• ETRI Journal
• /
• v.44 no.3
• /
• pp.476-490
• /
• 2022

With the recent advances in technology, automatic speech recognition (ASR) has been widely used in real-world applications. The efficiency of converting large amounts of speech into text accurately with limited resources has become more vital than ever. In this study, we propose a method to rapidly recognize a large speech database via a transformer-based end-to-end model. Transformers have improved the state-of-the-art performance in many fields. However, they are not easy to use for long sequences. In this study, various techniques to accelerate the recognition of real-world speeches are proposed and tested, including decoding via multiple-utterance-batched beam search, detecting end of speech based on a connectionist temporal classification (CTC), restricting the CTC-prefix score, and splitting long speeches into short segments. Experiments are conducted with the Librispeech dataset and the real-world Korean ASR tasks to verify the proposed methods. From the experiments, the proposed system can convert 8 h of speeches spoken at real-world meetings into text in less than 3 min with a 10.73% character error rate, which is 27.1% relatively lower than that of conventional systems.

• Geomechanics and Engineering
• /
• v.29 no.3
• /
• pp.229-236
• /
• 2022

In this study, a model experiment and field experiment was conducted to introduce the optimal tensile force when constructing a non-open cut tunnel according to the ground conditions of sandy soil. CMR (Concrete Modular Roof) method is economical because of the high precision and excellent durability, and corrosion resistance, and the inserted parts can be used as the main structure of a tunnel. In addition the CMR method has a stable advantage in interconnection because the concrete beam is press-fitted compared to the NTR (New Tubular Roof) method, and the need for quality control can be minimized. The ground conditions were corrected by adjusting the relative density of sandy soil during the construction of non-open cut tunnels, and after introducing various tensile forces, the surface settlement according to excavation was measured, and the optimal tensile force was derived. As a result of the experiment, the amount of settlement according to the relative density was found to be minor. Furthermore, analysis of each tensile force based on loose ground conditions resulted in an average decrease of approximately 22% in maximum settlement when the force was increased by 0.8 kN per segment. Considering these results, it is indicated that more than 2.0 kN tensile force per segment is recommended for settlement of the upper ground.

• Smart Structures and Systems
• /
• v.28 no.6
• /
• pp.827-838
• /
• 2021

The dynamic displacement is considered to be an important indicator of structural safety, and becomes an indispensable part of Structural Health Monitoring (SHM) system for high-speed railway bridges. This paper proposes an indirect strain based dynamic displacement reconstruction methodology for high-speed railway box girders. For the typical box girders under eccentric train load, the plane section assumption and elementary beam theory is no longer applicable due to the bend-torsion coupling effects. The monitored strain was decoupled into bend and torsion induced strain, pre-trained multi-output support vector regression (M-SVR) model was employed for such decoupling process considering the sensor layout cost and reconstruction accuracy. The decoupled strained based displacement could be reconstructed respectively using box girder plate element analysis and mode superposition principle. For the transformation modal matrix has a significant impact on the reconstructed displacement accuracy, the modal order would be optimized using particle swarm algorithm (PSO), aiming to minimize the ill conditioned degree of transformation modal matrix and the displacement reconstruction error. Numerical simulation and dynamic load testing results show that the reconstructed displacement was in good agreement with the simulated or measured results, which verifies the validity and accuracy of the algorithm proposed in this paper.

• Steel and Composite Structures
• /
• v.41 no.6
• /
• pp.787-803
• /
• 2021

The size-dependent nonlinear thermomechanical vibration analysis of pre- and post-buckled tapered two-directional functionally graded (2D-FG) microbeams is presented in this study. In the context of the modified couple stress theory, the formulations are derived based on the parabolic shear deformation beam theory and von Karman nonlinear strains. Different thermomechanical material properties are assumed to be temperature-dependent and smoothly vary in both length and thickness directions using the power law and the physical neutral axis concept is employed. The nonlinear governing equations are derived using the Hamilton principle and the resulting variable coefficient equations of motion are solved using the differential quadrature method (DQM) and iterative Newton's method for clamped-clamped and simply supported boundary conditions. Comparison studies are presented to validate the derived model and solution procedure. The impacts of induced thermal moments, temperature power index, two gradient indices, nonuniform cross-section, and microstructure length scale parameter on the frequency-temperature configurations are explored for both clamped and simply supported microbeams.

• Computers and Concrete
• /
• v.30 no.6
• /
• pp.433-443
• /
• 2022

In the current paper, the nonlinear resonance response of functionally graded graphene platelet reinforced (FG-GPLRC) beams by considering different boundary conditions is investigated using the Euler-Bernoulli beam theory. Four different graphene platelets (GPLs) distributions including UD and FG-O, FG-X, and FG-A are considered and the effective material parameters are calculated by Halpin-Tsai model. The nonlinear vibration equations are derived by Euler-Lagrange principle. Then the perturbation method is used to discretize the motion equations, and the loadings and displacement are all expanded, so as to obtain the first to third order perturbation equations, and then the asymptotic solution of the equations can be obtained. Then the nonlinear amplitude-frequency response is obtained with the help of the modified Lindstedt-Poincare method (Chen and Cheung 1996). Finally, the influences of the distribution types of GPLs, total GPLs layers, GPLs weight fraction, elastic foundations and boundary conditions on the resonance problems are comprehensively studied. Results show that the distribution types of GPLs, total GPLs layers, GPLs weight fraction, elastic foundations and boundary conditions have a significant effect on the nonlinear resonance response of FG-GPLRC beams.

• Structural Engineering and Mechanics
• /
• v.81 no.6
• /
• pp.751-767
• /
• 2022

In the hogging bending moment area, continuous composite beams are subjected to the ultimate limit state of lateral-torsional buckling (LTB), which depends on web stiffness as well as concrete slab and shear connection stiffnesses. The design of the LTB and the determination of the elastic critical moment are produced approximately, using the European Standard EN 1994-1-1:2004, for continuous composite steel beams, but is applicable only for those with a plane web steel profile. Also, and from the previous researches, the elastic critical moment of the continuous composite beams with corrugated sinusoidal web steel profiles was determined. In this paper, a finite element analysis (FEA) model was developed using the ANSYS 16 software, to determine the elastic critical moments of continuous composite steel beams with various corrugated web profiles, such as trapezoidal, zigzag, and rectangular profiles, which were evaluated against numerical data of the sinusoidal one from the literature. Ultimately, the failure load of a composite steel beam with various web profiles was predicted by studying 46 models, based on FEA modeling, and a procedure for predicting the elastic critical moment of composite beams with various web steel profiles was proposed. When compared to sinusoidal web profiles, the trapezoidal, zigzag, and rectangular web profiles required an average increase in load capacity and stiffness of 7%, 17.5%, and 28%, respectively, according to the finite element analysis. Also, the rectangular web steel profile has a greater stiffness and load capacity. In contrast, the sinusoidal web has lower values for these characteristics.

• Structural Engineering and Mechanics
• /
• v.82 no.1
• /
• pp.31-40
• /
• 2022

In structural engineering, the material properties of the structures such as elastic modulus, shear modulus, density, and size may not be deterministic and may vary at different locations. The dynamic response analysis of such structures may need to consider these properties as stochastic. This paper introduces a stochastic finite element method (SFEM) approach to analyze moving loads problems. Firstly, Karhunen-Loéve expansion (KLE) is applied for expressing the stochastic field of material properties. Then the mathematical expression of the random field is substituted into the finite element model to formulate the corresponding random matrix. Finally, the statistical moment of the dynamic response is calculated by the point estimation method (PEM). The accuracy and efficiency of the dynamic response obtained from the KLE-PEM are demonstrated by the example of a moving load passing through a simply supported Euler-Bernoulli beam, in which the material properties (including elastic modulus and density) are considered as random fields. The results from the KLE-PEM are compared with those from the Monte Carlo simulation. The results demonstrate that the proposed method of KLE-PEM has high accuracy and efficiency. By using the proposed SFEM, the random vertical deflection of a high-speed railway (HSR) bridge is analyzed by considering the random fields of material properties under the moving load of a train.

• Advances in nano research
• /
• v.13 no.2
• /
• pp.165-174
• /
• 2022

In the present study, a hybrid model of artificial neural network (ANN) and internet of things (IoT) is proposed to overcome the difficulties in deriving governing equations and numerical solutions of the dynamical behavior of the nano-systems. Nano-structures manifest size-dependent behavior in response to static and dynamic loadings. Nonlocal and length-scale parameters alongside with other geometrical, loading and material parameters are taken as input parameters of an ANN to observe the natural frequency and damping behavior of micro sensors made from nanocomposite material with piezoelectric layers. The behavior of a micro-beam is simulated using famous numerical methods in literature under base vibrations. The ANN was further trained to correlate the output vibrations to the base vibration. Afterwards, using IoT, the electrical potential conducted in the sensors are collected and converted to numerical data in an embedded mini-computer and transferred to a server for further calculations and decision by ANN. The ANN calculates the base vibration behavior with is crucial in mechanical systems. The speed and accuracy of the ANN in determining base excitation behavior are the strengths of this network which could be further employed by engineers and scientists.

• Current Optics and Photonics
• /
• v.6 no.1
• /
• pp.44-50
• /
• 2022

An off-axis zone plate is able to focus on a single order while neglecting the zeroth order in a visible imaging system. This allows one to enhance the contrast quality in diffractive images, which is the major advantage of this type of zone plate. However, most previous reflection zone plates are used in focusing X-rays with a small grazing incident angle and are intricately designed with the use of a local grating period. In this study, we suggest the design of an off-axis elliptical zone plate (EZP) that is used to focus a monochromatic light beam with separation between the first and unfocused orders under a large grazing incident angle of 45°. An assumption using the total grating period, which depends on the average and constant grating period, is proposed to calculate the desired distance between the first and zeroth order and to simplify the construction of a novel model off-center EZP. Four diffractive optical elements (DOEs) with different parameters were subsequently fabricated by direct laser lithography and then verified using a performance evaluation system to compare the results from the assumption with the experimental results.

• Advances in aircraft and spacecraft science
• /
• v.9 no.5
• /
• pp.377-400
• /
• 2022

The aim of this work is a numerical comparison (FEM) between lattice pyramidal-core panel and honeycomb core panel for different core thicknesses. By evaluating the mid-span deflection, the shear rigidity and the shear modulus for both core types and different core thicknesses, it is possible to define which core type has got the best mechanical behaviour for each thickness and the evolution of that behaviour as far as the thickness increases. Since a specific base geometry has been used for the lattice pyramidal core, the comparison gives us the opportunity to investigate the unit cell strut angle giving the higher mechanical properties. The presented work considers a detailed FEM modelling of a standard 3-point bending test (ASTM C393/C393M Standard Practice). Detailed FEM modelling addresses to detailed discretization of cores by means of beam elements for lattice core and shell elements for honeycomb core. Facings, instead, have been modelled by using shell elements for both sandwich panels. On lattice core structure, elements of core and facings are directly connected, to better simulate the additive manufacturing process. Otherwise, an MPC-based constraint between facings and core has been used for honeycomb core structure. Both sandwich panels are entirely built of Aluminium alloy. Prior to compare the two models, the FEM sandwich panel model with lattice pyramidal core needs to be validated with 3-point bending test experimental results, in order to ensure a good reliability of the FEM approach and of the comparison. Furthermore, the analytical validation has been performed according to Allen's theory. The FEM analysis is linear static with an increasing midspan load ranging from 50N up to 500N.