We present GekkoFS, a temporary, highly-scalable burst buffer file system which has been specifically optimized for new access patterns of data-intensive High-Performance Computing (HPC) applications. The file system provides relaxed POSIX semantics, only offering features which are actually required by most (not all) applications. It is able to provide scalable I/O performance and reaches millions of metadata operations already for a small number of nodes, significantly outperforming the capabilities of general-purpose parallel file systems.

The displacement response of structures can provide important information regarding their behaviours. Since image-based measurements have been considerably improved in recent years, in this research, a fiducial marker-based approach is exploited and a displacement measurement technique is proposed. With this method, an industrial camera can be used to measure the displacement of a structure through the detection of an ArUco marker. A comparison with an established reference technique has been carried out. The uncertainty due to repeatability measurement and the relevance of the acquisition and environmental parameters on the measurement results were researched.

Measurement of velocity fields is a fundamental topic in fluid dynamics. Image-based analysis methods such as Particle Image Velocimetry or Laser Doppler Velocimetry are usually used. However, these techniques need complex instrumentation and particular test conditions. In this work, a computer vision-based approach is developed in order to obtain vapour velocity field map in effective, robust and economic way. Moreover, iterative filtering algorithm is applied to improve the results. The implemented method is tested on a suction system for domestic use, and the obtained velocity maps are validated by hot-wire anemometry, leading to totally comparable results, both in terms of profile and mean velocity. Uncertainty analysis shows acceptable results, considering the random nature of the phenomenon.

No abstract

<p class="Abstract">Additive Manufacturing (AM) is becoming a widely employed technique also in mass production. In this field, compliances with geometry and mechanical performance standards represent a crucial constrain. Since 3D printed products exhibit a mechanical behaviour that is difficult to predict and investigate due to the complex shape and the inaccuracy in reproducing nominal sizes, optical non-contact techniques are an appropriate candidate to solve these issues. In this paper, 2D digital image correlation and thermoelastic stress analysis are combined to map the stress and the strain performance of an airless wheel prototype. The innovative airless wheel samples are 3D-printed by fused deposition modelling and stereolithography in poly-lactic acid and photopolymer resin, respectively. The static mechanical behaviour for different wheel-ground contact configurations is analysed using the aforementioned non-contact techniques. Moreover, the wheel-ground contact pressure is mapped, and a parametric finite element model is developed. The results presented in the paper demonstrate that several factors have great influence on 3D printed airless wheels: a) the type of material used for manufacturing the specimen; b) the correct transfer of the force line (i.e., the loading system); c) the geometric complexity of the lattice structure of the airless wheel. The work confirms the effectiveness of the proposed non-contact measurement procedures for characterizing complex shaped prototypes manufactured using AM.</p>