In this paper, we describe the structures that produce a spike-type route to rotating stall and explain the physical mechanism for their formation. The descriptions and explanations are based on numerical simulations, complemented and corroborated by experiments. It is found that spikes are caused by a separation at the leading edge due to high incidence. The separation gives rise to shedding of vorticity from the leading edge and the consequent formation of vortices that span between the suction surface and the casing. As seen in the rotor frame of reference, near the casing the vortex convects toward the pressure surface of the adjacent blade. The approach of the vortex to the adjacent blade triggers a separation on that blade so the structure propagates. The above sequence of events constitutes a spike. The computed structure of the spike is shown to be consistent with rotor leading edge pressure measurements from the casing of several compressors: the centre of the vortex is responsible for a pressure drop and the partially blocked passages associated with leading edge separations produce a pressure rise. The simulations show leading edge separation and shed vortices over a range of tip clearances including zero. The implication, in accord with recent experimental findings, is that they are not part of the tip clearance vortex. Although the computations always show high incidence to be the cause of the spike, the conditions that give rise to this incidence (e.g., blockage from a corner separation or the tip leakage jet from the adjacent blade) do depend on the details of the compressor.

BackgroundWith the maturation of next-generation DNA sequencing (NGS) technologies, the throughput of DNA sequencing reads has soared to over 600 gigabases from a single instrument run. General purpose computing on graphics processing units (GPGPU), extracts the computing power from hundreds of parallel stream processors within graphics processing cores and provides a cost-effective and energy efficient alternative to traditional high-performance computing (HPC) clusters. In this article, we describe the implementation of BarraCUDA, a GPGPU sequence alignment software that is based on BWA, to accelerate the alignment of sequencing reads generated by these instruments to a reference DNA sequence.FindingsUsing the NVIDIA Compute Unified Device Architecture (CUDA) software development environment, we ported the most computational-intensive alignment component of BWA to GPU to take advantage of the massive parallelism. As a result, BarraCUDA offers a magnitude of performance boost in alignment throughput when compared to a CPU core while delivering the same level of alignment fidelity. The software is also capable of supporting multiple CUDA devices in parallel to further accelerate the alignment throughput.ConclusionsBarraCUDA is designed to take advantage of the parallelism of GPU to accelerate the alignment of millions of sequencing reads generated by NGS instruments. By doing this, we could, at least in part streamline the current bioinformatics pipeline such that the wider scientific community could benefit from the sequencing technology.BarraCUDA is currently available from http://seqbarracuda.sf.net

The porting of two-and three-dimensional Euler solvers from a conventional CPU implementation to the novel target platform of the Graphics Processing Unit (GPU) is described. The motivation for such an effort is the impressive performance that GPUs offer: typically 10 times more floating point operations per second than a modern CPU, with over 100 processing cores and all at a very modest financial cost. Both codes were found to generate the same results on the GPU as the FORTRAN versions did on the CPU. The 2D solver ran up to 29 times quicker on the GPU than on the CPU; the 3D solver 16 times faster.

This paper describes the role of tip leakage flow in creating the leading edge separation necessary for the onset of spike-type compressor rotating stall. A series of unsteady multipassage simulations, supported by experimental data, are used to define and illustrate the two competing mechanisms that cause the high incidence responsible for this separation: blockage from a casing-suction-surface corner separation and forward spillage of the tip leakage jet. The axial momentum flux in the tip leakage flow determines which mechanism dominates. At zero tip clearance, corner separation blockage dominates. As clearance is increased, the leakage flow reduces blockage, moving the stall flow coefficient to lower flow, i.e., giving a larger unstalled flow range. Increased clearance, however, means increased leakage jet momentum and contribution to leakage jet spillage. There is thus a clearance above which jet spillage dominates in creating incidence, so the stall flow coefficient increases and flow range decreases with clearance. As a consequence, there is a clearance for maximum flow range; for the two rotors in this study, the value was approximately 0.5% chord. The chordwise distribution of the leakage axial momentum is also important in determining stall onset. Shifting the distribution toward the trailing edge increases flow range for a leakage jet dominated geometry and reduces flow range for a corner separation dominated geometry. Guidelines are developed for flow range enhancement through control of tip leakage flow axial momentum magnitude and distribution. An example is given of how this might be achieved.

A new three-dimensional Navier-Stokes solver for flows in turbomachines has been developed. The new solver is based on the latest version of the Denton codes, but has been implemented to run on Graphics Processing Units (GPUs) instead of the traditional Central Processing Unit (CPU). The change in processor enables an order-of-magnitude reduction in run-time due to the higher performance of the GPU. Scaling results for a 16 node GPU cluster are also presented, showing almost linear scaling for typical turbomachinery cases. For validation purposes, a test case consisting of a three-stage turbine with complete hub and casing leakage paths is described. Good agreement is obtained with previously published experimental results. The simulation runs in less than 10 minutes on a cluster with four GPUs.

In this paper we describe the structures that produce a spike-type route to rotating stall and explain the physical mechanism for their formation. The descriptions and explanations are based on numerical simulations, complemented and corroborated by experiments. It is found that spikes are caused by a loss of pressure rise capability in the rotor tip region, due to flow separation resulting from high incidence. The separation gives rise to shedding of vorticity from the leading edge and the consequent formation of vortices that span between the suction surface and the casing. As seen in the rotor frame of reference, near the casing the vortex convects toward the pressure surface of the adjacent blade. The approach of the vortex to the adjacent blade triggers a separation on that blade so the structure propagates. The above sequence of events constitutes a spike. The simulations show shed vortices over a range of tip clearances including zero. The implication is that they are not part of the tip clearance vortex, in accord with recent experimental findings. Evidence is presented for the existence of these vortex structures immediately prior to spike-type stall and, more strongly, for their causal connection with spike-type stall inception. Data from several compressors are shown to reproduce the pressure and velocity signature of the spike-type stall inception seen in the simulations.

Endwall loss, often termed “secondary loss”, in axial turbines has been intensively studied for many years, despite this the physical origin of much of the loss is not really understood. This lack of understanding is a serious impediment to our ability to predict the loss and to the development of methods for reducing it. This paper aims to study the origins of the loss by interrogating the results from detailed and validated CFD calculations. The calculation method is first validated by comparing its predictions to detailed measurements in a turbine cascade. Very good agreement between the calculations and the measurements is obtained. The solution is then examined in detail to highlight the sources of entropy generation in the cascade, several different sources of loss are found to be significant. The same blade row is then used to study the effects of the of the inlet boundary layer thickness on the loss. It is found that only the inlet boundary layer loss and the mixing loss vary greatly with inlet boundary layer thickness. Finally a complete 50% reaction stage, with identical stator and rotor blade profiles, is examined using both steady calculations, with a mixing plane model, and the time average of unsteady calculations. It is found that the endwall flow in the rotor is completely different from that in the stator. Because of this it is considered that results from endwall flow and loss measurements in cascades are of limited relevance to the endwall flow in a real turbine. The results are also used to discuss the validity of the mixing plane model.

The implementation of a two-dimensional Euler solver on graphics hardware is described. The graphics processing unit is highly parallelized and uses a programming model that is well suited to flow computation. Results for a transonic turbine cascade test-case are presented. For large grids (106 nodes) a 40 times speed-up compared with a Fortran implementation on a contemporary CPU is observed.