Advances in Engineering Software最新文献

The dynamics of systems of systems often involve complex interactions among the individual systems, making the implications of design choices challenging to predict. Design features in such systems may trigger unexpected behaviors or result in large variations in safety, performance or resilience. To provide a means of simulating such systems for aiding in these decisions, we have developed a prototype tool, the control-oriented dynamic computational modeling tool (CDCM). The CDCM provides rapid simulation capabilities to perform trade studies in systems of systems. The general class of systems of systems that we aim to examine involve multiple hazards, damage, cascading consequences, repair and recovery. We especially focus on systems-of-systems that incorporate a health management system (HMS) that can monitor the state of the habitat and make decisions about actions to take. In this paper we describe the features of the CDCM, the architecture we devised for simulation of systems-of-systems, the unique functionalities of this tool, and we provide a demonstration of the capabilities by performing two illustrative examples. We articulate the use of this tool for making early design decisions and demonstrate its use for trade studies that consider a model of a deep space habitat. We also share some experiences and lessons that may be useful for others seeking to address similar problems.

Deep learning algorithms have been employed for real-time concrete crack detection. However, many algorithms are not specifically tailored for this purpose. Moreover, their lightweight iterations are generally optimized at the macro-model level, leaving room for further lightweight enhancements at the block level. Therefore, this study developed an enhanced YOLOv3 (You Only Look Once Network v3) model, named YOLO-Crack. The structural optimization of the model takes into consideration the shapes of concrete cracks in the dataset. Meanwhile, two multiple branch-shaped blocks based on dilated convolutions, convolutions and pooling operations were proposed. The two blocks, incorporating depthwise separable convolutions and attention mechanisms, were used to rebuild the model at the block level. These enhancements significantly reduce the size and improve the detection performance of YOLO-Crack. Furthermore, YOLO-Crack was softwareized for real-time detection of concrete cracks. The software was designed to support parallel computing, allowing for real-time detection of concrete cracks even on laptops with limited computing power. It was utilized to detect cracks on concrete roads at a university in Nanjing, China, enabling real-time detection at a frame rate of 30 frames per second with satisfactory accuracy.

Modular deployable antennas represent an ideal structural form for the development of large-aperture antennas because of their flexibility, adaptability, and high versatility. To enhance the surface accuracy of the antenna after deployment, a comprehensive pre-tension design method that considers truss deformation and tension uniformity is proposed. First, the configuration design of the antenna cable net is constructed, and the mathematical models for cable length under surface accuracy requirements and boundary nodes are established considering catenary effects. Second, the distribution patterns of cable net structure nodes and segments are analyzed, leading to the creation of node coordinate matrices and cable net connection matrices. A basic model for cable net pre-tension design is developed based on the fundamental principles of force density. Furthermore, a multi-objective optimization of cable net pre-tension is performed using a genetic algorithm, considering truss structure deformation and tension uniformity as dual factors. Finally, the developed model is applied to design a single-module cable net structure, and numerical simulation is used for validation. Research results show that the overall surface form error is 0.32 mm, and the maximum tension ratio of cable net on the front cable net surface is 1.54, whereas the maximum tension ratio of tension ties is 2.28, thereby meeting the design requirements. Numerical simulation shows that the maximum deformation of the cable net structure is 0.16 mm, validating the correctness of the model. This research can provide valuable insights and references for the pre-tension design and research of cable net structures in other antennas.

This work proposes and investigates a new multi-field-multi-constraint coupled topology optimization problem, in which stress control, design castability, and geometry dimensional shrinkage issues that are of concern to practical engineering are simultaneously considered. In the optimization proposal considered, a pair of special twin designs are generated using a two-projected-field scheme, which maintains a consistent topological configuration and uniform dimensional shrinkage variations during the optimization process. The implicit correlation between these twin designs poses major challenges to their independent stress and castability control. To this end, an appropriate formulation is presented by reasonably integrating stress and casting constraints into the optimization proposal with dimensional shrinkage. And, special numerical techniques including stress penalization, aggregation approximation, approximation correction, and regional regularization are appropriately introduced to construct an effective solution strategy. Typical numerical examples are operated to demonstrate the validity of the proposed method and systematically evaluate its numerical properties. The results indicate that in the absence of necessary stress control measures, the obtained twin designs cannot avoid local high-stress concentration under uniform dimensional shrinkage. In contrast, the proposed method can effectively address this issue, but at the cost of the design stiffness under a given material volume limit. As a result, twin designs used for blueprint and model designs that simultaneously meet stress, castability, and uniform dimensional shrinkage requirements are now readily available.

Reliability analysis of large, complex structures with high reliability has always been a challenging task. The Subset Simulation (SUS) has proven highly effective for high-dimensional and small failure probability problems. However, the computational demands of time-consuming numerical simulations in the field of mechanical engineering remain substantial. Metamodeling techniques provide an effective means to reduce simulation computational costs. This paper introduces a novel approach, termed SUSAK, which combines the Subset Simulation with Kriging metamodels to address the challenge of small failure probability calculations. By using the original distribution of input variables representing the structural system, the SUSAK method establishes an adaptive metamodel and iteratively optimizes it with intermediate failure domain samples corresponding to the intermediate events. By replacing the performance function in subsequent calculations of the probability associated with intermediate events with metamodels, the enhanced method significantly reduces the computational burden compared to using the SUS method. This approach does not rely on solving for the design point, is not constrained by the shape of the failure domain, and retains the advantages of the SUS method in high-dimensional small failure probability calculations. As a result, it is well-suited for calculating small failure probabilities in nonlinear, discontinuous failure domains, multiple failure domains, and high-dimensional problems.

The concurrent optimization of structural topologies and fiber orientations offers an effective way for improving the mechanical performance of fiber-reinforced composite structures. However, its local optimal solution problem under stress constraints remains extremely challenging. To tackle this problem, this paper presents an orientation variable corrected principal stress direction (OVCPSD) method, which is composed by a two-step optimization strategy. In the first step, the fiber orientation for the shear-weak materials is roughly distributed along the principal stress direction, and the principal stress direction of the previous iteration step is used as the initial fiber orientation. In the second step, the principal stress direction is further modified by an orientation variable in a small subinterval. Finally, a new optimization model of structural topologies and fiber orientations for minimum structural compliance under stress constraints is established, and the sensitivities of the objective and constraint functions with respect to both the pseudo-density and orientation variable are derived. Three numerical examples are presented to verify the effectiveness of the OVCPSD method. The complete code is available from the website: https://github.com/TopOpt-lbg/OVCPSD_L-shaped-beam.

Collecting reliable pedestrian trajectories in pedestrian behavior analysis, trajectories broken by frame sampling and trajectories crossing in multi-object conditions often hinder their performance of existing pedestrian tracking models. Despite attempts to address these issues by performing detection and tracking simultaneously using deep learning algorithms, previous methods still struggle with errors such as mistaking a single pedestrian for multiple pedestrians. We propose a novel approach to efficiently collect and correct pedestrian trajectories with minimized practical errors in multi-object conditions for urban surveillance systems. Our system utilizes a single vision sensor to automatically collects trajectories of multiple pedestrians and employ simple, low-computational algorithms, particularly the Deep simple online real-time tracking (Deep SORT) method, to calibrate the trajectories from tracking-by-detection models. Additionally, our system identifies and merges broken pedestrian trajectories, treating them as potential single trajectories, while considering their spatiotemporal ranges. We evaluate the proposed system by implementing it on real testbed video footage. Our method significantly improves practical errors and achieves more accurate pedestrian trajectories compared to existing models, and exhibits robust characteristics, effectively handling complex situations such as occlusions and crowds.

The aerodynamic shape of blades in turbo machinery is a key factor that affects blade efficiency, pressure ratio and other performance indicators. To address the lack of global continuity and difficulty in ensuring fairing quality in blade modeling, this paper proposes a novel blade modeling method that uses curvature constraints and a Mid-Closest Criterion to achieve adaptive knot refinement in 2D profile reconstructing and 3D T-mesh establishing process respectively. To begin with, the proposed Curvature-Constrained Periodic Reconstruction (CCPR) is used to reconstruct discrete sampling points of the section profile into a globally fairing periodic curve, and meets minimal error requirement. Additionally, the distribution of control points matches with curvature variation of the curve itself. Next, the Double Curves-guided Mid-Closest T-spline Surface skinning (DCM-TS) is used to construct a T-mesh from the stacked cross-sectional profile curves automatically, and the control vertices of the T-mesh are in a controllable scale and reasonably distributed. Further, the modeling precision is allocated with the curvature information of the ideal blade contour, and based on that, a T-spline skinning surface is obtained through LSPIA method, which updates the model and realizes blade modeling that considers both the design efficiency and accuracy. Eventually the proposed method is verified and compared with several traditional surface skinning methods in practical modeling cases. The result unveils that the presented method is compatible with traditional aerofoil parameterization methods, generates surface with global continuity and high fairing quality, and has sharply reduced and evenly distributed control vertices.

The ability to predict the effective material property of composites with periodic micro-structures based on homogenization theory has been an effective method to analyze structures with complex heterogeneities. Homogenization codes have been made available for educational purposes including the homogenization code for the prediction of effective elasticity and thermal material properties in MATLAB. The aim of this educational paper is to present a Python version of the existing homogenization code and provide detailed diagrams of its key modules extending its ability to conduct analysis and design studies possibly via integration into commercial FEM software. Python has become a popular programming language due to its wide applicability to several disciplines, its portability, its flexibility by means of programming paradigms, its open-source nature, its well-documented libraries, and its easy-to-learn syntax. To increase the applicability and community reach of the homogenization algorithm presented, we provide a Python translation of the well-known MATLAB implementation. By doing so, we aim to increase the integration potential and adaptability of the homogenization approach to other computing packages and target adoption by a wider audience by leveraging the advantages of basing the solution on a free and open-source platform.

A novel enrichment technique is proposed for the solution quality enhancement near weak singularities along straight lines in second-order elliptic boundary value problems. The considered 3D media can generally be heterogeneous. Enrichment is applied through construction of proper 3D singular functions for heterogeneous media. An advantage over similar approaches is that the singular bases are constructed without any knowledge of the analytical singularity order. To this end, the governing PDE is considered in a cylindrical fictitious domain whose axis lies along the singular line, over which the 3D equilibrated singular basis functions are developed. Combined with the smooth solution, the total solution undergoes imposition of the boundary conditions, so that the proper singular functions are automatically built. The singular solution series is primarily made by a combination of first kind Chebyshev polynomials and trigonometric functions, subjected to weighted residual imposition of the PDE to extract the equilibrated singular functions. Meanwhile, the cumbersome 3D integrals break into algebraic combinations of 1D predefined integrals, so that no numerical quadrature will be needed. The bases are tested as a boundary method in the solution of multiple 3D problems including weak singularities, to show their accuracy and efficiency. The proposed singular bases may be implemented in enriched techniques such as XFEM.