• 한국컴퓨터정보학회논문지
• /
• 제15권4호
• /
• pp.57-63
• /
• 2010

Levenberg-Marquardt 알고리즘에서 감쇠 파라미터는 오류역전파 학습과 Gauss-Newton 학습의 스위치 역할을 하며 학습 속도에 영향을 준다. 이런 감쇠 파라미터를 고정시키는 것은 오차 함수의 진동을 유발하고 학습 속도를 감소시킨다. 따라서 본 논문은 오차 함수의 변화 과정을 참조하여 감쇠 파라미터를 가변적으로 적용하는 방법을 제안한다. 제안된 방법은 오차의 변화량이 크면 감쇠 파라미터를 크게, 오차의 변화량이 작으면 감쇠 파라미터를 작게 조정한다. 이것은 모멘텀과 유사한 역할을 하여 학습 속도를 향상시킨다. 제안된 방법의 검증을 위한 실험으로는 iris 분류 문제와 wine 분류 문제를 사용하였다. 제안된 방법은 iris 분류 문제에서는 67% 학습에서, wine 분류 문제에서는 78% 학습에서 학습 속도가 향상되었으며 기존 방법과 비교하여 오차의 진동도 적은 것을 확인할 수 있었다.

• KIEE International Transaction on Electrical Machinery and Energy Conversion Systems
• /
• 제4B권4호
• /
• pp.206-210
• /
• 2004

This paper deals with a parameter estimation of the DC-DC converter system for its diagnosis. Especially, we present the results of parameter estimation for the DC-DC converter model by the system identification method. The parameter estimation for the DC-DC converter system aims at the diagnosis of its operating status. For the operating status diagnosis of the DC-DC converter system, we assume that the DC-DC converter system is an equivalent model of the Buck converter and estimate the main parameter for on-line diagnosis. In addition, for verification of an estimated parameter, we compare a bode plot of the estimated system transfer function and measurement results of the HP4194 instrument. It is a control system analyzer for system transfer function measurement. Our results confirm that the main parameter for diagnosis of the DC-DC converter system can be estimated by the system identification method and that the aging status of the system can be predicted by these results on operating status.

• 한국군사과학기술학회지
• /
• 제20권4호
• /
• pp.520-527
• /
• 2017

In this study, the experimental methods are compared for obtaining the parameters of the Johnson-Cook constitutive model. The parameters used for numerical simulation are very important in making an accurate estimation of numerical simulation. So, the testing method of obtaining the parameters is also very important. We compared the difference of conventional method, compression method and tensile method of AISI-4340 steel at various strain rate by using MTS, SHPB and SHTB. Taylor impact test and M&S were carried out to compare differences among these three types of JC constitutive parameter.

• Journal of Electrical Engineering and Technology
• /
• 제5권4호
• /
• pp.597-605
• /
• 2010

This paper presents a contribution to parameter identification of a non-linear system using a new strategy to improve the genetic algorithm (GA) method. Since cost function plays an important role in GA-based parameter identification, we propose to improve the simple version of GA, where weights of the cost function are not taken as constant values, but varying along the procedure of parameter identification. This modified version of GA is applied to the induction motor (IM) as an example of nonlinear system. The GA cost function is the weighted sum of stator current and rotor speed errors between the plant and the model of induction motor. Simulation results show that the identification method based on improved GA is feasible and gives high precision.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1997년도 한국자동제어학술회의논문집; 한국전력공사 서울연수원; 17-18 Oct. 1997
• /
• pp.1432-1435
• /
• 1997

This paper proposes a fault detection and isolation metho by combining the parameter estimation method[4] with the observer method[2] to use merits of both methods. To verify the performance of the method proposed some simulations applied to remotely piloted vehicle are performed.

• Journal of Power Electronics
• /
• 제8권4호
• /
• pp.354-362
• /
• 2008

This paper presents neural load torque compensation method which is composed of a deadbeat load torque observer and gains compensation by a parameter estimator. As a result, the response of the PMSM (permanent magnet synchronous motor) obtains better precision position control. To reduce the noise effect, the post-filter is implemented by a MA (moving average) process. The parameter compensator with an RLSM (recursive least square method) parameter estimator is adopted to increase the performance of the load torque observer and main controller. The parameter estimator is combined with a high performance neural load torque observer to resolve problems. The neural network is trained in online phases and it is composed by a feed forward recall and error back-propagation training. During normal operation, the input-output response is sampled and the weighting value is trained multi-times by the error back-propagation method at each sample period to accommodate the possible variations in the parameters or load torque. As a result, the proposed control system has a robust and precise system against load torque and parameter variation. Stability and usefulness are verified by computer simulation and experiment.

• 전기학회논문지
• /
• 제59권4호
• /
• pp.788-792
• /
• 2010

The problem of delay-dependent and parameter-dependent robust stability condition for discrete-time uncertain singular systems with polytopic uncertainty and interval time-varying delay is considered. A new robust stability condition based on parameter-dependent Lyapunov function is derived in terms of LMI (linear matrix inequality). Moreover, the proposed robust stability condition is a general condition for both singular and non-singular systems. A numerical example is presented to demonstrate the effectiveness of the proposed method.

• 전기학회논문지P
• /
• 제63권4호
• /
• pp.290-294
• /
• 2014

PID(Proportional, Integral, Derivative) controller is the most popular process controllers in nuclear power plants. The optimized parameter setting of the process controller contributes to the stable operation and the efficiency of the operating nuclear power plants. PID parameter setting is tuned when new process control systems are installed or current process control systems are changed. When the nuclear plant is shut down, a lot of PID tuning methods such as the Trial and Error method, Ultimate Oscillation method operation, Ziegler-Nichols method, frequency method are used to tune the PID values. But inadequate PID parameter setting can be the cause of the unstable process of the operating nuclear power plant. Therefore the results of PID parameter setting should be simulated, optimized and finally verified. This paper introduces the simulation method of PID tuning to optimize the PID parameter setting and confirms them of the actual PID controller in the operating nuclear power plants. The simulation method provides the accurate process modeling and optimized PID parameter setting of the multi-loop control process in particular.

• 한국농공학회논문집
• /
• 제51권4호
• /
• pp.49-55
• /
• 2009

This paper is to induce design floods through L-moment with 3 and 4 parameter Kappa distributions including test of independence by Wald-Wolfowitz, homogeneity by Mann-Whitney and outlier by Grubbs-Beck on annual maximum flood flows at 9 water level gaging stations in Han, Nakdong and Geum Rivers of South Korea. After analyzing appropriateness of the data of annual maximum flood flows by Kolmogorov-Smirnov test, 3 and 4 Kappa distributions were applied and the appropriateness was judged. The parameters of 3 and 4 Kappa distributions were estimated by L-moment method and the design floods by water level gaging station was calculated. Through the comparative analysis using the relative root mean square errors (RRMSE) and relative absolute errors (RAE) calculated by 3 and 4 parameter Kappa distributions with 4 plotting position formulas, the result showed that the design floods by 4 parameter Kappa distribution with Weibull and Cunnane plotting position formulas are closer to the observed data than those obtained by 3 parameter Kappa distribution with 4 plotting position formulas and 4 parameter Kappa distribution with Hazen and Gringorten plotting position formulas.

• 호남수학학술지
• /
• 제36권4호
• /
• pp.755-766
• /
• 2014

In the discrete formulation of the bubble stabilized Legendre Galerkin methods, the system of equations includes the artificial viscosity term as the parameter. We investigate the estimation of this parameter to get the optimal solution which minimizes the maximum error. Some numerical results are reported.