settings. A key aspect is the design and
evaluation of telexistence interfaces for remote visual inspection, focusing on
interfaces that enhance operators' sense of presence and spatial awareness,
crucial for decision-making in risky conditions.
A proposed system architecture is validated with a real-world use case and used
as a basis for two user studies. The first study compares the effectiveness of a
2D and a VR interface in remote visual inspection in machinery spaces, leading
to recommendations for enhancing telexistence in maintenance scenarios. The
second study examined the impact of interface design on operators’ telexistence
capabilities at a cognitive workload level, using modalities like 2D feeds, 3D point
clouds, augmented virtuality, and mixed modalities.
A central theme is using VR to bridge the gap between human operators and
robots. VR's immersive nature improves operators' remote-control abilities, with
a digital representation in a VR control room mirroring the robot's actions in real-
time, enhancing remote-control effectiveness and providing a more immersive
telexistence experience.
Key findings demonstrate the advantages of immersive VR interfaces over
traditional 2D interfaces in remote visual inspection tasks. VR interfaces offer a
heightened sense of presence, reduced cognitive workload, and improved
usability, leading to more efficient and accurate task execution. Additionally, VR
users experienced fewer errors and completed tasks faster. Different
visualisation modalities also showed varied impacts on cognitive load, with 3D
modality imposing the most substantial demands.PhD in Manufacturin
Distributed optimal nonlinear dynamic inversion for multi-agents consensus
Get PDFIn this paper, we propose an optimal distributed controller based on Nonlinear Dynamic Inversion (NDI) theory and apply it to solve the consensus of nonlinear multi-agent systems (MASs). Our proposed method addresses the limitations of existing Distributed Nonlinear Dynamic Inversion (DNDI) techniques, which only apply to agents with square output. We formulated an optimal control problem to minimize a quadratic cost function while satisfying a set of linear constraints derived by simplifying the enforced consensus error dynamics. By relaxing the previous limitation, we introduced a distributed optimal framework called Distributed Optimal NDI (DONDI). This framework achieves consensus and incorporates additional objectives, such as minimizing control energy. The design of Optimal DNDI inherits all the advantages of NDI and provides an optimized allocation of control for achieving consensus in MAS. Also, we have shown how the controller handles the communication noise. This approach represents a significant advancement in multi-agent control, and our experimental results demonstrate its satisfactory performance and effectiveness.European Journal of Contro
Morphological changes of lenticels and their role in gas exchange and sprouting physiology of potato tubers during postharvest storage
Get PDFThe application of exogenous gases has been used to suppress sprouting in stored potato tubers. However, their efficacy in extending ecodormancy largely depends on achieving optimal gas exchange between the storage atmosphere and the tuber itself. This study aimed to investigate morphological variations and spatial distribution of lenticels and apical buds and to identify their potential role in tuber respiration rate and sprouting of five potato cultivars (‘Hermes’, ‘Lady Claire’, ‘Lady Rosetta’, ‘Saturna’, and ‘VR808’) during storage. Results revealed a consistent spatial pattern wherein the apical section of potato tubers exhibited significantly higher bud counts compared to lateral and stolon regions. ‘Lady Claire’ stood out as having the highest number of apical buds among the cultivars studied. Digital image analysis showed a seven times higher number of buds surrounding the apical eye and these were generally smaller than those distributed across the skin. ‘Saturna’ displayed double the lenticel density (12 lenticels cm-2) in smaller tubers, suggesting an inverse relationship between tuber size and lenticel density. ‘Lady Claire’ and ‘Saturna’ had respiration rates of 2.75 and 1.9 mL CO2 kg-1 h-1, respectively, and were selected for additional respiration and ethylene efflux analyses. In both cultivars, distinct spatial differences were observed, with the apical section exhibiting a seven-fold increase in lenticel density compared to the lateral and stolon sections. Respiration rate increased five-fold when apical lenticels were blocked, whereas it decreased 30-fold when the apical was the only unblocked section, suggesting differential physiological activity across lenticel locations. The apical sections, with the highest lenticel density, exhibited elevated respiration rates as a stress-induced physiological response upon blockage, compared to the lateral and stolon sections. Lenticels changed their morphology during storage, erupting before bud movement, suggesting lenticel eruption could be used as a pre-symptomatic visual marker of dormancy break. This study highlights the critical role that lenticel morphology and spatial distribution may have in determining potato tuber gas exchange and refining allied storage regimes.This study was supported by PepsiCo International Ltd. We thank the Biotechnology and Biological Sciences Research Council (BB/M027295/1) for partially sponsoring the research work.Frontiers in Plant Scienc
Sustainable wet-spun cellulose-Moringa oleifera composite fibres for potential water purification
No full textThis study explores a pioneering fabrication of novel cellulose-Moringa oleifera (M. oleifera) composite fibres (CeL-MoFs) and comparable pure regenerated cellulose fibres (CeFs) using the ionic liquid 1-ethyl-3-methylimidazolium diethyl phosphate (EMIM DEP) and the simple traditional wet-spinning process. The composites, CeL-MoFs at 0.5%, 1%, 2%, and 3%, were characterised. Fourier-transform infrared (FTIR) spectroscopy and scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM-EDX) confirmed the successful integration of M. oleifera seed powder (MoP) into the cellulose matrix. The results of preliminary adsorption studies demonstrated high selectivity for copper ions (Cu2+), with no detectable selectivity towards nickel (Ni2+) or cadmium (Cd2+). Thermogravimetric analysis (TGA) and derivative thermogravimetric (DTG) analysis revealed thermal stability variations with increasing MoP content, while atomic force microscopy (AFM) showed surface roughness and fibre defects. Rheological testing validated spinnability, and tensile analysis identified CeL-MoFs (2%) as the optimal composite, balancing mechanical strength and adsorption efficiency. These novel CeL-MoF composites, fabricated using EMIM DEP, are proposed as scalable, eco-friendly materials for selective heavy metal removal. Future work will focus on adsorption kinetics, thermodynamic modelling, and scaling production for industrial water purification applications.The authors acknowledge the Petroleum Technology Development Fund (PTDF) of Nigeria for funding through a PhD Overseas scholarship programme (No. PTDF/ED/OSS/PHD/AOO/1844/2020PHD152).RSC Advance