The failure accidents of prestressed concrete cylinder pipe (PCCP) seriously affect the economic feasibility construction site. traditional method needing to stop for inspection is time-consuming and laborious. This paper studies PCCP broken wire identification algorithm based on deep learning. A wire-breaking test platform was built; Distributed Fiber Acoustic Sensing Monitoring System (DAS) monitors wire-breakage events in DN4000mm PCCPs buried underground. collected signal creates a...

The low energy conversion efficiency of electromagnetic acoustic transducers (EMATs) is a critical issue in nondestructive testing applications. To overcome this shortcoming, butterfly coil EMAT was developed and optimized by numerical simulation based on 2-D finite element model. First, the effect structural parameters investigated orthogonal test theory. Then, modified designed that consists three-square permanent magnets with opposite polarity (TSPM-OP) to enhance signal amplitude....

Wire breakage is a major factor in the failure of prestressed concrete cylinder pipes (PCCP). In presented work, an automatic monitoring approach broken wires PCCP using fiber-optic distributed acoustic sensors (DAS) investigated. The study designs 1:1 prototype wire break experiment DN4000 mm buried underground simulated test environment. combines collected signals with previously noise operating pipe and transforms them into spectrogram as signal dataset. A deep learning-based target...

Abstract Simultaneous Localization and Mapping (SLAM) is one of the key technologies in robot navigation autonomous driving, playing an important role navigation. Due to sparsity LiDAR data singularity point cloud features, accuracy loss SLAM can occur during matching localization. In response these issues, this paper proposes a Measurement algorithm that integrates multi type geometric feature extraction optimized registration algorithms. This article first adopts advanced ground...

Slam (simultaneous localization and mapping) play an important role in the field of artificial driverless intelligence. A real-time dynamic visual SLAM algorithm based on object detection network is proposed to address robustness camera accuracy issues caused by objects indoor scenes. The YOLOv5s model, which has smallest depth feature map width YOLOv5 series, chosen as network. backbone replaced with lightweight ShuffleNetv2 Experimental results VOC2007 dataset show that YOLOv5-LITE model...

The optimal design of vehicle suspensions is a complex optimization problem which may have highly nonlinear spaces with many local optima. At present, there are two main problems for the suspensions. first that global solution hardly reached using numerical methods or simple genetic algorithms. second suspension models all lumped mass in Virtual prototyping software can deal multibody and provide good approximation to real system performance different simulated scenarios, but method method,...

The energy conversion of electromagnetic acoustic transducers (EMATs) is typically lower, which seriously restricts the application EMATs in field non-destructive testing and evaluation. In this work, parameters surface wave EMATs, including structural electrical parameters, are investigated using orthogonal test method to improve transducer's efficiency. Based on established finite element 2-D model amplitude displacement components at observation point a plate optimization objective be...

Currently, aeroplane images captured by camera sensors are characterized their small size and intricate backgrounds, posing a challenge for existing deep learning algorithms in effectively detecting targets. This paper incorporates the RFBNet (a coordinate attention mechanism) SIOU loss function into YOLOv5 algorithm to address this issue. The result is developing model undercarriage detection. primary goal synergize with algorithms, improving image capture precision. YOLOv5-RSC enhances...

Vibration is a major problem of switched reluctance motor drive system, which can cause undesirable acoustic noise. Acoustic noise severe when the periodic excitation machine normal force excites natural vibration mode stack. So, it very important for making frequency to avoid stator. In this paper, electromagnetic radial and modes are calculated respectively obtaining frequencies stator with windings based on 3D finite element. FFT used analyse dynamic response system excited by under...

An improved genetic algorithm based on J1 triangulation is proposed for multimodal optimization problems. And the fixed point theory introduced into this algorithm. The optimal problems are conversed to In paper, several typical functions used demonstrate effectiveness of algorithm, and testing results show that valid highly effective.

Undercarriage device is one of the essential parts an aeroplane, and accurate detection whether aeroplane undercarriage operating normally can effectively avoid accidents. To address problems low automation accuracy small target in existing methods, improved algorithm for YOLO V4 proposed. Firstly, convolutional network structure Inception-ResNet integrated into CSPDarkNet53 framework to improve algorithm’s ability extract semantic information features; then attention mechanism added path...

Abstract In the field of deep learning, point clouds are used as basic input format for 3D data, which can provide detailed geometric information about objects in original space. PointNet + is a learning network that uses cloud data an format, avoids losses associated with previous conversion into voxelization well collection 2D images. Although directly process various ways, due to disordered, irregular and unevenly distributed nature effect extracting features not ideal, large amount it...

<title>Abstract</title> This study innovatively improves the YOLOv8 target detection model, aiming to achieve fast and accurate of aircraft landing gear in natural environments. By introducing a small layer, dynamic serpentine convolutional CoTattention mechanism, successfully optimized original yolov8 model effectively detect small-sized gears when presented at distance. paper introduces layer 160x160 on top network, significantly improving performance airplane by fusing features from...

Abstract In deep learning, point clouds are used as the primary input format for 3D data, which can provide detailed geometric information about objects in original space. PointNet++ is a learning network that uses cloud data an format, avoids losses associated with previous conversion of into voxelization and collection 2D images. Although directly process various ways, due to disordered, irregular, unevenly distributed nature effect extracting features could be better. The large amount...

The mathematic model of a two-bar truss is built in MATLAB and the analysis carried out by genetic algorithm toolbox. In order to compare with each other, parametric planar also established ANSYS Parametric Design Language solutions are obtained using first-order method native ANSYS. comparison results shows that algorithms do not always display better properties than other for some problems. Finally, joint optimization which combines toolbox numerical based on quasi-Newton proposed. Then...

Aiming at the convergence precision defects of standard genetic algorithm, fixed point theory is introduced into algorithms. The population individual regarded as triangulation point. Hence vertex label information simplex would guide algorithm to optimization researching and judgment which could be calculated with hJ1 integer label. When loading simplexes individuals are transferred completely labeled simplexes, will terminated global optimal solution got. Finally, some functions used...

Stereo vision is increasingly applied to various fields of life, and three-dimensional (3D) reconstruction one the hottest topics in stereovision. Because active 3D system costly susceptible illumination, results have problems accuracy robustness, this paper proposes a passive method based on binocular stereo matching algorithm. precision sparse algorithm not high, most them are indoor scenes, so region-based dense adopted, which uses correlation degree gray information between windows...

With the emergence of new intelligent sensing technologies such as 3D scanners and stereo vision, high-quality point clouds have become very convenient lower cost. The research object recognition based on has also received widespread attention. Point are an important type geometric data structure. Because its irregular format, many researchers convert this into regular three-dimensional voxel grids or image collections. However, can lead to unnecessary bulk cause problems. In paper, we...

A virtual prototype model of a McPherson front suspension is built that used ADAMS/Car and based on multi-body system dynamics conducted the parallel wheel travel simulation. After sensitivity analysis design factor optimization multiple iterations using ADAMS/Insight, parameter has great influence for goal found under constraints finally. The problem positioning parameters better solved by optimizing parameters.

Establishing a 2-DOFs of 1/4 semi-active suspension system model, Using genetic algorithm approach, fuzzy logic controller the was designed, simulated in environment Matlab/Simulink. The results compared with passive demonstrate that this developed based on enhances performance full car significantly.

Piezoelectric materials are finding increasing applications in active vibration control of structures. A widely used modeling technique for piezoelectric flexible structures is the assumed modes method which results an infinite dimensional model composite structure. The paper builds and calculates modal frequency by application ANSYS software, extracts required according to reduction based on spatial H <inf xmlns:mml="http://www.w3.org/1998/Math/MathML"...

With the increase in tightening of carbon neutral policies, electric new energy vehicles have become a top priority development automotive industry. The integrated vehicle wheel (hub motor) can effectively optimize structure and reduce weight consumption. At same time, feature that four tires be driven separately is also conducive to application various control algorithms. In this paper, generalized adaptive PID experimented with speed hub motor, fuzzy control-based variable step method...