• Communications for Statistical Applications and Methods
• /
• v.24 no.4
• /
• pp.383-396
• /
• 2017

The model in our approach assumes that computer responses are a realization of a Gaussian processes superimposed on a regression model called a Gaussian process regression model (GPRM). Selecting a subset of variables or building a good reduced model in classical regression is an important process to identify variables influential to responses and for further analysis such as prediction or classification. One reason to select some variables in the prediction aspect is to prevent the over-fitting or under-fitting to data. The same reasoning and approach can be applicable to GPRM. However, only a few works on the variable selection in GPRM were done. In this paper, we propose a new algorithm to build a good prediction model among some GPRMs. It is a post-work of the algorithm that includes the Welch method suggested by previous researchers. The proposed algorithms select some non-zero regression coefficients (${\beta}^{\prime}s$) using forward and backward methods along with the Lasso guided approach. During this process, the fixed were covariance parameters (${\theta}^{\prime}s$) that were pre-selected by the Welch algorithm. We illustrated the superiority of our proposed models over the Welch method and non-selection models using four test functions and one real data example. Future extensions are also discussed.

• Proceedings of the KSR Conference
• /
• 1998.05a
• /
• pp.385-392
• /
• 1998

Train crashes involve complex interaction between deformable bodies in multiple collisions. The purpose of this study is to suggest the effective analytical procedure using simple model for the crashworthiness of motorized trailer of high speed train. The simple model, with very short modeling time and reduced computation time was adopted to extract the global behaviour and to perform a pre-optimization of the considered structure. Firstly, various types of crash events are investigated and the conditions for numerical simulation are defined. The simple model, using the beam and non-linear spring element, and shell element model are used to evaluated energy absorption and deformation mechanism in analyses. And aluminum is applied to real model after verification with square tube analyses. Finally, loading path and energy absorption of main components are evaluated. The analyses are aimed to ensure the crashworthiness design of high speed train.

• Structural Engineering and Mechanics
• /
• v.24 no.2
• /
• pp.227-245
• /
• 2006

The performance of a three dimensional non-linear finite element model for hyperelastic material considering the effect of compressibility is studied by analyzing rubber blocks under different modes of deformation. It includes simple tension, pure shear, simple shear, pure bending and a mixed mode combining compression, shear and bending. The compressibility of the hyperelastic material is represented in the strain energy function. The nonlinear formulation is based on updated Lagrangian (UL) technique. The displacement model is implemented with a twenty node brick element having u, ${\nu}$ and w as the degrees of freedom at each node. The results obtained by the present numerical model are compared with the analytical solutions available for the basic modes of deformation where the agreement between the results is found to be satisfactory. In this context some new results are generated for future references since the number of available results on the present problem is not sufficient enough.

• Advances in aircraft and spacecraft science
• /
• v.10 no.6
• /
• pp.521-543
• /
• 2023

Reynolds-averaged Navier-Stokes (RANS) models are extensively employed in industrial settings for the purpose of simulating intricate fluid flows. However, these models are subject to certain limitations. Notably, disparities persist in the Reynolds stresses when comparing the RANS model with high-fidelity data obtained from Direct Numerical Simulation (DNS) or experimental measurements. In this work we propose an approach to mitigate these discrepancies while retaining the favorable attributes of the Menter Shear Stress Transport (SST) model, such as its significantly lower computational expense compared to DNS simulations. This strategy entails incorporating an explicit algebraic model and employing a neural network to correct the turbulent characteristic time. The imposition of realizability constraints is investigated through the introduction of penalization terms. The assimilated Reynolds stress model demonstrates good predictive performance in both in-sample and out-of-sample flow configurations. This suggests that the model can effectively capture the turbulent characteristics of the flow and produce physically realistic predictions.

• Journal of Electrical Engineering and Technology
• /
• v.13 no.1
• /
• pp.418-429
• /
• 2018

This paper studies the decentralized position/force control problem for constrained reconfigurable manipulator without torque sensing. A novel joint torque estimation scheme that exploits the existing structural elasticity of the manipulator joint with harmonic drive model is applied for each joint module. Based on the estimated joint torque and dynamic output feedback technique, a decentralized position/force control strategy is presented. In order to solve the problem of controller parameter perturbation, the non-fragile robust technique is introduced into the dynamic output feedback controller. Subsequently, the stability of the closed-loop system is proved using the Lyapunov theory and linear matrix inequality (LMI) technique. Finally, two 2-DOF constrained reconfigurable manipulators with different configurations are applied to verify the effectiveness of the proposed control scheme in numerical simulation.

• Journal of Korea Water Resources Association
• /
• v.48 no.12
• /
• pp.1023-1035
• /
• 2015

This study applies 3-D N-S solver based on PBM (Porous Body Model), LED-WASS-3D ver 2.0 to directly analyze non-linear interaction of seawater-freshwater-coastal aquifer in order to simulate the seawater infiltration into coastal aquifer. This numerical simulation is the first trial in Korea, as well as unusual and new numerical analysis abroad. Firstly, to validate the applied numerical model, the validity and effectiveness was verified for the numerical model by comparing and considering it with the result of laboratory experiment for seawater-freshwater interface in coastal aquifer. And then it simulated the seawater infiltration into coastal aquifer considering the changed levels of seawater and groundwater in order to analyze the distribution characteristics of flow field and seawater-freshwater interface of coastal aquifer as the level difference between seawater and groundwater and rate of seawater level (${\Delta}h/h$) increased. In addition, the characteristics of seawater infiltration were analyzed from the vertical salinity in the coastal aquifer by ${\Delta}h/h$, which cannot be obtained from existing non-diffusion numerical models. Finally, it analyzed the effect of ${\Delta}h/h$ on the seawater infiltration distance in coastal aquifer, which was indexed.

• Earthquakes and Structures
• /
• v.6 no.1
• /
• pp.71-87
• /
• 2014

Interaction between closely-spaced buildings subject to earthquake induced strong ground motions, termed in the literature as "seismic pounding", occurs commonly during major seismic events in contemporary congested urban environments. Seismic pounding is not taken into account by current codes of practice and is rarely considered in practice at the design stage of new buildings constructed "in contact" with existing ones. Thus far, limited research work has been devoted to quantify the influence of slab-to-slab pounding on the inelastic seismic demands at critical locations of structural members in adjacent structures that are not aligned in series. In this respect, this paper considers a typical case study of a "new" reinforced concrete (R/C) EC8-compliant, torsionally sensitive, 7-story corner building constructed within a block, in bi-lateral contact with two existing R/C 5-story structures with same height floors. A non-linear local plasticity numerical model is developed and a series of non-linear time-history analyses is undertaken considering the corner building "in isolation" from the existing ones (no-pounding case), and in combination with the existing ones (pounding case). Numerical results are reported in terms of averages of ratios of peak inelastic rotation demands at all structural elements (beams, columns, shear walls) at each storey. It is shown that seismic pounding reduces on average the inelastic demands of the structural members at the lower floors of the 7-story building. However, the discrepancy in structural response of the entire block due to torsion-induced, bi-directionally seismic pounding is substantial as a result of the complex nonlinear dynamics of the coupled building block system.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.40 no.4
• /
• pp.353-362
• /
• 2016

An accurate hysteresis model of an elastomer is important for quantifying viscoelastic energy loss. We suggest a highly nonlinear hyper-viscoelastic constitutive model of elastomers. The model captures a nonlinear viscoelastic characteristic by combining Yeoh's hyperelastic model and Hoofatt's hysteresis model used Neo-Hookean hyperelastic model. Analytical and numerical models were generated from uniaxial cyclic tests of an elastomer under a sinusoidal load with a mean strain of 150%, amplitudes of 20~80%, and frequencies of 0.02~0.2Hz. The viscoelastic model can highly capture the viscoelastic energy loss up to a strain of 230%.

• Structural Engineering and Mechanics
• /
• v.26 no.4
• /
• pp.409-426
• /
• 2007

Following previous work carried out in Building Research Institute in Japan, finite element analyses of conceptual column designs are performed in this paper. The effectiveness of the numerical model is evaluated by experimental tests and parametric studies are conducted to determine influential factors in conceptual column designs. First, three different column designs are analysed: bonded, un-bonded, and un-bonded with additional reinforcing bars. The load-displacement curves and cracking patterns in concrete are obtained and compared with experimental ones. The comparisons indicate that the finite element model is able to reflect the experimental results closely. Both numerical and experimental results show that, the introduction of un-bonded zones in a column end can reduce cracking strains, accordingly reduce the stiffness and strength as well; the addition of extra reinforcement in the un-bonded zones can offset the losses of the stiffness and strength. To decide the proper length of the un-bonded zones and the sufficient amount of the additional reinforcing bars, parametric studies are carried out on their influences. It has been found that the stiffness of un-bonded designs slightly decreases with increasing the length of the un-bonded zones and increases with the size of the additional reinforcing bars.

• Journal of Korean Society of Transportation
• /
• v.9 no.2
• /
• pp.121-126
• /
• 1991

Wilson made an importent contribution to develop a trip distribution model with the general form of gravity model which is an entropy maximization program. Also Wilson suggested a technique which is called the "iterative balancing method" for soving the model. This te-chnique however is not stable to find solution because it is a heuristic method and sometimes does not converge to the correct solution. In this paper a new solution method using a numerical method for solving the non-linear simultaneous equation system is developed and evaluated in both computers VAX 8700 and PC/AT 286 The result of this method and Wilson's method are compared with each other. Wilson's method resulted in inferior solutions measured by the final norm of residuals.