Refine
Year of publication
- 2020 (171) (remove)
Institute
- Fachbereich Medizintechnik und Technomathematik (57)
- IfB - Institut für Bioengineering (33)
- Fachbereich Luft- und Raumfahrttechnik (30)
- Fachbereich Energietechnik (25)
- Fachbereich Elektrotechnik und Informationstechnik (20)
- ECSM European Center for Sustainable Mobility (14)
- Fachbereich Maschinenbau und Mechatronik (11)
- Fachbereich Wirtschaftswissenschaften (11)
- INB - Institut für Nano- und Biotechnologien (11)
- Fachbereich Chemie und Biotechnologie (10)
Language
- English (171) (remove)
Document Type
- Article (101)
- Conference Proceeding (46)
- Part of a Book (16)
- Book (2)
- Conference Poster (2)
- Doctoral Thesis (2)
- Conference: Meeting Abstract (1)
- Other (1)
Keywords
- MINLP (3)
- Additive manufacturing (2)
- Adjacent buildings (2)
- Experimental validation (2)
- Historical centres (2)
- Shake table test (2)
- Stone masonry (2)
- rebound-effect (2)
- sustainability (2)
- 3D printing (1)
Is part of the Bibliography
- no (171)
With the many achievements of Machine Learning in the past years, it is likely that the sub-area of Deep Learning will continue to deliver major technological breakthroughs [1]. In order to achieve best results, it is important to know the various different Deep Learning frameworks and their respective properties. This paper provides a comparative overview of some of the most popular frameworks. First, the comparison methods and criteria are introduced and described with a focus on computer vision applications: Features and Uses are examined by evaluating papers and articles, Adoption and Popularity is determined by analyzing a data science study. Then, the frameworks TensorFlow, Keras, PyTorch and Caffe are compared based on the previously described criteria to highlight properties and differences. Advantages and disadvantages are compared, enabling researchers and developers to choose a framework according to their specific needs.
The implementation of IO-Link in the automation industry has increased over the years. Its main advantage is it offers a digital point-to-point plugand-play interface for any type of device or application. This simplifies the communication between devices and increases productivity with its different features like self-parametrization and maintenance. However, its complete potential is not always used.
The aim of this paper is to create an Arduino based framework for the development of generic IO-Link devices and increase its implementation for rapid prototyping. By generating the IO device description file (IODD) from a graphical user interface, and further customizable options for the device application, the end-user can intuitively develop generic IO-Link devices. The peculiarity of this framework relies on its simplicity and abstraction which allows to implement any sensor functionality and virtually connect any type of device to an IO-Link master. This work consists of the general overview of the framework, the technical background of its development and a proof of concept which demonstrates the workflow for its implementation.
Water distribution systems are an essential supply infrastructure for cities. Given that climatic and demographic influences will pose further challenges for these infrastructures in the future, the resilience of water supply systems, i.e. their ability to withstand and recover from disruptions, has recently become a subject of research. To assess the resilience of a WDS, different graph-theoretical approaches exist. Next to general metrics characterizing the network topology, also hydraulic and technical restrictions have to be taken into account. In this work, the resilience of an exemplary water distribution network of a major German city is assessed, and a Mixed-Integer Program is presented which allows to assess the impact of capacity adaptations on its resilience.
In collaborative research projects, both researchers and practitioners work together solving business-critical challenges. These projects often deal with ETL processes, in which humans extract information from non-machine-readable documents by hand. AI-based machine learning models can help to solve this problem.
Since machine learning approaches are not deterministic, their quality of output may decrease over time. This fact leads to an overall quality loss of the application which embeds machine learning models. Hence, the software qualities in development and production may differ.
Machine learning models are black boxes. That makes practitioners skeptical and increases the inhibition threshold for early productive use of research prototypes. Continuous monitoring of software quality in production offers an early response capability on quality loss and encourages the use of machine learning approaches. Furthermore, experts have to ensure that they integrate possible new inputs into the model training as quickly as possible.
In this paper, we introduce an architecture pattern with a reference implementation that extends the concept of Metrics Driven Research Collaboration with an automated software quality monitoring in productive use and a possibility to auto-generate new test data coming from processed documents in production.
Through automated monitoring of the software quality and auto-generated test data, this approach ensures that the software quality meets and keeps requested thresholds in productive use, even during further continuous deployment and changing input data.
We present an automated pipeline for the generation of synthetic datasets for six-dimension (6D) object pose estimation. Therefore, a completely automated generation process based on predefined settings is developed, which enables the user to create large datasets with a minimum of interaction and which is feasible for applications with a high object variance. The pipeline is based on the Unreal 4 (UE4) game engine and provides a high variation for domain randomization, such as object appearance, ambient lighting, camera-object transformation and distractor density. In addition to the object pose and bounding box, the metadata includes all randomization parameters, which enables further studies on randomization parameter tuning. The developed workflow is adaptable to other 3D objects and UE4 environments. An exemplary dataset is provided including five objects of the Yale-CMU-Berkeley (YCB) object set. The datasets consist of 6 million subsegments using 97 rendering locations in 12 different UE4 environments. Each dataset subsegment includes one RGB image, one depth image and one class segmentation image at pixel-level.
Bacterial cellulose (BC) is a promising material for biomedical applications due to its unique properties such as high mechanical strength and biocompatibility. This article describes the microbiological synthesis, modification, and characterization of the obtained BC-nanocomposites originating from symbiotic consortium Medusomyces gisevii. Two BC-modifications have been obtained: BC-Ag and BC-calcium phosphate (BC-Ca3(PO4)2). Structure and physicochemical properties of the BC and its modifications were investigated by scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDX), atomic force microscopy (AFM), and infrared Fourier spectroscopy as well as by measurements of mechanical and water holding/absorbing capacities. Topographic analysis of the surface revealed multicomponent thick fibrils (150–160 nm in diameter and about 15 µm in length) constituted by 50–60 nm nanofibrils weaved into a left-hand helix. Distinctive features of Ca-phosphate-modified BC samples were (a) the presence of 500–700 nm entanglements and (b) inclusions of Ca3(PO4)2 crystals. The samples impregnated with Ag nanoparticles exhibited numerous roundish inclusions, about 110 nm in diameter. The boundaries between the organic and inorganic phases were very distinct in both cases. The Ag-modified samples also showed a prominent waving pattern in the packing of nanofibrils. The obtained BC gel films possessed water-holding capacity of about 62.35 g/g. However, the dried (to a constant mass) BC-films later exhibited a low water absorption capacity (3.82 g/g). It was found that decellularized BC samples had 2.4 times larger Young’s modulus and 2.2 times greater tensile strength as compared to dehydrated native BC films. We presume that this was caused by molecular compaction of the BC structure.
The production of dispatchable renewable energy will be one of the most important key factors of the future energy supply. Concentrated solar power (CSP) plants operated with molten salt as heat transfer and storage media are one opportunity to meet this challenge. Due to the high concentration factor of the solar tower technology the maximum process temperature can be further increased which ultimately decreases the levelized costs of electricity of the technology (LCOE). The development of an improved tubular molten salt receiver for the next generation of molten salt solar tower plants is the aim of this work. The receiver is designed for a receiver outlet temperature up to 600 °C. Together with a complete molten salt system, the receiver will be integrated into the Multi-Focus-Tower (MFT) in Jülich (Germany). The paper describes the basic engineering of the receiver, the molten salt tower system and a laboratory corrosion setup.
Biocomposite Materials Based on Carbonized Rice Husk in Biomedicine and Environmental Applications
(2020)
This chapter describes the prospects for biomedical and environmental engineering applications of heterogeneous materials based on nanostructured carbonized rice husk. Efforts in engineering enzymology are focused on the following directions: development and optimization of immobilization methods leading to novel biotechnological and biomedical applications; construction of biocomposite materials based on individual enzymes, multi-enzyme complexes and whole cells, targeted on realization of specific industrial processes. Molecular biological and biochemical studies on cell adhesion focus predominantly on identification, isolation and structural analysis of attachment-responsible biological molecules and their genetic determinants. The chapter provides a short overview of applications of the biocomposite materials based of nanostructured carbonized adsorbents. It emphasizes that further studies and better understanding of the interactions between CNS and microbial cells are necessary. The future use of living cells as biocatalysts, especially in the environmental field, needs more systematic investigations of the microbial adsorption phenomenon.
Background
Osteoporosis is associated with the risk of fractures near the hip. Age and comorbidities increase the perioperative risk. Due to the ageing population, fracture of the proximal femur also proves to be a socio-economic problem. Preventive surgical measures have hardly been used so far.
Methods
10 pairs of human femora from fresh cadavers were divided into control and low-volume femoroplasty groups and subjected to a Hayes fall-loading fracture test. The results of the respective localization and classification of the fracture site, the Singh index determined by computed tomography (CT) examination and the parameters in terms of fracture force, work to fracture and stiffness were evaluated statistically and with the finite element method. In addition, a finite element parametric study with different position angles and variants of the tubular geometry of the femoroplasty was performed.
Findings
Compared to the control group, the work to fracture could be increased by 33.2%. The fracture force increased by 19.9%. The used technique and instrumentation proved to be standardized and reproducible with an average poly(methyl methacrylate) volume of 10.5 ml. The parametric study showed the best results for the selected angle and geometry.
Interpretation
The cadaver studies demonstrated the biomechanical efficacy of the low-volume tubular femoroplasty. The numerical calculations confirmed the optimal choice of positioning as well as the inner and outer diameter of the tube in this setting. The standardized minimally invasive technique with the instruments developed for it could be used in further comparative studies to confirm the measured biomechanical results.
Purpose
This study aims to investigate the biomechanics of handcycling during a continuous load trial (CLT) to assess the mechanisms underlying fatigue in upper body exercise.
Methods
Twelve able-bodied triathletes performed a 30-min CLT at a power output corresponding to lactate threshold in a racing recumbent handcycle mounted on a stationary ergometer. During the CLT, ratings of perceived exertion (RPE), tangential crank kinetics, 3D joint kinematics, and muscular activity of ten muscles of the upper extremity and trunk were examined using motion capturing and surface electromyography.
Results
During the CLT, spontaneously chosen cadence and RPE increased, whereas crank torque decreased. Rotational work was higher during the pull phase. Peripheral RPE was higher compared to central RPE. Joint range of motion decreased for elbow-flexion and radial-duction. Integrated EMG (iEMG) increased in the forearm flexors, forearm extensors, and M. deltoideus (Pars spinalis). An earlier onset of activation was found for M. deltoideus (Pars clavicularis), M. pectoralis major, M. rectus abdominis, M. biceps brachii, and the forearm flexors.
Conclusion
Fatigue-related alterations seem to apply analogously in handcycling and cycling. The most distal muscles are responsible for force transmission on the cranks and might thus suffer most from neuromuscular fatigue. The findings indicate that peripheral fatigue (at similar lactate values) is higher in handcycling compared to leg cycling, at least for inexperienced participants. An increase in cadence might delay peripheral fatigue by a reduced vascular occlusion. We assume that the gap between peripheral and central fatigue can be reduced by sport-specific endurance training.