• Design & Manufacturing
• /
• v.17 no.4
• /
• pp.72-78
• /
• 2023

Injection molding is a process widely used in various industries because of its high production speed and ease of mass production during the plastic manufacturing process, and the product is molded by injecting molten plastic into the mold at high speed and pressure. Since process conditions such as resin and mold temperature mutually affect the process and the quality of the molded product, it is difficult to accurately predict quality through mathematical or statistical methods. Recently, studies to predict the quality of injection molded products by applying artificial neural networks, which are known to be very useful for analyzing nonlinear types of problems, are actively underway. In this study, structural optimization of neural networks was conducted by applying multi-task learning techniques according to the characteristics of the input and output parameters of the artificial neural network. A structure reflecting the characteristics of each process step was applied to the input parameters, and a structure reflecting the quality characteristics of the injection molded part was applied to the output parameters using multi-tasking learning. Building an artificial neural network to predict the three qualities (mass, diameter, height) of injection-molded product under six process conditions (melt temperature, mold temperature, injection speed, packing pressure, pacing time, cooling time) and comparing its performance with the existing neural network, we observed enhancements in prediction accuracy for mass, diameter, and height by approximately 69.38%, 24.87%, and 39.87%, respectively.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.6
• /
• pp.2689-2708
• /
• 2016

Multi-camera systems which integrate two or more low-cost digital cameras are adopted to reach higher ground coverage and improve the base-height ratio in low altitude remote sensing. To guarantee accurate multi-camera integration, the geometric relationship among cameras must be determined through platform calibration techniques. This paper proposed a combined two-step platform calibration method. In the first step, the static platform calibration was conducted based on the stable relative orientation constraint and convergent conditions among cameras in static environments. In the second step, a dynamic platform self-calibration approach was proposed based on not only tie points but also straight lines in order to correct the small change of the relative relationship among cameras during dynamic flight. Experiments based on the proposed two-step platform calibration method were carried out with terrestrial and aerial images from a multi-camera system combined with four consumer-grade digital cameras onboard an unmanned aerial vehicle. The experimental results have shown that the proposed platform calibration approach is able to compensate the varied relative relationship during flight, acquiring the mosaicing accuracy of virtual images smaller than 0.5pixel. The proposed approach can be extended for calibrating other low-cost multi-camera system without rigorously mechanical structure.

• Journal of the Korean Society of Visualization
• /
• v.12 no.3
• /
• pp.15-20
• /
• 2014

Direct numerical simulation (DNS) dataset of a turbulent boundary layer (TBL) with a step change from smooth to rough surface is analyzed to examine spatially developing flow characteristics. The roughness elements are periodically arranged two-dimensional (2-D) spanwise rods with a streamwise pitch of ${\lambda}=8k$ ($=12{\theta}_{in}$), and the roughness height is $k=15{\theta}_{in}$, where ${\theta}_{in}$ is the inlet momentum thickness. The step change is introduced $80{\theta}_{in}$ downstream from the inlet. For the first time, full images from the DNS data with the step change from the smooth to rough walls is present to get some idea of the geometry of turbulent coherent structures over rough wall, especially focusing on their existence and partial dynamics over the rough wall. The results show predominance of hairpin vortices over the rough wall and their spanwise scale growth mechanism by merging.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.20 no.5
• /
• pp.1661-1670
• /
• 1996

Incompressible flow over a backward-facing step is computed by low Reynolds number turbulence models in order to compare with direct simulation results. In this study, selected low Reynolds number 1st and 2nd (Algebraic Stress Model : ASM) moment closure turbulence models are adopted and compared with each other. Each turbulence model predicts different flow characteristics, different re-attachment point, velocity profiles and Reynolds stress distribution etc. Results by .kappa.-.epsilon. turbulence models indicate that predicted re-attachment lengths are shorter than those by standard model. Turbulent intensity and eddy viscosity by low Reynolds number .kappa.-.epsilon. models are still greater than DNS results. The results by algebraic stress model (ASM) are more reasonable than those by .kappa.-.epsilon. models. The convective scheme is QUICK (Quadratic Upstream Interpolation for Convective Kinematics) and SIMPLE algorithm is adopted. Reynolds number based on step height and inlet free stream velocity is 5100.

• Atmosphere
• /
• v.25 no.3
• /
• pp.473-482
• /
• 2015

In this study, street canyons with a higher downwind building (so called, step-up street canyons) are considered for understanding characteristics of flow and reactive pollutants' dispersion as a basic step to understand the characteristics in wider urban areas. This study used a CFD_NIMR_SNU coupled to a chemistry module just including simple $NO_X-O_3$ photochemical reactions. First, flow characteristics are analyzed in step-up street canyons with four aspect ratios (0.33, 0.47, 0.6, 0.73) defined as ratios of upwind building heights to downwind building height. The CFD_NIMR_SNU reproduced very well the main features (that is, vortices in the street canyons) which appeared in the wind-tunnel experiment. Wind speed within the street canyons became weak as the aspect ratio increased, because volume of flow incoming over the upwind building decreased. For each step-up street canyon, chemistry transport model was integrated up to 3600 s with the time step of 0.5 s. The distribution patterns of $NO_X$ and $O_3$ were largely dependent on the mean flow patterns, however, $NO_X$ and $O_3$ concentrations were partly affected by photochemical reactions. $O_3$ concentration near the upwind lower region of the street canyons was much lower than background concentration, because there was much reduction in $O_3$ concentration due to NO titration there. Total amount of $NO_X$ in the street canyons increased with the aspect ratio, resulting from the decrease of mean wind intensity.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.5
• /
• pp.45-54
• /
• 2009

In this paper, we propose an adult images detection algorithm using a mean intensity filter and an improved 2D Hough Transform. This paper is composed of three major steps including a training step, a recognition step, and a verification step. The training step generates a mean nipple variance filter that will be used for detecting nipple candidate regions in the recognition step. To make the mean variance filter, we converts an input color image into a gray scale image and normalize it, and make an average intensity filter for nipple areas. The recognition step first extracts edge images and finds connected components, and decides nipple candidate regions by considering the ratio of width and height of a connected component. It then decides final nipple candidates by calculating the similarity between the learned nipple average intensity filter and the nipple candidate areas. Also, it detects breast lines of an input image through the improved 2D Hough transform. The verification step detects breast areas and identifies adult images by considering the relations between nipple candidate regions and locations of breast lines.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.17 no.5
• /
• pp.1253-1261
• /
• 1993

Numerical analysis and measurements of the velocity and temperature distributions in buoyancy assisting laminar mixed convection flow over a vertically located, two-dimensional backward-facing step are reported. Laser-Doppler Velocimeter and Constant Temperature Anemometer operated in constant current were used to measure simultaneously the velocity and temperature distributions in the recirculation region downstream of the step. The reattachment length was measured by using flow visualization technique for different inlet velocities, wall temperatures and step heights. While the reattachment length $X_r$ increases as the inlet velocity or step height increase, it decreases as the buoyancy force increases, causing the size of the recirculation region to decrease. For the experimental range of $Gr_s$/$Re_{s}^{2}$$\times$$10^3$<17, a correlation equation for the reattachment length can be given by $X_{r}=1.05(2.13+0.021 Re_{s})exp$ $(-33.7_s^{-0.186}/Gr_{s}/Re_{s}^2).$ The Nusselt number is found to increase and the location of its maximum value moves closer to the step as the buoyancy force increases. The location of the maximum Nusselt number occurs downstream of the reattachment point, and distance between the reattachment point and the location of the maximum Nusselt mumber increases as the buoyancy force increases. Computational prediction agrees favorably well with measured results.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.8 no.8
• /
• pp.1179-1186
• /
• 2013

In this paper, an efficient vehicle detection and tracking algorithm for detection incident in tunnel is proposed. The proposed algorithm consists of three steps. The first one is a step for background estimates, low computational complexity and memory consumption Running Gaussian Average (RGA) is used. The second step is vehicle detection step, Adaboost algorithm is applied to this step. In order to reduce false detection from a relatively remote location of the vehicles, local features according to height of vehicles are used to detect vehicles. If the local features of an object are more than the threshold value, the object is classified as a vehicle. The last step is a vehicle tracking step, the Kalman filter is applied to track moving objects. Through computer simulations, the proposed algorithm was found that useful to detect and track vehicles in the tunnel.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.26 no.3
• /
• pp.423-431
• /
• 2002

An experimental study is conducted to investigate the effect of two-frequency forcing on turbulent flow behind a backward-facing step at the Reynolds number of 27000 based on the step height. The forcing is provided from a thin slit located at the edge of the backward-facing step to increase mixing behind the backward-facing step and consequently to reduce the reattachment length. With single frequency forcing, the minimum reattachment length is obtained at the non-dimensional forcing frequency (F) of St$\_$h/ = 0.29. With two-frequency forcing, a subharmonic frequency (F/2) or biharmonic frequency (2F) is combined with the fundamental frequency (F), i.e. (F, F/2) or (F, 2F) forcing is applied. In the case of (F, F/2) forcing, the reattachment length is not much sensitive to the phase difference between F and F/2. However, the reattachment length significantly depends on the phase difference between F and 2F in the case of (F, 2F) forcing. At a certain range of the phase difference, the reattachment length becomes smaller than that of the single frequency forcing.

• Korean Journal of Applied Biomechanics
• /
• v.18 no.1
• /
• pp.73-83
• /
• 2008

The purpose of this study was to investigate kinematic variables of three types of volleyball jump blocking motions through 3D video analysis. The subjects participated in this study were 7 male university volleyball players and 7 male physical education majors. 1. Regardless of blocking types, peak vertical velocity and jump height in the skilled group were faster and higher than the unskilled group. 2. The skilled group was flexed lower than the unskilled group at E2 of the hip, knee, and ankle joint during the cross over step blocking. 3. In all types of step, the peak angular velocity of the hip, knee, and ankle joint was showed at E2. 4. The skilled group was slower than the unskilled group at E2 of the angular velocity of the hip, knee, and ankle joint during cross over step blocking. In conclusion, the cross over step blocking in which the use of the upper extremities was relatively more available was effective to improve a defence skill for the jump height.