• Wind and Structures
• /
• 제39권4호
• /
• pp.271-285
• /
• 2024

The multiple-fan wind tunnel is an important facility for reproducing target wind field. Existing control methods for the multiple-fan wind tunnel can generate wind speeds that satisfy the target statistical characteristics of a wind field (e.g., power spectrum). However, the frequency-domain features cannot well represent the nonstationary winds of extreme storms (e.g., downburst). Therefore, this study proposes a multiple-fan wind tunnel control scheme based on Deep Reinforcement Learning (DRL), which will completely transform into a data-driven closed-loop control problem, to reproduce the target wind field in the time domain. Specifically, the control scheme adopts the Deep Deterministic Policy Gradient (DDPG) paradigm in which the strong fitting ability of Deep Neural Networks (DNN) is utilized. It can establish the complex relationship between the target wind speed time series and the current control strategy in the DRL-agent. To address the fluid memory effect of the wind field, this study innovatively designs the system state and control reward to improve the reproduction performance based on historical data. To validate the performance of the model, we established a simplified flow field based on Navier Stokes equations to simulate a two-dimensional numerical multiple-fan wind tunnel environment. Using the strategy of DRL decision maker, we generated a wind speed time series with minor error from the target under low Reynolds number conditions. This is the first time that the AI methods have been used to generate target wind speed time series in a multiple-fan wind tunnel environment. The hyperparameters in the DDPG paradigm are analyzed to identify a set of optimal parameters. With these efforts, the trained DRL-agent can simultaneously reproduce the wind speed time series in multiple positions.

• 대기
• /
• 제25권4호
• /
• pp.659-667
• /
• 2015

In this study, sensitivity of inflow wind speed and turbulent Schmidt number to pollutant dispersion in an urban street canyon is investigated, by comparing CFD-simulated results to wind-tunnel results. For this, we changed systematically inflow wind speed at the street-canyon height ($1.5{\sim}10.0m\;s^{-1}$ with the increment of $0.5m\;s^{-1}$) and turbulent Schmidt number (0.2~1.3 with interval of 0.1). Also, we performed numerical experiments under the conditions that turbulent Schmidt numbers selected with the magnitude of mean kinetic energy at each grid point were assigned in the street canyon. With the increase of the inflow wind speed, the model underestimated (overestimated) pollutant concentration in the upwind (downwind) side of the street canyon because of the increase of pollutant advection. This implies that, for more realistic reproduction of pollutant dispersion in urban street canyons, large (small) turbulent Schmidt number should be assigned for week (strong) inflow condition. In the cases of selectively assigned turbulent Schmidt number, mean bias remarkably decreased (maximum 60%) compared to the cases of constant turbulent Schmidt number assigned. At week (strong) inflow wind speed, root mean square error decreases as the area where turbulent Schmidt number is selectively assigned becomes large (small).

• 한국컴퓨터그래픽스학회논문지
• /
• 제28권3호
• /
• pp.67-77
• /
• 2022

자연풍에 의한 감각은 대부분의 사람들이 일상 생활에서 경험하는 가장 흔한 느낌 중 하나이다. 그러나 가상현실 환경에서 자연풍이 어떻게 재현될 수 있는지, 인공풍이 결합된 다감각 콘텐츠가 인간의 정서를 개선하는지에 대한 연구는 거의 수행되지 않았다. 이러한 문제를 다루기 위해, 본 연구는 녹화된 영상 및 바람 데이터를 기반으로 하는 Wind Reproduction VR System 을 제안하고, 이 시스템이 사용자의 정적 정서에 주는 영향을 확인하는 연구를 진행한다. 실험을 위해 요트상에서 360 도 영상과 함께 풍향 및 풍속 데이터를 수집했다. 수집한 데이터는 제안한 시스템을 통해 다감각 VR 환경을 만드는데 사용되었다. 총 19 명의 대학생들이 실험에 참여했으며, K-PANAS(Korean version of Positive and Negative Affect Schedule)를 통하여 참가자들의 정서 변화를 측정했다. 실험 결과, 인공풍이 추가된 요트 VR 콘텐츠 체험 이후 참가자의 '영감을 받다', '활기차다' 정서가 유의하게 증가하였다. 뿐만 아니라, 바람의 유무에 따라 '흥미롭다' 정서가 가장 큰 영향을 받는 것으로 확인하였다. 제안한 시스템은 인터랙티브 미디어, 체험형 콘텐츠와 같은 다양한 VR 응용 프로그램에서 효과적으로 활용될 수 있다.

• 한국해양공학회지
• /
• 제24권6호
• /
• pp.23-29
• /
• 2010

Korean shipbuilding companies have sometimes carried out sea trials to measure a vessel's speed performance around the western channel of the Straits of Korea, where the flow fields are very complicated because of the effect of various flows such as sea, tidal, geostrophic, and wind-driven currents. Because these flows seem to present significant interference to a ship, the numerical reproduction of the flow-fields in the vicinity of the target sites could provide a better understanding of the sea environments while performing sea trials. In this study, we used the MEC ocean model to simulate the tidal currents around Tsushima Island and compared the simulated tidal amplitudes and currents with the measurements of Teague et al. (2001). The tidal amplitudes of the present simulation results agreed well with the observations. Based on the numerical simulation, the optimal direction and proper sites for a speed trial are described.