• Journal of Korean Institute of Industrial Engineers
• /
• v.42 no.4
• /
• pp.263-269
• /
• 2016

The purpose of this study is to challenge a computational regression-type problem, that is handling large-size data, in which conventional metamodeling techniques often fail in a practical sense. To solve such problems, regression-type boosting, one of ensemble model techniques, together with bootstrapping-based re-sampling is a reasonable choice. This study suggests weight updates by the amount of the residual itself and a new error decision criterion which constructs an ensemble model of models selectively chosen by rejection limits. Through these ideas, we propose AdaBoost.RMU.R as a metamodeling technique suitable for handling large-size data. To assess the performance of the proposed method in comparison to some existing methods, we used 6 mathematical problems. For each problem, we computed the average and the standard deviation of residuals between real response values and predicted response values. Results revealed that the average and the standard deviation of AdaBoost.RMU.R were improved than those of other algorithms.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.34 no.E
• /
• pp.30-44
• /
• 1992

A modification of the TEACH-like computer program based on the k-$\varepsilon$ turbulence transport was applied for predicting air mixing patterns and temperature distributions in a rectangular, slot-ventilated enclosure having obstructions ; a rectangular obstacle with heat flux, solid walls separates the passage and the pig pens, and purlins beneath the ceiling. Air flow patterns were calculated for the cases with and without the purlin, extending 300mm beneath the ceiling. Comparisons of prediction data of Randall & Battams(1976) showed air flow pattern predicted well for the case without the purlin. Heat was accumulated at the corner of the left side of the solid wall and the right-upper region of the simulated pigs. However the air distribution pattern was completely different from data for the case with the purlin. The deviation from the observation may be attributed to the difference of the geometric configuation. Exploring the cause of the deviation should be conducted in a further study. Temperature stratification was also observed due to incomplete mixing. The obstruction in the route of the inlet air jet at inlet should be avoided since most of kinetic energy dissipates at the abstacle duet to impingement.

• Nuclear Engineering and Technology
• /
• v.52 no.12
• /
• pp.2901-2909
• /
• 2020

Thermal creep is a key property of zircaloy cladding. CZ developed by CGN is a new zircaloy used as PWR fuel cladding. This research is devoted to investigating the thermal creep behavior of CZ and build the thermal creep model of CZ. Twenty internal pressure creep tests were conducted, and the ranges of temperature and Tresca stress were 320-430 ℃ and 70-300 MPa, respectively. Real-time creep data were analyzed by separating primary creep and steady-state creep. Based on Soderberg model and creep test data, CZ thermal creep model is derived. As a whole, the mean value and the standard deviation of P/M of CZ saturated primary creep strain are very close to these from steady-state creep rate, however, the predictive effect of primary creep is less satisfactory. Four conditions, where there exists large deviation between predicted values and test data, are 320 ℃ and 300 MPa, 350 ℃ and 190 MPa, 380 ℃ and 160 MPa, 380 ℃ and 190 MPa, respectively. As primary creep was much smaller than steady-state creep in long-time operation, the thermal creep model built can be applied to predict the thermal creep behavior of CZ cladding.

• Nuclear Engineering and Technology
• /
• v.41 no.7
• /
• pp.875-884
• /
• 2009

We present three nuclear/hydrogen-related R&D activities being performed at KAIST: air-ingressed LOCA analysis code development, gas turbine analysis tool development, and hydrogen-production system analysis model development. The ICE numerical technique widely used for the safety analysis of water-reactors is successfully implemented into GAMMA, with which we solve the basic equations for continuity, momentum conservation, energy conservation of the gas mixture, and mass conservation of 6 species (He, N2, O2, CO, CO2, and H2O). GAMMA has been extensively validated using data from 14 test facilities. We developed a tool to predict the characteristics of HTGR helium turbines based on the throughflow calculation with a Newton-Raphson method that overcomes the weakness of the conventional method based on the successive iteration scheme. It is found that the current method reaches stable and quick convergence even under the off-normal condition with the same degree of accuracy. The dynamic equations for the distillation column of HI process are described with 4 material components involved in the HI process: H2O, HI, I2, H2. For the HI process we improved the Neumann model based on the NRTL (Non-Random Two-Liquid) model. The improved Neumann model predicted a total pressure with 8.6% maximum relative deviation from the data and 2.5% mean relative deviation, and liquid-liquid-separation with 9.52% maximum relative deviation from the data.

• Computers and Concrete
• /
• v.23 no.6
• /
• pp.433-444
• /
• 2019

As rice husk ash (RHA) is not produced in controlled manufacturing process like cement, its properties vary significantly even within the same lot. In fact, properties of Rice Husk Ash Based Concrete (RHABC) are largely dictated by uncertainty leading to huge deviations from their expected values. This paper proposes a Robust Cost Optimization (RCO) procedure for RHABC, which minimizes such unwanted deviation due to uncertainty and provides guarantee of achieving desired strength and workability with least possible cost. The RCO simultaneously minimizes cost of RHABC production and its deviation considering feasibility of attaining desired strength and workability in presence of uncertainty. RHA related properties have been modeled as uncertain-but-bounded type as associated probability density function is not available. Metamodeling technique is adopted in this work for generating explicit expressions of constraint functions required for formulation of RCO. In doing so, the Moving Least Squares Method is explored in place of conventional Least Square Method (LSM) to ensure accuracy of the RCO. The efficiency by the proposed MLSM based RCO is validated by experimental studies. The error by the LSM and accuracy by the MLSM predictions are clearly envisaged from the test results. The experimental results show good agreement with the proposed MLSM based RCO predicted mix properties. The present RCO procedure yields RHABC mixes which is almost insensitive to uncertainty (i.e., robust solution) with nominal deviation from experimental mean values. At the same time, desired reliability of satisfying the constraints is achieved with marginal increment in cost.

• Bulletin of the Korean Chemical Society
• /
• v.26 no.11
• /
• pp.1783-1789
• /
• 2005

Modeling of a steam generator crevice in a nuclear power system needs to take into account both thermalhydraulic and chemical phenomena. As a first step towards developing a reliable model, a chemical equilibrium model was developed to predict chemical speciation in a magnetite-packed crevice by adopting the “tableau” method. The model was benchmarked with the available experimental data and the maximum deviation did not exceed two orders of magnitude. The developed model was applied to predict the chemical speciation in a magnetite-packed crevice. It was predicted that caustic environment was developed by the concentration of NaOH and the dissolution of magnetite. The model indicated that the dominant aqueous species of iron in the caustic crevice was $FeO_2\;^-$. The increase of electrochemical corrosion potential observed in the experiment was rationalized by the decrease of dissolved hydrogen concentration due to a boiling process. It was predicted that under the deaerated condition magnetite was oxidized to hematite.

• Fire Science and Engineering
• /
• v.31 no.3
• /
• pp.19-24
• /
• 2017

Flammable substances are used in laboratories and industrial process. The flash point (FP) is one of the most important physical properties used to determine the potential for characterizing the fire and explosion hazard of liquids. The FP data at 101.3 kPa were measured for the binary systems {toluene+ethylbenzene}, {methlycyclohenxane+ethylbenzene} and {n-heptane+ ethylbenzene}. The experiments were performed according to the standard test method (ASTM D 3278) using a SETA closed cup flash point tester. The measured FPs were compared with the values predicted using the following activity coefficient models: Wilson, Non-Random Two Liquid (NRTL), and UNIversal QUAsiChemical (UNIQUAC). The average absolute deviation between the predicted and measured lower FP was less than 1.74 K.

• International Journal of Concrete Structures and Materials
• /
• v.1 no.1
• /
• pp.63-73
• /
• 2007

Optimum multi-layered feed-forward neural network (NN) models using a resilient back-propagation algorithm and early stopping technique are built to predict the shear capacity of reinforced concrete deep and slender beams. The input layer neurons represent geometrical and material properties of reinforced concrete beams and the output layer produces the beam shear capacity. Training, validation and testing of the developed neural network have been achieved using 50%, 25%, and 25%, respectively, of a comprehensive database compiled from 631 deep and 549 slender beam specimens. The predictions obtained from the developed neural network models are in much better agreement with test results than those determined from shear provisions of different codes, such as KBCS, ACI 318-05, and EC2. The mean and standard deviation of the ratio between predicted using the neural network models and measured shear capacities are 1.02 and 0.18, respectively, for deep beams, and 1.04 and 0.17, respectively, for slender beams. In addition, the influence of different parameters on the shear capacity of reinforced concrete beams predicted by the developed neural network shows consistent agreement with those experimentally observed.

• The KSFM Journal of Fluid Machinery
• /
• v.3 no.3 s.8
• /
• pp.31-40
• /
• 2000

A computer program is developed for the prediction of the aerodynamic performance and the noise characteristics in the basic design step of axial flow fan. The flow field and the performance of fan are analyzed by using the streamline curvature computing scheme with total pressure loss and flow deviation models. Fan noise is assumed to be generated due to the pressure fluctuations induced by wake vortices of fan blades and to radiate via dipole distribution. The vortex-induced fluctuating pressure on blade surface is calculated by combining thin airfoil theory and the predicted flow field data. The predicted performances, sound pressure level and noise directivity patterns of fan by the present method are favorably compared with the test data of actual fan. Furthermore, the present method is shown to be very useful in optimizing design variables of fan with high efficiency and low noise level.

• 유체기계공업학회:학술대회논문집
• /
• 2001.11a
• /
• pp.37-42
• /
• 2001

A computerized axial flow fan design system is developed with the capabilities for predicting the aerodynamic performance and the noise characteristics of fan. In the present study, the basic fan blading design is made by combining vortex distribution scheme with camber line design, airfoil selection, blade thickness distribution and stacking of blade elements. With the designed fan blade geometry, the through-flow field and the performance of fan are analyzed by using the streamline curvature computing scheme with spanwise total pressure loss and flow deviation models. Fan noise is assumed to be generated due to the pressure fluctuation induced by wake vortices of fan blades and to radiate as dipole distribution. The vortex-induced fluctuating pressure on blade surface is calculated by combining thin airfoil theory and the predicted flow field data. The predicted performances, sound pressure level and noise directivity patterns of fan by the present method are favorably compared with the test data of actual fans. Furthermore, the present method is shown to be very useful in designing the blade geometry of new fan and optimizing design variables of the fan to achieve higher efficiency and lower noise level.