• Journal of the Society of Naval Architects of Korea
• /
• v.51 no.2
• /
• pp.162-170
• /
• 2014

A CFD(Computational Fluid Dynamics) analysis is presented to predict hydrodynamic characteristics of a marine propeller. A commercial RANS(Reynolds Averaged Navier-Stokes equation) solver, namely FLUENT, is utilized in conjunction with fully unstructured meshes around rotating propeller. Mesh generation process is greatly accelerated by using fully unstructured meshes composed of both isotropic and anisotropic tetrahedral elements. The anisotropic tetrahedral elements were used in the flow domain near the blade and shaft, where the viscous effect is important, having complex shape yet resolving the thin boundary layers. For other regions, isotropic tetrahedral elements are utilized. Two different approaches simulating rotational effect of the propeller are employed, namely Moving reference frame technique for steady simulation, and Sliding mesh technique for unsteady simulation. Both approaches are applied to the propeller open water (POW) test simulation. The current results, which are thrust and torque coefficients, are compared with available experimental data.

• International Journal of Fluid Machinery and Systems
• /
• v.4 no.1
• /
• pp.199-208
• /
• 2011

This work presents numerical results, using OpenFOAM, of the flow in the swirl flow generator test rig developed at Politehnica University of Timisoara, Romania. The work shows results computed by solving the unsteady Reynolds Averaged Navier Stokes equations. The unsteady method couples the rotating and stationary parts using a sliding grid interface based on a GGI formulation. Turbulence is modeled using the standard k-${\varepsilon}$ model, and block structured wall function ICEM-Hexa meshes are used. The numerical results are validated against experimental LDV results, and against design velocity profiles. The investigation shows that OpenFOAM gives results that are comparable to the experimental and design profiles. The unsteady pressure fluctuations at four different positions in the draft tube is recorded. A Fourier analysis of the numerical results is compared whit that of the experimental values. The amplitude and frequency predicted by the numerical simulation are comparable to those given by the experimental results, though slightly over estimated.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.1-16
• /
• 2022

Despite recent breakthroughs in deep learning and computer vision fields, the pixel-wise identification of tiny objects in high-resolution images with complex disturbances remains challenging. This study proposes a modified U-net for tiny crack segmentation in real-world steel-box-girder bridges. The modified U-net adopts the common U-net framework and a novel Self-Attention-Self-Adaption (SASA) neuron as the fundamental computing element. The Self-Attention module applies softmax and gate operations to obtain the attention vector. It enables the neuron to focus on the most significant receptive fields when processing large-scale feature maps. The Self-Adaption module consists of a multiplayer perceptron subnet and achieves deeper feature extraction inside a single neuron. For data augmentation, a grid-based crack random elastic deformation (CRED) algorithm is designed to enrich the diversities and irregular shapes of distributed cracks. Grid-based uniform control nodes are first set on both input images and binary labels, random offsets are then employed on these control nodes, and bilinear interpolation is performed for the rest pixels. The proposed SASA neuron and CRED algorithm are simultaneously deployed to train the modified U-net. 200 raw images with a high resolution of 4928 × 3264 are collected, 160 for training and the rest 40 for the test. 512 × 512 patches are generated from the original images by a sliding window with an overlap of 256 as inputs. Results show that the average IoU between the recognized and ground-truth cracks reaches 0.409, which is 29.8% higher than the regular U-net. A five-fold cross-validation study is performed to verify that the proposed method is robust to different training and test images. Ablation experiments further demonstrate the effectiveness of the proposed SASA neuron and CRED algorithm. Promotions of the average IoU individually utilizing the SASA and CRED module add up to the final promotion of the full model, indicating that the SASA and CRED modules contribute to the different stages of model and data in the training process.

• Journal of Korea Spatial Information System Society
• /
• v.11 no.2
• /
• pp.89-98
• /
• 2009

As data stream is entered into system continuously and the memory space is limited, the data exceeding the memory size cannot be processed. In order to solve the problem, load shedding methods which drop a part of data to prevent exceeding the storage space have been researched. Generally, a traditional load shedding method uses random sampling with optimized rate according to data deviation. The method samples data not to distinguish those used in spatial query because the method uses only a random sampling with optimized rate according to data deviation. Therefore, the accuracy of query was reduced in u-GIS environment including spatial query. In this paper, we researched a new load shedding method improving accuracy of the query in u-GIS environment which runs spatial query and aspatial query simultaneously. The method uses a new sampling method that samples data having low probability used in query. Therefore proposed method improves spatial query accuracy and query processing speed as applying spatial filtering operation to sampling operator.

• International Journal of Fluid Machinery and Systems
• /
• v.6 no.4
• /
• pp.200-205
• /
• 2013

In order to investigate the characteristics of unsteady flow in a mixed flow pump guide vane under the small flow conditions, several indicator points in a mixed flow pump guide vane was set, the three-dimensional unsteady turbulence numerical value of the mixed flow pump which is in the whole flow field will be calculated by means of the large eddy simulation (LES), sub-grid scale model and sliding mesh technology. The experimental results suggest that the large eddy simulation can estimate the positive slope characteristic of head & capacity curve. And the calculation results show that the pressure fluctuation coefficients of the middle section in guide vane inlet will decrease firstly and then increase. In guide vane outlet, the pressure fluctuation coefficients of section will be approximately axially symmetrical distribution. The pressure fluctuation minimum of section in guide vane inlet is above the middle location of the guide vane suction surface, and the pressure fluctuation minimum of section in which located the middle and outlet of guide vane. When it is under the small flow operating condition, the eddy scale of guide vane is larger, and the pressure fluctuation of the channel in guide vane being cyclical fluctuations obviously which leads to the area of eddy expanding to the whole channel from the suction side. The middle of the guide vane suction surface of the minimum amplitude pressure fluctuation to which the vortex core of eddy scale whose direction of fluid's rotation is the same to impeller in the guide vane adhere.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2017.05a
• /
• pp.260-260
• /
• 2017

This study aims to compare the performance of TRIGRS (Transient Rainfall Infiltration and Grid-based Regional Slope-stability model) and TiVaSS (Time-variant Slope Stability model) in the prediction of rainfall-induced shallow landslides. TRIGRS employs one-dimensional (1-D) subsurface flow to simulate the infiltration rate, whereas a three-dimensional (3-D) model is utilized in TiVaSS. The former has been widely used in landslide modeling, while the latter was developed only recently. Both programs are used for the spatiotemporal prediction of shallow landslides caused by rainfall. The present study uses the July 2011 landslide event that occurred in Mt. Umyeon, Seoul, Korea, for validation. The performance of the two programs is evaluated by comparison with data of the actual landslides in both location and timing by using a landslide ratio for each factor of safety class ( index), which was developed for addressing point-like landslide locations. In addition, the influence of surface flow on landslide initiation is assessed. The results show that the shallow landslides predicted by the two models have characteristics that are highly consistent with those of the observed sliding sites, although the performance of TiVaSS is slightly better. Overland flow affects the buildup of the pressure head and reduces the slope stability, although this influence was not significant in this case. A slight increase in the predicted unstable area from 19.30% to 19.93% was recorded when the overland flow was considered. It is concluded that both models are suitable for application in the study area. However, although it is a well-established model requiring less input data and shorter run times, TRIGRS produces less accurate results.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.47 no.4
• /
• pp.390-402
• /
• 2011

A preliminary study on species composition of a Korean purse seine catch landed at cannery was conducted in April 2011. In the cannery, all tuna catch are sliding through a sorting grid panel that filters and drops fish in the buckets by size class (above 9kg, 3.4-9kg, 1.8-3.4kg, 1.4-1.8kg and below 1.4kg). In cannery processing, species sorting was made for skipjack tuna and yellowfin tuna only from catches greater than 3.4kg during filtering but not for bigeye tuna because of difficulties in species identification between bigeye tuna and yellowfin tuna under frozen state. As no species identification was carried out for catch groups less than 3.4kg in the cannery process, this study focused on sorting out skipjack tuna and yellowfin tuna from these groups and then identifying bigeye tuna from all size groups of yellowfin tuna. Using the mixture rate of species obtained from the samples taken, species composition of the landed catch was estimated. As results, cannery research showed 95% for skipjack tuna, 3% for yellowfin tuna and 2% for bigeye tuna in species composition, while vessel logbook data represented 96%, 3% and 1% for skipjack tuna, yellowfin tuna and bigeye tuna, respectively. The proportion of bigeye tuna identified in the cannery was slightly higher than shown in logbook data by 1%.

• Journal of the Microelectronics and Packaging Society
• /
• v.12 no.4 s.37
• /
• pp.289-299
• /
• 2005

The ball shear force was investigated in terms of test parameters, i.e. displacement rate and probe height, with an experimental and non-linear finite element analysis for evaluation of the solder joint integrity in area array packages. The increase in the displacement rate and the decrease in the probe height led to the increase in the shear force. Excessive probe height could cause some detrimental effects on the test results such as unexpected high standard deviation and probe sliding from the solder ball surface. The low shear height conditions were favorable for assessing the mechanical integrity of the solder joints. The mechanical and electrical properties of the Sn-37Pb/Cu and Sn-3.5Ag/Cu BGA solder joints were also investigated with the number of reflows. The total thickness of the intermetallic compound (IMC) layers, consisting of Cu6Sn5 and Cu3Sn, was increased as a function of cubic root of reflow time. The shear force was increased up to 3 or 4 reflows, and then was decreased with the number of reflows. The fracture occurred along the bulk solder, in irrespective of the number of reflows. The electrical resistivity was increased with increasing the number of reflows.

• Journal of Navigation and Port Research
• /
• v.46 no.4
• /
• pp.367-374
• /
• 2022

The purpose of this study was to propose a deep learning algorithm that applies to the fault diagnosis of fuel pumps and purifiers of autonomous ships. A deep learning algorithm reflecting the time dependence of the measured signal was configured, and the failure pattern was trained using the vibration signal, measured in the equipment's regular operation and failure state. Considering the sequential time-dependence of deterioration implied in the vibration signal, this study adopts Conv1D with sliding window computation for fault detection. The time dependence was also reflected, by transferring the measured signal from two-dimensional to three-dimensional. Additionally, the optimal values of the hyper-parameters of the Conv1D model were determined, using the grid search technique. Finally, the results show that the proposed data preprocessing method as well as the Conv1D model, can reflect the sequential dependency between the fault and its effect on the measured signal, and appropriately perform anomaly as well as failure detection, of the equipment chosen for application.