• Wind and Structures
• /
• v.36 no.6
• /
• pp.423-434
• /
• 2023

Aerodynamic force coefficients are generally obtained from traditional wind tunnel tests or computational fluid dynamics (CFD). Unfortunately, the techniques mentioned above can sometimes be cumbersome because of the cost involved, such as the computational cost and the use of heavy equipment, to name only two examples. This study proposed to build a deep neural network model to predict the aerodynamic force coefficients based on data collected from CFD simulations to overcome these drawbacks. Therefore, a series of CFD simulations were conducted using different geometric parameters to obtain the aerodynamic force coefficients, validated with wind tunnel tests. The results obtained from CFD simulations were used to create a dataset to train a multilayer perceptron artificial neural network (ANN) model. The models were obtained using three optimization algorithms: scaled conjugate gradient (SCG), Bayesian regularization (BR), and Levenberg-Marquardt algorithms (LM). Furthermore, the performance of each neural network was verified using two performance metrics, including the mean square error and the R-squared coefficient of determination. Finally, the ANN model proved to be highly accurate in predicting the force coefficients of similar bridge sections, thus circumventing the computational burden associated with CFD simulation and the cost of traditional wind tunnel tests.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.11
• /
• pp.3083-3098
• /
• 2023

As demands for efficient and echo-friendly production of marine products increase, smart aquaculture based on information and communication technology (ICT) has become a promising trend. The smart aquaculture is expected to control fundamental farm environment variables including water temperature and dissolved oxygen (DO) levels with less human intervention. A recirculating aquaculture system (RAS) is required for the smart aquaculture which utilizes a purification tank to reuse water drained from the water tank while blocking the external environment. Elaborate water treatment should be considered to properly operate RAS. However, analyzing the water treatment performance is a challenging issue because fish farm circumstance continuously changes and recursively affects water fluidity. To handle this issue, we introduce computational fluid dynamics (CFD) aided water treatment analysis including water fluidity and the solid particles removal efficiency. We adopt RAS parameters widely used in the real aquaculture field to better reflect the real situation. The simulation results provide several indicators for users to check performance metrics when planning to select appropriate RAS without actually using it which costs a lot to operate.

• Journal of the Semiconductor & Display Technology
• /
• v.22 no.1
• /
• pp.134-138
• /
• 2023

This study focuses on the analysis of the results of computational fluid dynamics simulations of mist-chemical vapor deposition for the growth of an epitaxial wafer in power semiconductor technology using artificial intelligence techniques. The conventional approach of predicting the uniformity of the deposited layer using computational fluid dynamics and design of experimental takes considerable time. To overcome this, artificial intelligence method, which is widely used for optimization, automation, and prediction in various fields, was utilized to analyze the computational fluid dynamics simulation results. The computational fluid dynamics simulation results were analyzed using a supervised deep neural network model for regression analysis. The predicted results were evaluated quantitatively using Euclidean distance calculations. And the Bayesian optimization was used to derive the optimal condition, which results obtained through deep neural network training showed a discrepancy of approximately 4% when compared to the results obtained through computational fluid dynamics analysis. resulted in an increase of 146.2% compared to the previous computational fluid dynamics simulation results. These results are expected to have practical applications in various fields.

• 유체기계공업학회:학술대회논문집
• /
• 1997.02a
• /
• pp.137-144
• /
• 1997

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.28 no.7
• /
• pp.1274-1280
• /
• 2022

The resistance performance evaluation of general ships using computational fluid dynamics requires a lot of time and cost, and various methods are being studied to reduce the time and cost. Existing methods using main particulars or cross sections of ships have limitations in estimating resistance performance that is greatly dependent on the shape of the ship. In this paper, we propose a deep neural network model that can quickly predict the resistance performance of the hull surface by inputting the geometric information of the hullform mesh. The proposed deep neural network model based on Perceiver IO can immediately predict resistance performance, unlike computational fluid dynamics techniques that require calculation in each time step. It shows the result of estimating the resistance performance with an average error of less than 1% in the data set for a 50 K tanker ship, a type of low-speed full ship.

• Proceedings of the KSME Conference
• /
• 2004.11a
• /
• pp.1454-1459
• /
• 2004

The major parameters governing the fluid dynamical and thermo-dynamical behavior in the large pipeline network system are friction loss and the pipeline length. But in local pipeline networks and relatively short distance pipeline system, secondary loss and the considerations of the moving states of the fluid machine are also important. One of the major element in local pressure control system is pressure regulator. It causes the variations of the physical properties in that pipeline system. When it is under working, the accurate analysis of the flow properties is so difficult. In this study, some numerical approaches to investigate the critical-flow-characteristics of the pressure regulator have been done according to the variations of the opening ratio or cross-sectional area and the detail examinations and considerations of the pressure regulator as a pipeline network elements have been carried. Finally the flow-flied distributions and critical-flow-characteristics have been presented in detail and the critical flow phenomena and the relation to the opening ratio or cross-sectional-area ratio have been studied.

• Journal of the Korean Institute of Gas
• /
• v.1 no.1
• /
• pp.14-20
• /
• 1997

Pipeline network analysis requires fluid mechanics. A lot of equations have been used for flow analysis according to the behavior of fluid in pipelines and the operative situations. In this paper, simulation and analysis have been performed for the pipeline network system in Kyungin area using a steady-state mathematical model. Then, a statistical model using partial least squares(PLS) method has been developed with the data obtained from the developed mathematical model. The results showed that it is possible to simulate and analyze pipeline network systems using statistical approach.

• Korea-Australia Rheology Journal
• /
• v.14 no.2
• /
• pp.49-55
• /
• 2002

To understand the large amplitude oscillatory shear (LAOS) behavior of complex fluids, we have investigated the flow behavior of a network model in the LAOS environment. We applied the LAOS flow to the model proposed by Vaccaro and Marrucci (2000), which was originally developed to describe the system of associating telechelic polymers. The model was found to predict at least three different types of LAOS behavior; strain thinning (G' and G" decreasing), strong strain overshoot (G' and G" increasing followed by decreasing), and weak strain overshoot (G' decreasing, G" increasing followed by decreasing). The overshoot behavior in the strain sweep test, which il often observed in some complex fluid systems with little explanation, could be explained in terms of the model parameters, or in terms of the overall balance between the creation and loss rates of the network junctions, which are continually created and destroyed due to thermal and flow energy. This model does not predict strain hardening behavior because of the finitely extensible nonlinear elastic (FENE) type nonlinear effect of loss rate. However, the model predicts the LAOS behavior of most of the complex fluids observed in the experiments.he experiments.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.53 no.1
• /
• pp.1-7
• /
• 2011

The nature of a network implies movement among vertices, and can be regarded as flows. Based on the flow concept which network follows the hydraulic fluid principle, we develop a spatial network model using Bernoulli equation. Then we explore the organization of spatial network and growth by the velocity of network flows. Results show that flow velocity determines network connections or influence of a vertex up to a point, and that the overall network structure is the result of pull force (pressure) and flow velocity. We demonstrate how one vertex can monopolize connections within a network.

• Journal of the Korean Statistical Society
• /
• v.32 no.4
• /
• pp.359-371
• /
• 2003

We consider the probability that the total population of a stable Jackson network reaches a given large value. By using the fluid limit of the reversed network, we derive new upper and lower bounds on this probability, which are sharper than those in Glasserman and Kou (1995). In particular, the improved lower bound is useful for analyzing the performance of an importance sampling estimator for the overflow probability in Jackson tandem networks. Bounds on the expected time to overflow are also obtained.