Pinpointing subcellular protein localizations from microscopy images is easy to the trained eye, but challenging to automate. Based on the Human Protein Atlas image collection, we held a competition to identify deep learning solutions to solve this task. Challenges included training on highly imbalanced classes and predicting multiple labels per image. Over 3 months, 2,172 teams participated. Despite convergence on popular networks and training techniques, there was considerable variety among the solutions. Participants applied strategies for modifying neural networks and loss functions, augmenting data and using pretrained networks. The winning models far outperformed our previous effort at multi-label classification of protein localization patterns by ~20%. These models can be used as classifiers to annotate new images, feature extractors to measure pattern similarity or pretrained networks for a wide range of biological applications.
This paper investigates the influence of boundary layer skew on flow structure, total pressure loss, and flow control technique numerically on a high-loaded axial-flow compressor cascade. We have developed two new models respectively about loss evaluation and end wall flow control mechanism for more specific analysis. The result shows that boundary layer skew weakens the secondary flow and delays the generation of passage vortex when incidence approaches 0°. This results in a reduction of total pressure loss mainly (89.4%) due to relieved corner separation. However, as incidence exceeds a certain value (+7°), severe corner separation or even earlier corner stall can be induced by inlet boundary layer skew. Optimization procedure for profiled end wall at inflow condition of +7° incidence is further carried out to investigate the impact of boundary layer skew on flow control technique. The result shows that boundary layer skew should be counted in the optimization design of profiled end wall because of its significant influence on the development of end wall flow. The optimum profiled end walls for cases with and without boundary layer skew show great difference in the manner of end wall flow control. According to the improvement of cascades’ performance, end wall profiling seems more efficient in reducing loss when influenced by the boundary layer skew.
This paper presents numerical and experimental investigations about grooved casing treatment with the help of a high-speed small-scale compressor rotor. First, the numerical investigation seeks to offer a contribution of understanding the working mechanism by which circumferential grooves improve stall margin. It is found that stall margin gain due to the presence of circumferential grooves arises from the suction-injection effect and the near-tip unloading effect. Based on that, the philosophy of design of experiment is then set up. Finally, parametric studies are carried out through systematical experiments. It is found that the orthogonal experiment and the factorial analyses are successful in identifying the “best casing configuration” in terms of stall margin improvement. However, the ineffectiveness of the deduction from simulations suggests that the secondary flow circulations on stall margin gain should not be neglected, and the overall contribution of each groove to stall margin gain depends on its unloading effect and the compound effect of suction-injection. Further numerical investigation will focus on how to set up quantitative criteria to evaluate the compound effect of suction-injection and the unloading effect on stall margin gain respectively in each groove.
Three-dimensional unsteady calculations were performed to explore the unsteady nature of tip clearance flow and its possible linking with the spike stall inception. It was found that the interaction of the broken-down leakage vortex with the tip clearance flow formed another distinctive vortex, denoted as the tip separation vortex. It formed below the rotor blade tip section and propagated diagonally inward. This vortex propagated across the rotor passage from the pressure side to the suction side with its vortex core filled with low-energy fluid. The spike emergence during stall inception included the breakdown of the tip leakage vortex and the formation and movement of the tip separation vortex.
The use of slots and grooves in the shroud over the tips of compressor blades, known as casing treatment, is known as a powerful method to control tip leakage flow through the clearance gap and enhance the flow stability in compressors. This paper present a detailed steady and unsteady numerical studies of the coupled flow through rotor blade passages and two different types of casing treatment for a modern subsonic axial-flow compressor rotor. Particular attention was given to examining the interaction between the tip leakage flow and the casing treatment. In order to validate the multi block model applied in the rotor blade end-wall region, the computational results for the modern subsonic compressor rotor both with and without casing treatment were correlated with available experimental test data for estimation of the global performance. Detailed analyses of the flow visualization at the tip have exposed the different tip flow topologies between the cases with casing treatment and with untreated smooth wall. It was found that the primary stall margin enhancement afforded by the casing treatment is a result of the tip clearance flow manipulation. The repositioning of the tip clearance vortex further towards the trailing edge of the blade passage and delaying the movement of incoming/tip clearance flow interface to the leading edge plane are the physical mechanisms responsible for extending the compressor stall margin.
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.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.