Ingenieurwissenschaften und zugeordnete Tätigkeiten
Filtern
Dokumenttyp
- Zeitschriftenartikel (3)
- Vortrag (3)
- Posterpräsentation (1)
- Forschungsdatensatz (1)
Sprache
- Englisch (8)
Schlagworte
- Software (8) (entfernen)
Organisationseinheit der BAM
Eingeladener Vortrag
- nein (3)
We present Jobflow, a domain-agnostic Python package for writing computational workflows tailored for high-throughput computing applications. With its simple decorator-based approach, functions and class methods can be transformed into compute jobs that can be stitched together into complex workflows. Jobflow fully supports dynamic workflows where the full acyclic graph of compute jobs is not known until runtime, such as compute jobs that launch
other jobs based on the results of previous steps in the workflow. The results of all Jobflow compute jobs can be easily stored in a variety of filesystem- and cloud-based databases without the data storage process being part of the underlying workflow logic itself. Jobflow has been intentionally designed to be fully independent of the choice of workflow manager used to dispatch the calculations on remote computing resources. At the time of writing, Jobflow
workflows can be executed either locally or across distributed compute environments via an adapter to the FireWorks package, and Jobflow fully supports the integration of additional workflow execution adapters in the future.
AbstractLaser beam welding has become widely applied in many industrial fields in recent years. Solidification cracks remain one of the most common welding faults that can prevent a safe welded joint. In civil engineering, convolutional neural networks (CNNs) have been successfully used to detect cracks in roads and buildings by analysing images of the constructed objects. These cracks are found in static objects, whereas the generation of a welding crack is a dynamic process. Detecting the formation of cracks as early as possible is greatly important to ensure high welding quality. In this study, two end-to-end models based on long short-term memory and three-dimensional convolutional networks (3D-CNN) are proposed for automatic crack formation detection. To achieve maximum accuracy with minimal computational complexity, we progressively modify the model to find the optimal structure. The controlled tensile weldability test is conducted to generate long videos used for training and testing. The performance of the proposed models is compared with the classical neural network ResNet-18, which has been proven to be a good transfer learning model for crack detection. The results show that our models can detect the start time of crack formation earlier, while ResNet-18 only detects cracks during the propagation stage.
McSAS3 is a refactored software package for fitting large batches of (X-ray or Neutron) scattering data. It uses a Monte-Carlo acceptance-rejection algorithm to optimize model parameters - ideal for analysis of size-disperse scatterers.
The refactored code can exploit multiprocessing, traceably stores (multiple) results in the output file, and allows for re-histogramming of previous optimizations. Besides analysis of large batches, it can also be integrated in automated data processing pipelines.
The live demonstration will show how to use the software, what its limitations are, and what outcomes can look like for batches of results.
Simulates X-ray and Neutron scattering patterns from arbitrary shapes defined by STL files.
Features:
- Uses multithreading to compute a number of independent solutions, then uses the variance of the results to estimate an uncertainty on the output.
- Can be launched from the command line using an excel sheet to define settings, or from a jupyter notebook.
- Outputs scattering patterns in absolute units if the contrast is set.
- A Gaussian size distribution is available, where the relative scaling of objects for each repetion can be varied. Recommended to be used with limited width (max. 10%) to avoid artefacts.
- Writes results with settings to an archival HDF5 file.
Application examples:
This software has been used in several studies to date. For example, it has been used here to simulate a model scattering pattern for a cuboid shape, which was then fed forward into the McSAS3 analysis program for analyzing scattering patterns of polydisperse cuboids. A second use is here, where it was used for the modeling of flattened helices. In this paper, scattering pattern features could be matched with particular morphological changes in the structure. Lastly, this paper has an example where it was used to validate the analytical analysis model, and explore the realistic limits of application of the analytical model.
Optimized planning and evaluation of dental implant fatigue testing: A specific software application
(2020)
Mechanical complications in implant-supported fixed dental prostheses are often related to implant and prosthetic design. Although the current ISO 14801 provides a framework for the evaluation of dental implant mechanical reliability, strict adherence to it may be difficult to achieve due to the large number of test specimens which it requires as well as the fact that it does not offer any probabilistic reference for determining the endurance limit. In order to address these issues, a new software program called ProFatigue is presented as a potentially powerful tool to optimize fatigue testing of implant-supported prostheses. The present work provides a brief description of some concepts such as load, fatigue and stress-number of cycles to failure curves (S-N curves), before subsequently describing the current regulatory situation. After analyzing the two most recent versions of the ISO recommendation (from 2008 and 2016), some limitations inherent to the experimental methods which they propose are highlighted. Finally, the main advantages and instructions for the correct implementation of the ProFatigue free software are given. This software will contribute to improving the performance of fatigue testing in a more accurate and optimized way, helping researchers to gain a better understanding of the behavior of dental implants in this type of mechanical test.
The SPONGE
(2020)
Small-angle scattering (SAS) offers a reliable route to characterize the nanostructure of large amounts of material with a minimum of tedium, for example, easily extracting size distributions and volume fractions. There are a variety of analysis programs available while the evaluation of SAS measurements has been dominated by the classical curve fitting approach. SASfit represents such a classical curve fitting toolbox: it is one of the mature programs for SAS data analysis and has been available and used for many years. The latest developments will be presented and a scattering function of a mass fractal model of branched polymers in solution is provided as an example for implementing a plug-in.
Alternatively to classical curve fitting, part two presents the latest developments of the user-friendly open-source Monte Carlo regression package McSAS. The form-free Monte Carlo nature of McSAS means, it is not necessary to provide further restrictions on the mathematical form of the parameter distribution: without prior knowledge, McSAS is able to extract complex multimodal or odd- shaped parameter distributions from SAS data. The headless mode is presented by an example of operation within interactive programming environments such as a Jupyter notebook.