To investigate the influence of throttling processes on dynamic characteristics of compressor stability, the rotating stall development of National Aeronautics and Space Administration Stage 35 was simulated with full-annulus Unsteady Reynolds-Averaged Navier–Stokes under different throttling processes. The numerical methods were verified. By combining Dynamic Mode Decomposition and flow field evolution research, the flow structures and dynamic characteristics of “critical mass flow” under different throttling processes were deeply studied; the flow mechanism of flow instabilities under different throttling processes was explored. It is found that the “critical mass flow” corresponds to the beginning of a rapid decrease in mass flow, mainly characterized by shock forward movement and a larger range of spillage flow. Around “critical mass flow,” if the throttle is still tightening, it presents stall pattern 2; otherwise, it presents stall pattern 1. During the pre-stall, both patterns are dominated by tip clearance vortex (TCV)-shock interference. Stall inception disturbance is generated from TCV-shock interference; pattern 1 presents a single disturbance, while pattern 2 presents multiple disturbances. Subsequently, the TCV-shock interference gradually weakens. The single stall disturbance of pattern 1 gradually develops and stabilizes. The multiple stall disturbances in pattern 2 undergo processes including fusion and disappearance, ultimately developing into a single stall cell. During the stable stall, the throttling processes have no significant impact on the speed of the stall cell, and the flow in the un-stalled region is basically consistent with the speedline. However, the tighter the throttle is, the larger the stalled region, and the weaker the flow capacity of the un-stalled region.

ASPAC is an in-house computational fluid dynamics (CFD) software for the simulation of flow in turbomachinery. In this paper, with a dual-level hybrid and heterogeneous programming method, we optimized the ASPAC software and ran it on the Sunway TaihuLight supercomputer. Then the unsteady implicit simulation of the 13-stage twin-spool compressor with 6.116 billion grid cells was realized using 611,000 cores. This is an important step closer to solving one of the four challenges mentioned in the “CFD Vision 2030 Study” proposed by NASA. A general solution method for the twin-spool compressor was first proposed. In addition, an efficient large-scale parallel processing method of row interface data exchange was proposed. After that, a series of optimization methods, such as DMA transmission, data packaging, and linear equation system solving process reconstruction, were explored. The test results show that the proposed parallel processing method of row interface data exchange greatly reduces the memory occupation, and the speed of establishing the row interface exchange relationship is increased by two orders of magnitude. Strong scalability testing in the twin-spool 13-stage compressor showed that for 0.765 billion grid cells, ASPAC can achieve 85.0% parallel efficiency when the cores are increased from 10,400 to 624,000; for 6.161 billion grid cells, ASPAC can achieve nearly 100% parallel efficiency when the cores are increased from 208,000 to 611,000. The test shows that the heterogeneous parallel modification of ASPAC on Sunway does not affect the calculation accuracy, and has high scalability, which can play a major role in the large-scale simulation of compressors.

In order to solve the problem of target detection in infrared images, a single-frame infrared point target detection method based on YOLO-v5 is proposed. Firstly, the problems and difficulties of single-frame infrared target detection are analyzed, and the principles of three commonly used single-frame detection algorithms based on background prediction are introduced. Then the algorithm principle and network structure of YOLO-v5 are elaborated in detail. Finally, 12 groups of infrared images with different backgrounds are constructed for target detection experiments, and the experimental results prove that YOLO-v5 has great advantages over traditional single-frame detection algorithms.

The influence of rotating inlet distortion (RID) on compressor stability has attracted widespread attention in academia, mainly about stall margin and onset of rotating stall. However, most of the experiments are conducted on low-speed compressors, most of the simulations are conducted on rotors, and only one distorted region exists at the inlet. This paper aims to explore the influence of RID with more than one distorted region on rotating stall development in the high-speed multi-row compressor. Flow instability of NASA stage 35 under uniform inlet condition and RID with different rotational speeds are simulated. Results indicate that, when rotating stall is fully developed, the RID has little influence on the rotational frequency and number of stall cells, which may be the inherent characteristics of the compressor. However, the onset and the initial state of rotating stall will be primarily influenced. RID with lower rotational speed will delay the occurrence of rotating stall, RID with higher rotational speed will exceed the occurrence of rotating stall but will experience two types of stall states. The causes for these flow phenomena are then discussed in detail. This research’s findings can help understand the instability behavior of downstream compressor in multi-spool engine, under the distortion flow induced by upstream compressor.

Compared to triangular meshes, quadrilateral meshes offer numerous advantages, such as enhanced fitting accuracy, the capacity to maintain geometric features, and alignment with specific problem orientations. However, there is currently no flawless method for the automatic generation of full quadrilateral meshes, which often leads to challenges such as subpar mesh quality and inadequate algorithm robustness when addressing regions with intricate boundaries. To overcome the limitations associated with full quadrilateral mesh generation, this paper proposes a parallel and anisotropic quad-dominant mesh generation framework based on Riemann frame fields (RFF). This framework (referred to as RFF-meshing) utilizes the triangulation of the geometric model as input and works following the four steps. Firstly, the geometric constraint boundaries are extracted through feature recognition. Secondly, RFF is computed on the surface using combinatorial alignment strategies. Thirdly, the background mesh is partitioned into sub-regions according to frame fields and feature edges. Finally, the anisotropic size field computation is conducted and the quadrilateral mesh is generated in parallel for all partitions. The proposed method has been evaluated with multiple computational fluid dynamics models with varying features, demonstrating its effectiveness and efficiency.

Abstract To achieve a large-scale and efficient numerical solution to supersonic combustion chemical reaction problems, a numerical simulation algorithm for combustion chemical reactions was established on a graphics processing unit (GPU). Numerical simulation of the combustion chemical reaction of a hydrocarbon fuel/air mixture in a supersonic combustor was conducted to verify the accuracy of GPU parallel calculation results and analyze its parallel performance. Numerical results show that GPU parallel computing can accurately simulate the complex flowfield in the combustor. The distribution law of wall pressure is basically consistent with the experimental values, and the numerical agreement is good, indicating that the numerical simulation algorithm for the multi-GPU supersonic combustion chemical reaction established in this paper is correct and reliable. On the Tesla V100 GPU parallel computing platform, the speedup of GPU parallel computing can reach over 100 times, greatly improving computational efficiency. When the grid size is 18.64 million, the parallel efficiency of single and double precision floating-point operations on four GPUs is 66.2% and 69.1%, respectively, which demonstrates that the multi-GPU parallel algorithm has good scalability.

To address the new functional requirements brought by the introduction of new weapons and new combat modes, a comprehensive survey of the research progress in the area of combat simulation software is performed from the perspective of software engineering. First, the top-level specification, simulation engine, and simulation framework of combat simulation software are reviewed. Then, several typical combat simulation software systems are demonstrated, and the relevant software frameworks are analyzed. Finally, combining the application prospect of artificial intelligence, metaverse, and other new technologies in combat simulation, the development trends of combat simulation software are presented, namely intellectualization, adaptation to an LVC (live, virtual, and constructive) system, and a more game-based experience. Based on a comprehensive comparison between the mentioned simulation frameworks, we believe that the AFSIM (Advanced framework for simulation, integration, and modeling) and the E-CARGO (Environments—classes, agents, roles, groups, and objects) are appropriate candidates for developing distributed combat simulation software.

The assembly of overlapping grids is a key technology to deal with the relative motion of multi-bodies in computational fluid dynamics. However, the conventional implicit assembly techniques for overlapping grids are often confronted with the problem of complicated geometry analysis, and consequently, they usually have a low parallel assembly efficiency resulting from the undifferentiated searching of grid nodes. To deal with this, a parallel implicit assembly method that employs a two-step node classification scheme to accelerate the hole-cutting operation is proposed. Furthermore, the aforementioned method has been implemented as a library, which can be conveniently integrated into the existing numerical simulators and enable efficient assembly of large-scale multi-component overlapping grids. The algorithm and relevant library are validated with a seven-sphere configuration and multi-body trajectory prediction case in the aspects of parallel computing efficiency and interpolation accuracy.

Although significant advances have been made in full-annulus simulation of flow in single-spool compressors. However, no work has been devoted to full-annulus simulation of flow in multi-spool compressors. The aim of the paper is to develop techniques for full-annulus simulation of unsteady flows in multi-spool compressor. The techniques for automatic generation of large-scale full-annulus grid of multi-spool compressor, for uniting simulation in LPC(Low Pressure Compressor) and HPC(High Pressure Compressor) in one solver, for processing LPC/HPC interface in multi-spool compressor in parallel environment, for managing vast amounts of flow field data in multi-spool compressor with low space complexity are proposed. Based on an in-house software ASPAC, these techniques are then applied to full-annulus unsteady simulation of flow in a 13-stage multi-spool compressor with 800 million grid cells using 10240 processes. The numerical simulation results are validated through compared with results of empirical model. Meanwhile, it indicates that results of large-scale unsteady simulation are more precise than results of steady simulation.

This paper presents an efficient nearest point projection method based on improved Newton iteration. This method is different from the traditional Newton iterative method. For any complex curve and surface, the initial value fast positioning method is used to obtain the initial iterative point set that may meet the conditions. The initial point set is screened by the circle and sphere clipping algorithm to obtain the initial iterative points to be selected. Then the improved unidirectional Newton iterative algorithm is applied to search the region, and the parameter values in accordance with precision requirements are obtained. The corresponding points are calculated according to the parameter values. Finally, through a large number of test cases, it is shown that the method in this paper is robust and efficient in adapting to the shape change of curves and surfaces.

HRADesign系统作为通用旋翼翼型气动设计和评估系统,研制目的主要是为工业设计环境提供通用、高效、鲁棒的优化设计架构,应用于各类旋翼翼型族的设计,提高旋翼翼型设计的效率和精度,以满足先进直升机对高性能旋翼翼型的迫切需求.旋翼翼型设计技术是直升机旋翼设计的核心技术,旋翼翼型的优化设计具有多点、多目标、强约束的特点.HRADesign系统针对旋翼翼型设计的特点,发展了多目标进化算法、PCA多目标降维技术、Kriging代理模型、基于CST方法的翼型参数化技术以及高精度CFD等优化设计技术,构建了基于进化算法的多目标优化流程.通过详细介绍平台的系统架构、主要的功能模块以及多目标优化流程,展现了系统架构设计的灵活性和功能模块的完备性.通过ADODG基准测试算例、某厚度旋翼翼型常规多目标优化算例和考虑多目标降维的优化算例进行了系统功能验证,优化结果表明,在满足约束的条件下,优化后的旋翼翼型和基准翼型相比,综合性能都有明显改善,验证了优化设计系统的有效性、可靠性.

重叠网格装配是处理计算流体力学领域多体相对运动问题的关键技术之一.针对常见重叠网格隐式装配方法中几何分析过程复杂、节点无差别并行查找操作影响并行装配效率等问题,提出一种高度自动化的重叠网格隐式装配方法.首先,基于协方差分析、切割盒子等快速算法,将壁面距离计算与贡献单元存在性判断解耦,实现网格组动态重叠关系的自动化识别;其次,结合集合分析,设计出并行化的自动挖洞算法;最后,通过快速查询方法建立重叠单元与贡献单元的插值关系.针对所实现的并行重叠网格隐式装配工具库,采用某五球体部件验证了自动挖洞逻辑的准确性,并通过机翼-挂架-外挂物(WPFS)模型检验了重叠单元与贡献单元插值关系的准确性.

There are many multifaceted problems in the aerospace engineering domain, which demand the multidisciplinary simulations for coupling analysis and the development of coupling frameworks. In order to address the issue of lacking domestic coupling framework, we propose an aerodynamics-centric framework for multidisciplinary coupling analysis, and then demonstrate the usage of it in developing a structured-unstructured grid coupling software, which is validated with an external store separation case study.

多学科优化设计框架在飞行器外形设计领域发挥着至关重要的作用.针对国内缺少自主MDO框架的状况,基于工作流技术开发出一款MDO框架软件.该软件完全遵循软件工程思想,集成了多个学科分析工具链和多套优化算法,支持本地/集群模式的学科分析计算,并借助数据库实现了数据的持久性存储.针对某飞行器进行了多目标多学科的优化设计实践,结果表明该框架具有较高的自动化程度和良好的通用性.

Aiming at the detection of dim and small moving point targets in space-based infrared (IR) imaging systems, a detection method based on spatial-temporal local contrast (STLC) is presented. Firstly, use the spatial and temporal information to estimate the background gray value of each pixel, and obtain two prediction results respectively. Then, the spatial local contrast (SLC) and the temporal local contrast (TLC) are defined according to the original image and two estimated backgrounds. Next, the STLC is defined as the product of SLC and TLC. Finally, a threshold value is set in the STLC for target detection. Four groups of experiments are conducted, and the simulation results indicate that the proposed method has a great advantage in terms of background suppression factor (BSF) and gain of signal-to-noise (GSNR).

量子计算是最重要的后摩尔计算技术之一,拥有经典计算机无可比拟的超强计算能力,未来能够对各行业应用产生颠覆性的影响.针对量子计算给空气动力学带来的机遇和挑战,详细综述了量子计算机、量子算法、量子底层软件栈等方面的研究进展.结合空气动力学领域常用的基础方法,在综述量子计算线性方程组求解、插值操作、数值积分、优化搜索等最新进展的基础上,结合典型量子算法深入分析了量子计算在空气动力学领域的应用前景,并指出了需要重点关注的研究方向.

Meshes are the input for computational fluid dynamics (CFD) analysis, whose size and quality have important impact on the simulation results. With the continuous advancement of high-fidelity CFD simulation, the scale of needed computational meshes has become larger and larger, which poses great challenges to the development of interactive mesh generation software. To address this issue, a parallel and distributed framework called PadMesh is proposed, which performs as the infrastructure for developing various interactive mesh generation software (e.g., structured, unstructured and Cartesian). First, the framework PadMesh is demonstrated, including the introduction of design principles, software architecture and key components, namely message-oriented middleware, client application, server application and parallel supporting module. Second, a parallel and distributed structured mesh generation software called PGridStar is developed based on PadMesh. Strategies are investigated on managing the visual data and distributed data for structured meshes. Finally, two functionalities of the preliminary PGridStar are presented, which validate the usability of PadMesh in developing interactive mesh generation software.

Mesh generation is a key preprocessing step in the computational fluid dynamics (CFD). The scale and quality of the generated mesh have an important influence on the CFD simulation results. In order to meet the requirements of large-scale unstructured mesh generation, the application of PadMesh (a distributed and parallel mesh generation software framework based on C/S architecture) in developing parallel unstructured mesh generation software is investigated. Two domain decomposition strategies have been proposed, namely the region growing strategy and the bounding box strategy. Three practical configurations have been utilized to verify the two proposed strategies. Experimental results show that the region growing strategy can generate subdomains with good load balance, while the bounding box strategy is suitable for the configurations with symmetrical features and can obtain subdomains with high load balance and good data locality.

The structured meshes are widely utilized in the aircraft industry because of their efficiency in achieving accurate solutions, especially for viscous flows of complex configurations. However, they are usually generated by engineers through enormous and tedious interaction with the graphical user interface of mesh generation software. To address this issue, we implement one piece of interactive structured mesh generation software called NNW-GridStar. First, we present its design details, including the software architecture and data structure. Second, we demonstrate its implementation and emphasize on the accelerated techniques proposed in our software, which consist of automatic boundary-layer mesh generation, rapid block assembly, multiblock stretching, and O-type block stretching. Finally, we perform a deep comparison between the developed NNW-GridStar and another two pieces of well known commercial software, and then demonstrate the application of NNW-GridStar in the productional CFD simulations. Experimental results show that NNW-GridStar has an outstanding generation speed and can produce meshes with high quality.

未来航空工业的发展,需要解决多学科综合设计的关键问题,为新型高性能飞行器的设计提供有力的设计方法和设计工具.DIPasda作为复杂外形设计的通用飞行器多学科优化设计平台,研制目的主要是提供一套新型通用、鲁棒、高效的优化设计架构,应用于通用飞行器工业设计环境,改善传统设计耗时低效的状况,提高新型飞行器设计的效率和精度.DIPasda平台系统包含了优化设计过程中所需用到的各类方法,主要包括数值优化算法、几何模型参数化方法、代理模型方法、高精度的学科分析工具等.通过详细介绍平台的系统架构、主要的功能模块、伴随优化设计和多目标优化设计流程,展现了DIPasda平台系统架构设计的灵活性和功能模块的完备性.最后通过优化算例展示了系统的综合优化设计能力.