• Journal of Institute of Control, Robotics and Systems
• /
• v.5 no.2
• /
• pp.189-199
• /
• 1999

This paper presents an optimization algorithm for a stable Self Dynamic Neural Network(SDNN) using genetic algorithm. Optimized SDNN is applied to a problem of controlling nonlinear dynamical systems. SDNN is dynamic mapping and is better suited for dynamical systems than static forward neural network. The real-time implementation is very important, and thus the neuro controller also needs to be designed such that it converges with a relatively small number of training cycles. SDW has considerably fewer weights than DNN. Since there is no interlink among the hidden layer. The object of proposed algorithm is that the number of self dynamic neuron node and the gradient of activation functions are simultaneously optimized by genetic algorithms. To guarantee convergence, an analytic method based on the Lyapunov function is used to find a stable learning for the SDNN. The ability and effectiveness of identifying and controlling a nonlinear dynamic system using the proposed optimized SDNN considering stability is demonstrated by case studies.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.14 no.2
• /
• pp.324-330
• /
• 1990

In this study, in order to perform dynamic design of machine tools reasonably and effectively, a method was formulated to be applicable to the damped structures connected by joints having elasticity and damping by using substructure synthesis method. And a nonlinear solution method was proposed and it formulates the nonlinear parts by describing functions and uses the reducing transformation matrix by the substructure synthesis method. The results of frequency response analysis of a machine tool, where an NC lathe was partitioned by three parts of spindle, housing and bed-base part and the nonlinearity of bearing parts between spindle and housing was modelled, showed force dependency of the response.

• Structural Engineering and Mechanics
• /
• v.58 no.6
• /
• pp.1109-1126
• /
• 2016

The present paper focuses on size optimization of double layer barrel vaults considering nonlinear behavior. In order to tackle the optimization problem an improved colliding bodies optimization (ICBO) algorithm is proposed. The important task that should be achieved before optimization of structural systems is to determine the best form having the least cost. In this study, an attempt is done to find the best form then it is optimized considering linear and non-linear behaviors. In the optimization process based on nonlinear behavior, the geometrical and material nonlinearity effects are included. A large-scale double layer barrel vault is presented as the numerical example of this study and the obtained results indicate that the proposed ICBO has better computational performance compared with other algorithms.

• Earthquakes and Structures
• /
• v.3 no.6
• /
• pp.869-887
• /
• 2012

Ground motion scaling techniques are actively debated in the earthquake engineering community. Considerations such as what amplitude, over what period range and to what target spectrum are amongst the questions of practical importance. In this paper, the effect of various ground motion scaling approaches are explored using three reinforced concrete prototypical building models of 8, 12 and 20 stories designed to respond nonlinearly under a design level earthquake event in the seismically active Southern California region. Twenty-one recorded earthquake motions are selected using a probabilistic seismic hazard analysis and subsequently scaled using four different strategies. These motions are subsequently compared to spectrally compatible motions. The nonlinear response of a planar frameidealized building is evaluated in terms of plasticity distribution, floor level acceleration and uncorrelated acceleration amplification ratio distributions; and interstory drift distributions. The most pronounced response variability observed in association with the scaling method is the extent of higher mode participation in the nonlinear demands.

• Journal of Institute of Control, Robotics and Systems
• /
• v.9 no.12
• /
• pp.977-983
• /
• 2003

An output feedback stabilizing controller far non-minimum phase nonlinear systems is presented. We first perform the standard input-output linearization of the system and then transform the zero dynamics into a special normal form in which the antistable part is not affected by the stable part and the antistable part is given in approximately linear form. Under the assumption that the nonlinear system satisfies the observability rank condition, we can design an observer f3r the extended system that is made of the augmentation of a chain of integrators. The proposed output feedback stabilizing controller can then be designed by combining the observer and the state feedback controller.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2006.04a
• /
• pp.1017-1024
• /
• 2006

Capacity spectrum method(CSM) of ATC-40(1996) and displacement coefficient method(DCM) of FEMA-273(1997) are applied to evaluate the seismic performance of bridges. In this study, equivalent response is obtained from nonlinear static analysis for the 3spans continues bridge and nonlinear maximum displacement response is calculated using CSM and DCM. Nonlinear maximum displacement response of DCM is larger than this of CSM. It is method that DCM can evaluate target displacement and ductility of structural to be easy and simple, but tend to overestimate the maximum displacement response. Therefore, this method is mainly used at preparation design level to evaluate the structural response. It is not desirable to evaluate the seismic performance using DCM.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2001.09a
• /
• pp.243-250
• /
• 2001

Industrial machines are sometimes exposed to the danger of earthquake. In the design of a mechanical system, this factor should be accounted for from the viewpoint of reliability. A method to analyze a complex nonlinear structure system under random excitation is proposed. First, the actual random excitation, such as earthquake, is approximated to the corresponding Gaussian process far the statistical analysis. The modal equations of overall system are expanded sequentially. Then, the perturbed equations are synthesized into the overall system and solved in probabilistic way. Several statistical properties of a random process that are of interest in random vibration applications are reviewed in accordance with nonlinear stochastic problem. The obtained statistical properties of the nonlinear random vibration are evaluated in each substructure. Comparing with the results of the numerical simulation proved the efficiency of the proposed method.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2001.09a
• /
• pp.311-319
• /
• 2001

Recently, the number of seismically isolated bridges increased suddenly since the occurrence of strong earthquakes. However, because isolator lies between pier and girder, the response of the superstructure of seismically isolated bridge may be magnified and induce risk of unseating girder Consequently, the response of girder constitutes a crucial factor in designing bridge. In the case of frictional bearing, the inherent nonlinearity makes the use of former linear response spectrum unable to estimate the maximum response of the bearing, and nonlinear tlme history analysis shall be applied. In this paper, nonlinear response spectrum considering frictional element is established, and simple analysis method using such nonlinear spectrum is proposed to estimate the maximum response of the superstructure.

• Transactions on Control, Automation and Systems Engineering
• /
• v.2 no.2
• /
• pp.130-135
• /
• 2000

A two parameter model of a single fictitious weak acid with unknown dissociation constant has been successfully applied to design a neutralization system for many multi-component acid streams. But there are some processes for which above two parameter model is not satisfactory due to poor approxmation of the nonlinearity of pH process. Here, for etter control of wide class of multi-component acid streams, a three parameter model of a strong acid and a weak acid with unknown dissociation constant is proposed. The model approximates effectively three types of largest gain variation nonlinearities. Based on this model a nonlinear pH control system is designed. Parameters can eeasily estimated since their combinations appear linearly in the model equations and nonlinear adaptive control system may also be constructed just as with the two parameter model.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.33 no.8
• /
• pp.779-785
• /
• 2009

This paper is concerned with a computational approach for topology optimization of geometrically nonlinear structures following specific load-displacement trajectories. In our previous works, attention was paid to stabilize topology optimization involving large displacement and a method called the element connectivity parameterization was developed. Here, we aimed to extend the element connectivity parameterization method to find an optimal geometrically nonlinear structure yielding a specific load-displacement trajectory. In contrast to designing a stiffest structure, the trajectory design problem requires special consideration in topology optimization formulation and solution procedure. Some numerical problems were considered to test the developed element connectivity parameterization based formulation.