• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2005년도 추계학술대회논문집
• /
• pp.963-966
• /
• 2005

Model updating method is known to the area to correct finite element models by the results of the experimental modal analysis. Most common methods in model updating depend on a parametric model of the structure. In this case, the number of parameters is normally smaller than that of modal data obtained from an experiment. In order to overcome this limitation, many researchers are trying to get modal data as many as possible to date. 1 want to name this method multiple modified-system generation method. These Methods consist of direct system modification method and feedback controller method. The direct system modification Is to add a mass or stiffness on the original structure or perturb the boundary conditions. The feedback controller method is to make the closed food system with sensor and actuator so as to get the closed loop modal data. In this paper, we need to focus on the feedback controller method because of its simplicity. Several methods related the feedback controller methods are virtual passive controller (VPC) sensitivity enhancement controller (SEC) and mode decoupling controller (MDC). Among them, we will apply MDC to the model updating problem. MDC has various advantages compared with other controllers, such as VPC and SEC. To begin with, only the target mode can be changed without changing modal property of non-target modes. In addition, it is possible to fix any modes if the number of sensors is equal to that of the system modes. Finally, the required control power to achieve desired change of target mode is always lower than those of other methods such as VPC. However, MDC can make the closed loop system unstable when using incomplete modal data. So we need to take action to avoid undesirable instability from incomplete modal data. In this paper, we address the method to design the unique and robust MDD obtained from incomplete modal data. The associated simulation will be Incorporated to demonstrate the usefulness of this method.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2000년도 추계학술대회논문집
• /
• pp.282-286
• /
• 2000

In this paper, magneto-rheological(MR) damper is studied for vibration control of large infra structures under earthquake. Generally, active control devices need a large control force and a high power supply system to reduce the vibration effectively. Large and miss tuned control force may induce the dangerous situation such that the generated large control force acts to amplify the structural vibration. Recently, to overcome the weaknesses of the active control, the semi-active control method is suggested by many researchers. Semi-active control uses the passive control device of which the characteristics can be modified. Control force of the semi-active device is not generated from the actuator with power supply. It is generated as a dynamic reaction force of the device same as in the passive control case, so the control system is inherently stable and robust. Unlike the case of passive control, control force of semi-active control is adjusted depending on the measured response of the structure, so the vibration can be reduced more effectively against various unknown environmental loads. Magneto-rheological(MR) damper is one of the semi-active devices. Dynamic characteristics of the MR material can be changed by applying the magnetic fields. So the control of MR damper needs only small power. Response time of MR to the input voltage is very short, so the high performance control is possible. MR damper has a high force capacity so it is adequate to the vibration control of large infra structure. Because MR damper has a nonlinear property, normal control method used in active control may not be effective. Clipped optimal control, modified bang-bang control etc. have been suggested to MR damper by many researchers. In this study, sliding mode fuzzy control(SMFC) is applied to MR damper. Genetic algorithm is used for the controller tuning. To verify the applicability of MR damper and suggested algorithm, numerical simulation on the aseismic control is carried out. Simulation model is three-story building structure, which was used in the paper of Dyke, et al. The control performance is compared with clipped optimal control. The present results indicate that the SMFC algorithm can reduce the earthquake-induced vibration very effectively.

• Structural Engineering and Mechanics
• /
• 제85권4호
• /
• pp.469-484
• /
• 2023

A deep recursive bidirectional Cuda Deep Neural Network Long Short Term Memory (Bi-CuDNNLSTM) layer is recruited in this paper to predict the entire force time histories, and the corresponding hysteresis and backbone curves of reinforced concrete (RC) bridge piers using experimental fast and slow cyclic tests. The proposed stacked Bi-CuDNNLSTM layers involve multiple uncertain input variables, including horizontal actuator displacements, vertical actuators axial loads, the effective height of the bridge pier, the moment of inertia, and mass. The functional application programming interface in the Keras Python library is utilized to develop a deep learning model considering all the above various input attributes. To have a robust and reliable prediction, the dataset for both the fast and slow cyclic tests is split into three mutually exclusive subsets of training, validation, and testing (unseen). The whole datasets include 17 RC bridge piers tested experimentally ten for fast and seven for slow cyclic tests. The results bring to light that the mean absolute error, as a loss function, is monotonically decreased to zero for both the training and validation datasets after 5000 epochs, and a high level of correlation is observed between the predicted and the experimentally measured values of the force time histories for all the datasets, more than 90%. It can be concluded that the maximum mean of the normalized error, obtained through Box-Whisker plot and Gaussian distribution of normalized error, associated with unseen data is about 10% and 3% for the fast and slow cyclic tests, respectively. In recapitulation, it brings to an end that the stacked Bi-CuDNNLSTM layer implemented in this study has a myriad of benefits in reducing the time and experimental costs for conducting new fast and slow cyclic tests in the future and results in a fast and accurate insight into hysteretic behavior of bridge piers.

• 한국산학기술학회논문지
• /
• 제13권3호
• /
• pp.1255-1262
• /
• 2012

현재 가동 중인 월성 원자력 발전의 핵연료 교체로봇 시스템을 살펴보면 핵연료 교환에 필요한 구동압력 제어를 위해 PI제어기를 사용한다. PI제어는 구조가 간단하고 이득 설정을 통해 시스템 요구조건에 만족하는 제어 성능을 낼 수 있지만 밸브와 관로 등의 파라미터 변화로부터 적절한 이득 변경 없이 안정한 제어가 힘들다. 이러한 문제를 해결하기 위해 PI제어기 이득을 동적으로 변경 하거나 PI제어기 출력을 보상하도록 제어기를 구성하는 것이 바람직하다. 본 연구개발의 목적은 파라미터 변화에도 안정한 제어가 가능하도록 제어기를 설계하여 오차와 진동현상을 줄이는데 있다. 제안한 PI/NN제어 기법은 PI제어기와 신경회로망 제어기를 병렬 결합한 구조로 신경회로망 제어기가 PI제어기 출력을 보상하여 파라미터 변화에 강인하도록 설계 하였다. 제어기의 성능평가를 위해 직접 실 공정에 테스트하기가 힘들기 때문에 공정의 특성을 반영하여 모델링한 시뮬레이터를 개발하였고, 시뮬레이션 결과를 실 공정데이터와 비교하여 공정 특성을 모사함을 보였으며, 파라미터 변화에 PI/NN제어기가 오차 및 진동현상을 줄이는 것을 확인 하였다. 또한, 실 공정에서 사용 중인 PI제어기를 주 제어기로 사용하면서 파라미터 변화에 대한 비선형성을 보상하는 제어기 역할을 하기 때문에 신경회로망을 단독으로 사용하였을 때 보다 더 신뢰성 있고 안정적인 제어가 가능하다.

• 대한기계학회논문집A
• /
• 제41권7호
• /
• pp.575-583
• /
• 2017

본 논문에서는 고속도로의 합류지점 상황에서 자율주행을 위한 운전 모드 결정 알고리즘의 개발 및 평가를 진행하였다. 합류 상황을 위한 자율주행 알고리즘 개발에 있어 적절하게 합류를 결정하는 운전 모드 결정이 필수적이다. 운전자 모드는 총 2가지로 차선 유지, 차선 변경(합류)이다. 합류 모드 결정은 주변 차량의 정보 및 합류 차선에 남은 거리를 기반으로 결정된다. 합류 모드 결정 알고리즘에서는 합류 가능 여부를 판단하고 합류가 가능할 때, 안전하고 빠르게 합류하기 위한 최적의 위치를 찾는다. 안전 주행 영역은 주변 차량의 정보 및 주행 모드를 기반으로 정의된다. 안전 주행 영역으로 자율주행 차량을 유지하기 위한 조향각과 종방향 가속도를 얻기 위해 여러 제한 조건이 더해진 강건 모델 예측기법이 사용되었다. 본 논문에서 제안된 알고리즘은 컴퓨터 시뮬레이션을 이용해 검증되었다.