Physics-based analyses have the potential to consolidate and substantiate medical diagnoses in rhinology. Such methods are frequently subject to intense investigations in research. However, they are not used in clinical applications, yet. One issue preventing their direct integration is that these methods are commonly developed as isolated solutions which do not consider the whole chain of data processing from initial medical to higher valued data. This manuscript presents a workflow that incorporates the whole data processing pipeline based on a environment. Therefore, medical image data are fully automatically pre-processed by machine learning algorithms. The resulting geometries employed for the simulations on high-performance computing systems reach an accuracy of up to 99.5% compared to manually segmented geometries. Additionally, the user is enabled to upload and visualize 4-phase rhinomanometry data. Subsequent analysis and visualization of the simulation outcome extend the results of standardized diagnostic methods by a physically sound interpretation. Along with a detailed presentation of the methodologies, the capabilities of the workflow are demonstrated by evaluating an exemplary medical case. The pipeline output is compared to 4-phase rhinomanometry data. The comparison underlines the functionality of the pipeline. However, it also illustrates the influence of mucosa swelling on the simulation.
Computational fluid dynamics (CFD) is the main field of computational mechanics that has historically benefited from advances in high-performance computing. Highperformance computing involves several techniques to make a simulation efficient and fast, such as distributed memory parallelism, shared memory parallelism, vectorization, memory access optimizations, etc. As an introduction, we present the anatomy of supercomputers, with special emphasis on HPC aspects relevant to CFD. Then, we develop some of the HPC concepts and numerical techniques applied to the complete CFD simulation framework: from preprocess (meshing) to postprocess (visualization) through the simulation itself (assembly and iterative solvers).
In this paper, deep learning (DL) methods are evaluated in the context of turbulent flows. Various generative adversarial networks (GANs) are discussed with respect to their suitability for understanding and modeling turbulence. Wasserstein GANs (WGANs) are then chosen to generate small-scale turbulence. Highly resolved direct numerical simulation (DNS) turbulent data is used for training the WGANs and the effect of network parameters, such as learning rate and loss function, is studied. Qualitatively good agreement between DNS input data and generated turbulent structures is shown. A quantitative statistical assessment of the predicted turbulent fields is performed.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.