• 한국지진공학회:학술대회논문집
• /
• 한국지진공학회 2006년도 학술발표회 논문집
• /
• pp.545-552
• /
• 2006

Small-scale models have been frequently used for seismic performance tests because of limited testing facilities and economic reasons. However, there are not also enough studies on similitude law for analogizing prototype structures accurately with small-scale models, although conventional similitude law based on geometry similitude is not well consistent in their inelastic seismic behaviors. When fabricating prototype and small-scale model of reinforced concrete structures by using the same material, added mass is demanded from a volumetric change and scale factor could be limited due to aggregate size. Therefore, it is desirable to use different materials for small-scale model. In our recent study, a modified similitude law was derived depending on geometric scale factor, equivalent modulus ratio and ultimate strain ratio. And quasi-static and pseudo-dynamic tests on the specimens are carried out using constant and variable modulus ratios, and correlation between prototype and small-scale model is investigated based on their test results. In this study, tests on scaled model of different concrete compressive strength aye carried out. In shaking table tests, added mass can not be varied. Thus, constant added mass on expected maximum displacement was applied and the validity was verified in shaking table tests. And shaking table tests on non-artificial mass model is carried out to settle a limitation of acceleration and the validity was verified in shanking table tests.

• 인터넷정보학회논문지
• /
• 제22권3호
• /
• pp.9-16
• /
• 2021

This study aims to develop a human activity recognition (HAR) system as a Deep-Learning (DL) classification model, distinguishing various human activities. We solely rely on the signals from a wristband accelerometer worn by a person for the user's convenience. 3-axis sequential acceleration signal data are gathered within a predefined time-window-slice, and they are used as input to the classification system. We are particularly interested in developing a Deep-Learning model that can outperform conventional machine learning classification performance. A total of 13 activities based on the laboratory experiments' data are used for the initial performance comparison. We have improved classification performance using the Convolutional Neural Network (CNN) combined with an auto-encoder feature reduction and parameter tuning. With various publically available HAR datasets, we could also achieve significant improvement in HAR classification. Our CNN model is also compared against Recurrent-Neural-Network(RNN) with Long Short-Term Memory(LSTM) to demonstrate its superiority. Noticeably, our model could distinguish both general activities and near-identical activities such as sitting down on the chair and floor, with almost perfect classification accuracy.

• 한국분무공학회지
• /
• 제27권3호
• /
• pp.155-160
• /
• 2022

Recently, major developed countries have strengthened automobile fuel efficiency regulations and carbon dioxide emission allowance standards to curb climate change caused by global warming worldwide. Accordingly, research and manufacturing on electric vehicles that do not emit pollutants during actual driving on the road are being conducted. Several automobile companies are producing and testing electric vehicles to commercialize them, but it takes a lot of manpower and time to test and evaluate mass-produced electric vehicles with driving mileage of more than 300km on a per-charge. Therefore, in order to reduce this, a simulation model was developed in this study. This study used vehicle information and MCT speed profile of small electric vehicle as basic data. It was developed by applying Simulink, which models the system in a block diagram method using MATLAB software. Based on the vehicle dynamics, the simulation model consisted of major components of electric vehicles such as motor, battery, wheel/tire, brake, and acceleration. Through the development model, the amount of change in battery SOC and the mileage during driving were calculated. For verification, battery SOC data and vehicle speed data were compared and analyzed using CAN communication during the chassis dynamometer test. In addition, the reliability of the simulation model was confirmed through an analysis of the correlation between the result data and the data acquired through CAN communication.

• 로봇학회논문지
• /
• 제18권4호
• /
• pp.427-435
• /
• 2023

This paper proposes a method of estimating the pose of a mobile robot by using a learning model. When estimating the pose of a mobile robot, wheel encoder and inertial measurement unit (IMU) data are generally utilized. However, depending on the condition of the ground surface, slip occurs due to interaction between the wheel and the floor. In this case, it is hard to predict pose accurately by using only encoder and IMU. Thus, in order to reduce pose error even in such conditions, this paper introduces a pose estimation method based on a learning model using data of the wheel encoder and IMU. As the learning model, long short-term memory (LSTM) network is adopted. The inputs to LSTM are velocity and acceleration data from the wheel encoder and IMU. Outputs from network are corrected linear and angular velocity. Estimated pose is calculated through numerically integrating output velocities. Dataset used as ground truth of learning model is collected in various ground conditions. Experimental results demonstrate that proposed learning model has higher accuracy of pose estimation than extended Kalman filter (EKF) and other learning models using the same data under various ground conditions.

• International Journal of Naval Architecture and Ocean Engineering
• /
• 제10권5호
• /
• pp.545-565
• /
• 2018

A numerical code is developed based on potential flow theory to investigate nonlinear sloshing in rectangular Liquefied Natural Gas (LNG) tanks under forced excitation. Using this code, internal free-surface elevation and sloshing loads on liquid tanks can be obtained both in time domain and frequency domain. In the mathematical model, acceleration potential is solved in the calculation of pressure on tanks and the artificial damping model is adopted to account for energy dissipation during sloshing. The Boundary Element Method (BEM) is used to solve boundary value problems of both velocity potential and acceleration potential. Numerical calculation results are compared with published results to determine the efficiency and accuracy of the numerical code. Sloshing properties in partially filled rectangular and membrane tank under translational and rotational excitations are investigated. It is found that sloshing under horizontal and rotational excitations share similar properties. The first resonant mode and excitation frequency are the dominant response frequencies. Resonant sloshing will be excited when vertical excitation lies in the instability region. For liquid tank under rotational excitation, sloshing responses including amplitude and phase are sensitive to the location of the center of rotation. Moreover, experimental tests were conducted to analyze viscous effects on sloshing and to validate the feasibility of artificial damping models. The results show that the artificial damping model with modifying wall boundary conditions has better applicability in simulating sloshing under different fill levels and excitations.

• 제어로봇시스템학회논문지
• /
• 제19권7호
• /
• pp.620-625
• /
• 2013

Mammals such as dogs and cheetahs change their gait from trot to gallop as they run faster. However, lizards always trot for various speeds of running. When mammals run slowly with trot gait, their fore leg and hind leg generate the required force for acceleration or deceleration such that the yaw moments created by these forces cancel each other. On the other hand, when lizards run slowly, their fore legs and hind legs generate the forces for deceleration and acceleration, respectively. In this paper, the yaw motion of a lizard model is controlled by the movement of their waist and tail, and the reaction moment from the ground produced by the hind legs in simulation. The simulation uses the whole body dynamics of a lizard model, which consists of 4 links based on the Callisaurus draconoides. The results show that the simulated trotting of the model is similar to that of a real lizard when the movement of the model is optimized to minimize the reaction moment from the ground. It means that the body of a lizard moves in such a way that the reaction moment from the ground is minimized. This demonstrates our hypothesis on how lizards trot using body motion.

• 대한기계학회논문집A
• /
• 제41권10호
• /
• pp.967-974
• /
• 2017

본 연구에서는 횡가속도 센서 계측 신호 기반의 기준 차량 요레이트 모델을 활용하여 횡경사 유무에 관계 없이 견실한 성능을 보장하는 제동기반 요 모멘트 제어시스템을 개발하였다. 2자유도 single track 모델과 횡가속도 센서 계측 신호를 융합하여 새로운 기준 요레이트 모델을 설계하였고 이를 기반으로 요 모멘트 제어기를 설계하였다. 또한 외란 관측기를 적용하여 요레이트 동역학에 존재하는 차량 파라미터 오차를 보상하고 제어기의 성능을 개선하였다. 다자유도 차량동역학 해석 SW인 CARSIM을 이용하여 평지 및 횡경사 노면을 반영한 다양한 검증 시나리오 조건에서 제안된 제어기를 검증하였다. 그 결과 기준 차량모델에 횡가속도 계측 신호를 반영하고 외란 관측기를 통해 모델 파라미터 오차를 보상하는 것을 특징으로 하는 새롭게 제안된 횡안정성 제어기가 도로 횡경사에 관계없이 다양한 주행상황에서 차량의 횡안정성을 견실하게 유지할 수 있음을 확인하였다.

• 대한토목학회논문집
• /
• 제29권1A호
• /
• pp.61-73
• /
• 2009

본 논문에서는 강판형교의 주된 두 손상유형인 거더의 휨 강성 저하와 지점부의 손상을 검색하기 위해 가속도-임피던스 특성을 이용한 하이브리드 구조건전성 모니터링 기법을 제안하였다. 하이브리드 기법은 1) 전역적인 방법으로 손상의 발생을 경보하고, 2) 구조물의 구조 부재내의 발생된 손상을 분류하며, 3) 구조 부재에 따라 적절한 방법을 이용하여 세부적으로 분류된 손상을 평가하는 크게 3단계로 구성되었다. 첫 번째 단계에서는 가속도 특성 변화를 모니터링하여 전역적인 손상의 발생을 경보한다. 두 번째 단계에서는 임피던스 특성 변화를 모니터링하여 경보된 손상유형을 분류한다. 세 번째 단계에서는 모드변형에너지기반 손상지수법과 RMSD 기법을 이용하여 손상의 위치와 크기를 평가한다. 몇몇의 손상 시나리오에 의해 측정된 하이브리드 가속도-임피던스 신호를 이용한 모형 강판형교 실험을 통해 제안된 하이브리드 기법의 유용성을 평가하였다. 또한, 온도변화 및 지점손상 조건에 대한 실험을 통해 임피던스기반 손상모니터링의 정확도에 미치는 온도유발 영향을 검토하였다.

• 한국항공우주학회지
• /
• 제50권8호
• /
• pp.531-540
• /
• 2022

본 논문에서는 IMM 필터 기반으로 장사정포의 탄종을 식별하고 탄착점을 신속하게 예측하는 알고리즘을 제시한다. 탄도궤적 방정식을 시스템 모델로 사용하고, 각각 다른 탄도계수 값을 갖는 3가지 모델을 IMM 필터에 적용한다. 가속도를 중력, 공기저항, 양력에 의한 3가지 성분으로 나누고 양력가속도를 새로운 상태변수로 추가하여 추정한다. 속도벡터와 양력가속도가 수직이라는 운동학 조건을 유사 측정값으로 추가한 측정방정식을 다룬다. IMM 필터를 통해 추정된 상태변수와 모드 확률이 가장 높은 모델의 탄도계수를 기반으로 탄착점을 예측한다. 탄착점 예측을 위해 일반적으로 사용되는 룽게-쿠타 수치적분 대신, 준해석적인 방법을 사용하여 적은 계산량으로 탄착점을 예측할 수 있음을 설명한다. 마지막으로 최소제곱법을 이용한 상태변수 초기화 방법에 대해 제안하고 성능을 확인하였다. 탄종식별, 탄착점 예측 및 초기화를 포함한 통합 알고리즘을 제시하고 시뮬레이션을 통해 제안한 방법의 타당성을 검증하였다.

• 한국산업보건학회지
• /
• 제16권4호
• /
• pp.398-412
• /
• 2006

The purpose of this study is to investigate the relationship between the acceleration of vibration by the powered hand tools used in the shipbuilding industry, and to develop the prospective prevalence model for the hand-arm vibration syndrome among the shipbuilding workers.The acceleration levels and frequencies of six types of grinder were measured using the ISO5349 method along with the time of exposure to the vibration from the powered hand tools. Medical examination for 114 workers were performed using the cold provocation test. Comparisons were made between the estimated prevalence of hand-arm vibration syndrome from ISO5349 and the observed values from the medical examinations. By multiple regression, we developed the prospective prevalence model of hand-arm vibration syndrome produced by the hand tools used in the shipbuilding industry. 4 hour-energy-equivalent frequency-weighted accelerations were $6.23m/s^2$ in the grinding job done after welding, and $13.39m/s^2$ in the grinding job done before painting. The mean exposure time while holding powered hand tools was 4.64 hours. Prevalence rates of Raynaud's Phenomenon were 12.04% in the grinding after soldering, and 42.9% in the grinding before painting measured using the ISO5349 method. After exposure to vibration for 10.79 years, about a half of the workers in the grinding after welding could developed Raynaud's Phenomenon. For the workers in the grinding before painting, the latency was 5.02 years. The ISO equation for dose response relationship was not significantly correlated with observed recovery rates of finger skin temperatures, blood flows and amplitudes of nerve conduction velocities. A multiple regression model for dose-response relationship was proposed from the results. Recovery rate of the skin temperatures = -0.668+ 0.337 ${\times}$ 4 hour energy equivalent frequency-weighted accelerations + 0.767 ${\times}$ duration of vibration exposure(years) The validity was proved by multiple regression analysis after correlation transformation and regression results based on model-building data and validation data.