Tosson, Amir Samy

In this article, we present the experimental protocol and data-processing framework for megahertz X-ray Photon Correlation Spectroscopy (MHz XPCS) experiments on soft matter samples implemented at the Materials Imaging and Dynamics (MID) instrument of the European X-ray Free-Electron Laser (EuXFEL). Due to the introduction of a standard configuration and the implementation of a highly automated data-processing pipeline, MHz-XPCS measurements can now be conducted and analyzed with minimal user intervention. A key challenge lies in managing the extremely large data volumes generated by the Adaptive Gain Integrating Pixel Detector (AGIPD) – often reaching several petabytes within a single experiment. We describe the technical implementation, discuss the hardware requirements related to effective parallel data processing and propose strategies to enhance data quality, in particular related to data reduction strategies and an improvement of the signal-to-noise ratio. Finally, we address strategies for making the processed data FAIR (Findable, Accessible, Interoperable, Reusable), in alignment with the goals of the DAPHNE4NFDI project.

Big data will be a source of new economic value and innovation. But even more is at stake. Big data’s ascendancy represents […] shifts in the way we analyze information that transforms how we understand and organize society”. This important statement by Mayer-Schönberger and Cukier highlights the crucial benefit this field offers to the research and business communities. It is opening up an entirely new horizon for economic and innovative solutions. It also sheds light on the main challenge researchers and business analytics across the world have been facing during the recent years, namely big data handling, extraction and employing. Digitization enabling the real-time data visualization for the further analysis is now a vital prerequisite for a successful and innovative scientific community. This work aims to contribute to the transformation of the crystallography community, dealing with an enormous volume of the measured data obtained due to technological advancement in radiation and detection, into a smarter scientific environment. \\ Specifically, this dissertation has a twofold aim. First, it is intended to provide an introduction and implementation guideline to scientists in the crystallography community and across the fields demonstrating how trend technologies might be applied to design cutting-edge research projects. Second, considering the Energy Dispersive Laue Diffraction (EDLD) as a case study, it introduces an innovative approach to exploit high-performance computing to develop a novel framework for data processing within the time frame of a few seconds compared to the traditional analytic systems requiring a few hours to process the same amount of data. The framework developed is based on multiple artificial intelligence algorithms, designed to perform tasks that cannot be processed by basic programming techniques. Extending this approach, data clouding has been employed to establish the communication channels between scientists and the collected data. This computing solution helps to achieve commoditization of computational resources, implementation of open-source software, data virtualization, globalization of workforce, establishing a data-sharing point. On the whole, due to these technological improvements, the crystallography community might gain maximum benefit from the collected data. As a proof of concept, the reliability, efficiency, and performance of the entire work has been verified by involving the system in a challenging task, namely: the one-shot analysis of the micro texture in polycrystalline materials.

Tosson, Amir Samy

Filters

Settings

Sort By

Results per page