• 유공압시스템학회논문집
• /
• 제5권2호
• /
• pp.20-26
• /
• 2008

The most vehicle active suspension system is activated by a hydraulic source and transmission system which has nonlinear characteristics. Even though we have designed a proper controller for this system, it sometimes cannot show remarkable performance characteristics because of many factors that undercut the performance of the hydraulic system, such as nonlinearity, modelling errors, parameter variations etc. So, the robust controller that prevents a system from lowering its performance is needed. In this study, the sliding mode control which is the representative one of robust controllers is adopted to investigate system parameter sensibility. As a result, the sliding mode controller shows robustness to the system parameters variations relative to the other controllers.

• 대한기계학회논문집
• /
• 제18권9호
• /
• pp.2312-2321
• /
• 1994

In this work the dynamics of an electromagnetic levitation system is described by a set of three first order nonlinear ordinary differential equations. The objective is to design a digital linear controller which takes the inherent instability of the uncontrolled system and the disturbing force into consideration. The controller is made by employing digital linear quadratic(LQ) design methodology and the unknown state variables are estimated by the kalman filter. The state estimation is performed using not only an air gap sensor but also both an air gap sensor and a piezoelectric accelerometer. The design scheme resulted in a digital linear controller having good stability and performance robustness in spite of various modelling errors. In case of using both a gap sensor and an accelerometer for the state estimation, the control input was rather stable than that in a system with gap sensor only and the controller dealt with the disturbing force more effectively.

• Journal of Mechanical Science and Technology
• /
• 제20권7호
• /
• pp.917-928
• /
• 2006

In recent decades, Artificial Neural Networks (ANNs) have become the focus of considerable attention in many disciplines, including robot control, where they can be used to solve nonlinear control problems. One of these ANNs applications is that of the inverse kinematic problem, which is important in robot path planning. In this paper, a neural network is employed to analyse of inverse kinematics of PUMA 560 type robot. The neural network is designed to find exact kinematics of the robot. The neural network is a feedforward neural network (FNN). The FNN is trained with different types of learning algorithm for designing exact inverse model of the robot. The Unimation PUMA 560 is a robot with six degrees of freedom and rotational joints. Inverse neural network model of the robot is trained with different learning algorithms for finding exact model of the robot. From the simulation results, the proposed neural network has superior performance for modelling complex robot's kinematics.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2001년도 춘계학술대회논문집
• /
• pp.27-33
• /
• 2001

Linear harmonic analysis is a convenient and generally accurate method to use for the acoustic analysis of intake and exhaust silencers for IC engines. The major uncertainty in this form of modelling is the characterisation of the source, which is inherently nonlinear and time-variant. Experimental methods are generally used to determine the source characteristics, and in particular the indirect method is most suitable for an IC-engine source. With reference to an idealised linear time-variant source, it is found that the characteristics of a time-variant source as determined by the indirect method have no physical relevance. The direct method of experimental measurement appears to have some advantage over the indirect method, although in practice it is difficult to apply to an IC engine source. Again, an idealised linear time-variant source can be used to indicate that the characteristics of a time-variant source as determined by the direct method also have no physical relevance. Strangely, these meaningless measured source properties can nevertheless be used to accurately predict the radiated noise from an IC engine and silencer system.

• Smart Structures and Systems
• /
• 제14권4호
• /
• pp.643-658
• /
• 2014

This paper presents the design, modelling and, simulation and experimental results of a shape memory alloy (SMA) actuator based critical motion control application. Dynamic performance of SMA and its ability in replacing servo motor is studied for which the famous open loop unstable balancing ball and beam system direct driven by antagonistic SMA is designed and developed. Simulation uses the mathematical model of ball and beam structure derived from the first principles and model estimated for the SMA actuator by system identification. A PID based cascade control system consisting of two loops is designed and control of ball trajectory for various target positions with settling time as control parameter is verified experimentally. The results demonstrate the performance of SMA for a complicated i.e., under actuated, highly nonlinear unstable system, and thereby it's dynamic behaviour. Control strategies bring out the effectiveness of the actuator and its possible application to much more complex applications such as in aerospace control and robotics.

• Structural Engineering and Mechanics
• /
• 제63권2호
• /
• pp.207-215
• /
• 2017

The paper presents the studies carried out on low velocity impact of Ultra high performance concrete (UHPC) panels of size $350{\times}350{\times}10mm^3$ and $350{\times}350{\times}15mm^3$. The panels are cast with 2 and 2.5% micro steel fibre and compared with UHPC without fiber. The panels are subjected to low velocity impact, by a drop-weight hemispherical impactor, at three different energy levels of 10, 15 and 20 J. The impact force obtained from the experiments are compared with numerically obtained results using finite element method, theoretically by energy balance approach and empirically by nonlinear multi-genetic programming. The predictions by these models are found to be in good coherence with the experimental results.

• 전력전자학회:학술대회논문집
• /
• 전력전자학회 2002년도 전력전자학술대회 논문집
• /
• pp.33-37
• /
• 2002

This paper presents a simple torque estimation method and the switching angle control of SRM using Neural Network. SRM has gaining much interest as industrial applications due to the simple structure and high efficiency. Adaptive switching angle control is essential for the optimal driving of a SRM because of the driving characteristic varies with the load and speed. The proper switching angle which can increase the efficiency was investigated in this paper Neural Network was adapted to regulate the switching angle and nonlinear inductance modelling. Experimental result shows the validity of the switching angle controller.

• 한국해안해양공학회지
• /
• 제1권1호
• /
• pp.87-92
• /
• 1989

극지에서의 충돌 빙하중의 크기를 구하기 위해 지반, 구조 및 구조물을 일체로 하는 동적 모델을 제시하였다. 지반은 감쇠 효과를 고려한 비선형 스프링으로 이상화하였으며, 부빙과 구조물의 상호작용은 얼음의 특성을 고려한 비선형 스프링으로 이상화하였다. 이 모델에 의한 여름과 겨울의 부빙의 질량과 속도에 따른 영향을 모래지반과 시일토지반에 대해 각각 해석하여 에너지 평형 방법과 비교하였다. 해석 결과 모래지반의 경우는 대체적으로 에너지 평형 방법에 의한 값이 동적 모델에 의한 값보다 작게 나타났으며, 시일 토지반의 경우는 그 반대로 나타났다.

• Coupled systems mechanics
• /
• 제7권2호
• /
• pp.111-139
• /
• 2018

A successful methodology for modelling controlled destruction and progressive collapse of 2D reinforced concrete frames is presented in this paper. The strategy is subdivided into several aspects including the failure mechanism creation, and dynamic motion in failure represented with multibody system (MBS) simulation that are used to jointly capture controlled demolition. First phase employs linear elasto-plastic analysis with isotropic hardening along with softening plastic hinge concept to investigate the complete failure of structure, leading to creation of final failure mechanism that behaves like MBS. Second phase deals with simulation and control of the progressive collapse of the structure up to total demolition, using the nonlinear dynamic analysis, with conserving/decaying energy scheme which is performed on MBS. The contact between structure and ground is also considered in simulation of collapse process. The efficiency of the proposed methodology is proved with several numerical examples including six story reinforced concrete frame structures.

• Computers and Concrete
• /
• 제1권2호
• /
• pp.151-168
• /
• 2004

A two-dimensional nonlinear finite element model is developed to simulate time-dependent cracking of reinforced concrete members under service loads. To predict localized cracking, the crack band model is employed to model individual crack opening. In conjunction with the crack band model, a bond-interface element is used to model the slip between concrete and reinforcing steel permitting large slip displacements between the concrete element nodes and the steel truss element nodes at crack openings. The time-dependent effects of concrete creep and shrinkage are incorporated into the smeared crack model as inelastic pre-strains in an iterative solution procedure. Two test examples are shown to verify the finite element model with good agreement between the model and the observed test results.