• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.35 no.7
• /
• pp.647-654
• /
• 2007

A Discrete continuation Method/homotopy approaches are studied for energy/fuel optimal low thrust Earth escape trajectory by solving a two point boundary value problem(TPBVP). Recently, maneuvers using low thrust propulsion system have been identified as emerging technologies. The low thruster is considered as the main actuator for orbit maneuvers. The cost function consists of a energy/fuel consumption function, and constraints are position and velocity vectors at the terminal escape point. Solving the minimum energy/fuel problem directly is not an easy task, so we adopt the homotopy analysis. Using a solution of the minimum energy, which is solved by discrete continuation method, we obtain the solution of the minimum fuel problem.

• Journal of Korean Association for Spatial Structures
• /
• v.23 no.2
• /
• pp.29-36
• /
• 2023

Reinforcement learning (RL) is widely applied to various engineering fields. Especially, RL has shown successful performance for control problems, such as vehicles, robotics, and active structural control system. However, little research on application of RL to optimal structural design has conducted to date. In this study, the possibility of application of RL to structural design of reinforced concrete (RC) beam was investigated. The example of RC beam structural design problem introduced in previous study was used for comparative study. Deep q-network (DQN) is a famous RL algorithm presenting good performance in the discrete action space and thus it was used in this study. The action of DQN agent is required to represent design variables of RC beam. However, the number of design variables of RC beam is too many to represent by the action of conventional DQN. To solve this problem, multi-agent DQN was used in this study. For more effective reinforcement learning process, DDQN (Double Q-Learning) that is an advanced version of a conventional DQN was employed. The multi-agent of DDQN was trained for optimal structural design of RC beam to satisfy American Concrete Institute (318) without any hand-labeled dataset. Five agents of DDQN provides actions for beam with, beam depth, main rebar size, number of main rebar, and shear stirrup size, respectively. Five agents of DDQN were trained for 10,000 episodes and the performance of the multi-agent of DDQN was evaluated with 100 test design cases. This study shows that the multi-agent DDQN algorithm can provide successfully structural design results of RC beam.

• International Journal of CAD/CAM
• /
• v.7 no.1
• /
• pp.41-49
• /
• 2007

In density-based topology optimization, 0/1 solutions are sought. Discrete topological problems are often relaxed with continuous design variables so that they can be solved using continuous mathematical programming. Although the relaxed methods are practical, grey areas appear in the optimum topologies. SIMP (Solid Isotropic Microstructures with Penalization) employs penalty schemes to suppress the intermediate densities. SRV (the Sum of the Reciprocal Variables) drives the solution to a 0/1 layout with the SRV constraint. However, both methods cannot effectively remove all the grey areas. SRV has some numerical aspects. In this work, a new scheme SIMP-SRV is proposed by combining SIMP and SRV approaches, where SIMP is employed to generate an intermediate solution to initialize the design variables and SRV is then adopted to produce the final design. The new method turned out to be very effective in conjunction with the method of moving asymptotes (MMA) when using for the stiffness topology optimization of continuum structures for minimum compliance. The numerical examples show that the hybrid technique can effectively remove all grey areas and generate stiffer optimal designs characterized with a sharper boundary in contrast to SIMP and SRV.

• The Transactions of the Korean Institute of Electrical Engineers P
• /
• v.58 no.4
• /
• pp.455-460
• /
• 2009

A number of factors combine to make ultraviolet radiation a superior means of water purification for ground water, rainwater harvesting systems and so on. Ultraviolet radiation is capable of destroying all types of bacteria. Additionally, ultraviolet radiation disinfects rapidly without the use of heat or chemical additives which may undesirably alter the composition of water. In a typical operation, water enters the inlet of a UV lamp and flows through the annular space between the quartz sleeve and the outside chamber wall. The irradiated water leaves through the outlet nozzle. Several design features are combined to determine the dosage delivered. The first is Wavelength output of the lamp, the Second is Length of the lamp - when the lamp is mounted parallel to the direction of water flow, the exposure time is proportional to the length of the lamp, the third is Design water flow rate - exposure time is inversely related to the linear flow rate, the forth is Diameter of the purification chamber - since the water itself absorbs UV energy, the delivered dosage diminishes logarithmically with the distance from the lamp. In this paper, It describe the how to design optimal UV disinfection device for ground water and rainwater. To search the optimal design method, it was performed computer simulation with 3D-CFD discrete ordinates model and manufactured prototype. Using proposed design method manufactured prototype applied to disinfection test and proved satisfied performance.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.29 no.9 s.240
• /
• pp.1262-1269
• /
• 2005

Thrust of linear motor is one of the important factor to specify motor performance. Maximum thrust can be obtained by increasing the current in conductor and is relative to the sizes of conductor and magnet. But, the current and the size of conductor have an effect on temperature of linear motor. Therefore, it is practically important to find design results that can effectively maximize the thrust of linear motor within limited range of temperature. Finite element analysis was applied to calculate thrust and the temperature of the conductor was calculated by the thermal resistance. The diameter of copper wire among design variables has discrete value and number of turns must be integer. Considering these facts, special techinque for optimum design is presented. To reduce excessive computation time of thrust in optimization, the design variables was redefined by analysis of variance and second order regression model for thrust was determined by response surface metheodology. As a result, it is shown that the proposed method has an advantage in optimum design of linear motor.

• Proceedings of the KIPE Conference
• /
• 2018.07a
• /
• pp.344-345
• /
• 2018

In order to alleviate the negative impacts of harmonically distorted grid condition on grid-connect inverters, an optimal control design-based gain selection scheme of an LCL-filtered grid-connected inverter and its ability to compensate selective harmonics are presented in this paper. By incorporating resonant terms into the control structure in the state-space to provide infinity gain at selected frequencies, the proposed control offers an excellent steady-state response even under distorted grid voltage. The proposed control scheme is achieved by using a state feedback controller for stabilization purpose and by augmenting the resonant terms as well as intergral term into a control structure for reference tracking and harmonic compensation. Furthermore, the optimal linear quadratic control approach is adopted for choosing an optimal feedback gain to ensure an asymptotic stability of the whole system. A discrete-time full state observer is also introduced into the proposed control scheme for the purpose of reducing a total number of sensors used in the inverter system. The simulation results are given to prove the effectiveness and validity of the proposed control scheme.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.16 no.1
• /
• pp.1-12
• /
• 1991

In this paper the design and construction of discrete model following servo dontroller on the position control system is proposed. The operational time delay of the plant in the controller which is proposed, is considered and the system which is added by the integral compensation in first order difference equation is constructed. By applying the optimal regulator method to the system, the method which find the optimal state feedback gain is developed theoretically. The output of a model which is correspond to a DC Servo motor follow quickly the speed response of a DC Servo motor and the velocity error in ansteady-state is reduced in zero and the position response is controlled correctly, the performance of the controller is contoller is confirmed by Computer Simulation.

• Journal of Power Electronics
• /
• v.18 no.6
• /
• pp.1771-1779
• /
• 2018

Modern trams with a super capacitor have gained a lot of attention in recent years due to its reliability, convenience, energy conservation and environmental friendliness. Because of its special charging characteristic, the traditional charging structure and control strategy cannot satisfy its charging requirements. This paper presents a new charging topology for fast charging modern trams with a super capacitor and it designs a controller using continuous control set model predictive control (CCS-MPC). There are three contributions in this paper. First, a new charging structure is designed and its mathematics model is derived. The cascade structure is adopted instead of the parallel structure to simplify the control process and to keep the rated power of the controllable part low. Second, a MPC control strategy is proposed to satisfy the charging characteristic. The optimal control signal can be obtained by solving the designed optimization problem. The optimal control signal is related to the discrete control action. In addition, mapping between the continuous control signal and the discrete control action is designed. Third, a semi-physical experimental platform is built to verify the proposed topology and control method. The simulation model and experiment platform are built to verify the correctness of the new structure and its control method. The results obtained show that the new topology can work effectively.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.15 no.2
• /
• pp.315-327
• /
• 2002

The main objective of this study is to develop an discrete optimization algorithm of plane steel frames with rigid using second-order-elastic-plastic hinge analysis which is considering panel zone. Conventional analyses of steel frame are usually tarried out without considering the effect of panel zone deformation on frame behavior The validity of this model is established by comparison without panel zone on joint models is analyzed numerically to demonstrate the importance of using realistic models in steel frame analysis. The objective function is taken as Weight of steel frames and the constraints we formulated based on AISC-LRFD(1994). The validity of the developed algorithm we demonstrate by comparing the result with those of SAP2000. The result of the study indicates that the optimal design algorithm considering of panel zone behavior more economic design than simple steel frame design methods.

• Computers and Concrete
• /
• v.9 no.5
• /
• pp.327-340
• /
• 2012

The search for a design that meets both performance and safety, with minimal cost and lesser environmental impact was always the goal of structural engineers. In general, the design of conventional reinforced concrete structures is an iterative process based on rules of thumb established from the personal experience and intuition of the designer. However, such procedure makes the design process exhaustive and only occasionally leads to the best solution. In such context, this work presents the development and implementation of a mathematical formulation for obtaining optimal sections of reinforced concrete columns subjected to uniaxial flexural compression, based on the verification of strength proposed by the Brazilian standard NBR 6118 (ABNT 2007). To minimize the cost of the reinforced concrete columns, the Simulated Annealing optimization method was used, in which the amount and diameters of the reinforcement bars and the dimensions of the columns cross sections were considered as discrete variables. The results obtained were compared to those obtained from the conventional design procedure and other optimization methods, in an attempt to verify the influence of resistance class, variations in the magnitudes of bending moment and axial force, and material costs on the optimal design of reinforced concrete columns subjected to uniaxial flexural compression.