• Journal of the Korean Society for Precision Engineering
• /
• v.12 no.6
• /
• pp.120-127
• /
• 1995

A nonlinear empirical state-space model of the Artificial Neural Network(ANN) has been developed. The nonlinear model structure incorporates characteristic, so as to enable identification of the transient response, as well as the steady-state response of a dynamic system. A hybrid feedfoward/feedback neural network, namely a Local Time Delayed Recurrent Multi-layer Perception(RMLP), is the model structure developed in this paper. RMLP is used to identify nonlinear dynamic system in an input/output sense. The feedfoward protion of the network architecture provides with the well-known curve fitting factor, while local recurrent and cross-talk connections provides the dynamics of the system. A dynamic learning algorithm is used to train the proposed network in a supervised manner. The derived dynamic learning algorithm exhibit a computationally desirable characteristic; both network sweep involved in the algorithm are performed forward, enhancing its parallel implementation. RMLP state-space and its associate learning algorithm is demonstrated through a simple examples. The simulation results are very encouraging.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.21 no.1
• /
• pp.43-50
• /
• 2013

The stringent emission regulation and future shortage of fossil fuel motivate the research of alternative powertrain. In this study, a system of proton exchange membrane fuel cell has been modeled to analyze the performance of the fuel cell system for automotive application. The model is composed of the fuel cell stack, air compressor, humidifier, and intercooler, and hydrogen supply which are implemented by using the Matlab/Simulink(R). Fuel cell stack model is empirical model but the water transport model is included so that the system performance can be predicted over various humidity conditions. On the other hand, the model of air compressor is composed of motor, static air compressor, and some manifolds so that the motor dynamics and manifold dynamics can be investigated. Since the model is concentrated on the strategic operation of compressor to reduce the power consumption, other balance of components (BOP) are modeled to be static components. Since the air compressor model is empirical model which is based on curve fitting of experiments, the stack model is validated with the commercial software and the experiments. The dynamics of air compressor is investigated over unit change of system load. The results shows that the power consumption of air compressor is about 12% to 25% of stack gross power and dynamic response should be reduced to optimize the system operation.

• Journal of IKEEE
• /
• v.18 no.2
• /
• pp.240-247
• /
• 2014

Intraocular pressure is measured after a cornea air puff by observing biomechanical properties such as thickness or displacement of the cornea. In this paper, we deal with a frequency domain analysis of corneal deformation in the air puff tonometry that is used to diagnose glaucoma or lasik. We distinguish the patient from the normal by measuring the oscillation frequency in the neighborhood of the central cornea section. A binary image was obtained from the video images, and cornea vertical oscillation profile was extracted from the difference between the vertical displacement data and the curve fitting. In terms of Fourier transform, a vibration frequency of 479.2Hz for the patient was obtained as well as more higher 702.8Hz for the normal due to stiffness. Hilbert-Huang transform's empirical mode decomposition generally describes local, nonlinear, and nonstationary data. After the data were decomposed into intrinsic mode functions, a spectrum and power were analysed. Finally, we confirm that the patient has 6 times more higher power ratio for the specific intrinsic mode function between the patient and the normal.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.28 no.3C
• /
• pp.179-186
• /
• 2008

Pressuremeter test estimates the deformational properties of soil from the relationship between applied pressure and the displacement of cavity wall. It is general to utilize the reloading curve for the estimation of deformational properties of soil because the initial loading curve can be affected by the disturbance caused by boring. On the other hand, the instrumental resolution or the variation of measured data makes it hard to estimate the maximum shear modulus from pressuremeter test results. This study suggested the methodology estimating the maximum shear modulus from pressuremeter test directly, based on the curve fitting of reloading curve. In addition, the difference was taken into account between the stress state around the probe in reloading and that of the in-situ state. Pressuremeter tests were conducted for 15 cases using a large calibration chamber, together with a number of reference tests. The maximum shear moduli taken from suggested method were compared with those from empirical correlation and bender element test.

• Journal of Korea Water Resources Association
• /
• v.55 no.10
• /
• pp.843-854
• /
• 2022

For the rip current warning system to reduce rip-current accidents, the implementation method producing the risk index was modified. To produce fast response from the warning system based on real-time observations, the method employed the numerical results (i.e., rip current likelihoods according to the possible scenario) obtained in advance. In this study, instead of using the empirical curve-fitting functions of the previous method, the present modification utilized two-dimensional distributions (i.e., wave height and period, wave height and tidal elevation, wave height and direction, wave height and spreading of frequency-directional spectrum) of rip current likelihoods stacked in a database of the system. The wave and tidal observations in 2021 at the Haeundae coast were applied to the modified system, and its performances at several real events recorded in CCTV images were presented.