• 한국지능시스템학회논문지
• /
• 제24권5호
• /
• pp.536-541
• /
• 2014

기상레이더를 통해 취득된 데이터에는 지형에코, 파랑에코, 이상에코, 그리고 청천에코등이 존재한다. 각 에코는 여러 종류의 비강수에코이고, 이 비강수에코를 제거하기 위해 각 에코들의 특성을 분석하였다. 기상레이더 데이터는 매우 방대한 양이기 때문에 전처리 절차를 통해 분석된다. 본 논문에서는 클러스터링 기반 방사형 기저함수 신경회로망(RBFNNs : Radial Basis Function Neural Networks)과 에코 판단 모듈을 이용하여 기상레이더 데이터에서 강수에코와 비강수에코들을 구별하기 위한 에코 패턴분류기를 설계하였다. HCM(Hard C-Mean) 클러스터링 기반 RBFNNs 와 FCM(Fuzzy C-Mean) 클러스터링 기반 RBFNNs를 이용하여 출력성능은 비교 및 분석된다.

• 한국유체기계학회 논문집
• /
• 제18권6호
• /
• pp.12-18
• /
• 2015

Recently, the development trend of turbomachinery is high capacity and high efficiency. Most of turbomachinery in the market are adopting ball bearings or air foil bearings. However, ball bearings have a limit for high speed product over $2.0{\times}10^6DN$(product of the inner diameter of the bearing in mm (D) and the maximum speed in rpm (N)). Air foil bearings have a limit for high axial load for high power products over 200~300 HP(horse power). Magnetic bearing is one of the solutions to overcome the limits of high speed and high axial load. Because magnetic bearings have no friction between the rotor and the bearings, they can reduce the load of the motor and make it possible to increase the rotating speed up to $5.0{\times}10^6DN$. Moreover, they can have high axial load capacity, because the axial load capacity of magnetic bearing depends on the capacity of the designed electromagnet. In this study, the radial and thrust magnetic bearings are designed to be applied to the 200 HP class turbo blower, and their performance was evaluated by the experiment. Based on the tests up to 26,400 rpm and 21,000 rpm under the no-load and load condition, respectively, it was verified that the magnetic bearings are stably support the rotor of the turbo blower.

• 한국지능시스템학회논문지
• /
• 제21권1호
• /
• pp.12-18
• /
• 2011

본 논문에서는 RBF 뉴럴 네트워크에서 은닉층 활성함수에 Interval type-2 퍼지개념을 적용한 새로운 RBF 뉴럴 네트워크를 설계하였다. 퍼지 시스템 분야에서 불확실한 정보에 대한 Type-1 퍼지집합의 성능을 보안하고자 Type-2 퍼지집합이 제안되었으며, 멤버쉽함수 안에 다시 멤버쉽함수를 생성함으로써 불확실한 정보를 좀 더 효과적으로 다루고자 하였다. 따라서 본 논문에서는 RBF 뉴럴 네트워크의 은닉층 활성함수에 type-2 퍼지집합의 개념을 적용하여 불확실한 정보에 대한 모델 성능을 개선하고자 하였다. 나아가 연결가중치를 상수항이 아닌 1차식으로 구성된 다항식을 사용하여 최종출력을 입력-출력의 관계식으로 표현하였다. 연결가중치는 기존의 경사하강법(Gradient Descent Method; GDM) 대신 conjugate gradient method(CGM)을 사용하여 파라미터를 동조하고, 은닉층의 활성함수는 공간탐색 진화 알고리즘(Space Search Evolutionary Algorithm; SSEA)을 이용하여 가우시안 함수의 중심점 및 분포상수를 동조하여 모델의 성능을 개선시킨다. 제안된 모델의 성능을 평가하기 위해 가스로 시계열 데이터를 사용하였으며, 결과를 기존 모델과 비교하였다.

• Journal of Electrical Engineering and Technology
• /
• 제8권2호
• /
• pp.252-261
• /
• 2013

In this paper, a nonlinear adaptive damping controller based on radial basis function neural network (RBFNN), which can infinitely approximate to nonlinear system, is proposed for thyristor controlled series capacitor (TCSC). The proposed TCSC adaptive damping controller can not only have the characteristics of the conventional PID, but adjust the parameters of PID controller online using identified Jacobian information from RBFNN. Hence, it has strong adaptability to the variation of the system operating condition. The effectiveness of the proposed controller is tested on a two-machine five-bus power system and a four-machine two-area power system under different operating conditions in comparison with the lead-lag damping controller tuned by evolutionary algorithm (EA). Simulation results show that the proposed damping controller achieves good robust performance for damping the low frequency oscillations under different operating conditions and is superior to the lead-lag damping controller tuned by EA.

• Steel and Composite Structures
• /
• 제27권4호
• /
• pp.427-438
• /
• 2018

In this paper, the single and multi-objective optimization of thin-walled conical tubes with different types of indentations under axial impact has been investigated using surrogate models called metamodels. The geometry of tapered thin-walled tubes has been studied in order to achieve maximum specific energy absorption (SEA) and minimum peak crushing force (PCF). The height, radius, thickness, tapered angle of the tube, and the radius of indentation have been considered as design variables. Based on the design of experiments (DOE) method, the generated sample points are computed using the explicit finite element code. Different surrogate models including Kriging, Feed Forward Neural Network (FNN), Radial Basis Neural Network (RNN), and Response Surface Modelling (RSM) comprised to evaluate the appropriation of such models. The comparison study between surrogate models and the exploration of indentation shapes have been provided. The obtained results show that the RNN method has the minimum mean squared error (MSE) in training points compared to the other methods. Meanwhile, optimization based on surrogate models with lower values of MSE does not provide optimum results. The RNN method demonstrates a lower crashworthiness performance (with a lower value of 125.7% for SEA and a higher value of 56.8% for PCF) in comparison to RSM with an error order of $10^{-3}$. The SEA values can be increased by 17.6% and PCF values can be decreased by 24.63% by different types of indentation. In a specific geometry, higher SEA and lower PCF require triangular and circular shapes of indentation, respectively.

• 대한기계학회논문집B
• /
• 제35권8호
• /
• pp.799-804
• /
• 2011

본 논문에서는 자연채광 시스템 LFLP(Linear Fresnel Light Pipe)와 DBLP(Double Blind Light Pipe)시스템을 비교하였다. LFLP시스템은 평행한 빛을 선형프레넬렌즈를 이용하여 선형 형태로 빛으로 집광하여 자연채광에 이용하는 시스템이며, DBLP시스템은 베네시안 형태의 블라인드를 이용하여 빛을 반사시켜 자연채광에 이용하는 시스템이다. DBLP시스템은 LFLP시스템을 개선한 것으로 시스템 앞쪽에 위치한 블라인드는 태양의 고도에 따른 빛을, 뒤쪽에 위치한 블라인드는 태양의 방위각에 따른 빛을 변광부로 반사시키도록 설계 되었다. DBLP시스템의 변광부는 콘모양으로 이루어 져 있으며 블라인드에 의해 반사된 빛을 산광부로 보내주는 역할을 하며, 산광부로 들어온 빛은 실내조명에 사용된다. 따라서 맑은날(clear sky)을 기준으로 두 시스템의 효율을 비교하면 DBLP시스템이 LFLP시스템보다 세배 높게 나오는 것으로 나타났다.

• 한국추진공학회지
• /
• 제20권2호
• /
• pp.39-45
• /
• 2016

동축형 다공성재 분사기에서는 기체가 분사기 중심을 지나는 액체 제트 주위를 둘러싼 다공성재를 통해 액체 제트를 향해 분사된다. 분사 방법의 차이로 인한 전단동축 분사기와 동축형 다공성재 분사기의 분열 메커니즘 차이를 살펴보기 위해, 전단동축 분사기와 동축형 다공성재 분사기를 사용하여 수류시험과 2-D 축대칭 수치해석을 진행하였다. 같은 유량조건에서의 가시화 이미지와 분무 평균입경을 비교하였으며, 수치해석을 통해 분사기 내부에서의 속도 분포가 액체 제트에 어떤 영향을 끼치는지 고찰하였다. 결과적으로, 기체의 유량이 늘어날수록 분사기 내부 속도가 낮음에도 동축형 다공성재 분사기의 미립화 및 혼합 성능이 전단동축 분사기에 비해 유리함을 확인할 수 있었다.

• Steel and Composite Structures
• /
• 제20권2호
• /
• pp.379-397
• /
• 2016

Due to creep and shrinkage of the concrete core, concrete-filled steel tubular (CFST) arches continue to deform in the long-term under sustained loads. This paper presents analytical investigations of the effects of geometric nonlinearity on the long-term in-plane structural performance and stability of three-pinned CFST circular arches under a sustained uniform radial load. Non-linear long-term analysis is conducted and compared with its linear counterpart. It is found that the linear analysis predicts long-term increases of deformations of the CFST arches, but does not predict any long-term changes of the internal actions. However, non-linear analysis predicts not only more significant long-term increases of deformations, but also significant long-term increases of internal actions under the same sustained load. As a result, a three-pinned CFST arch satisfying the serviceability limit state predicted by the linear analysis may violate the serviceability requirement when its geometric nonlinearity is considered. It is also shown that the geometric nonlinearity greatly reduces the long-term in-plane stability of three-pinned CFST arches under the sustained load. A three-pinned CFST arch satisfying the stability limit state predicted by linear analysis in the long-term may lose its stability because of its geometric nonlinearity. Hence, non-linear analysis is needed for correctly predicting the long-term structural behaviour and stability of three-pinned CFST arches under the sustained load. The non-linear long-term behaviour and stability of three-pinned CFST arches are compared with those of two-pinned counterparts. The linear and non-linear analyses for the long-term behaviour and stability are validated by the finite element method.

• Journal of Biosystems Engineering
• /
• 제34권2호
• /
• pp.95-105
• /
• 2009

A method was proposed which employed control of the drop location of fertilizer particles on a spinner disc to optimize the spread pattern uniformity. The system contained an optical sensor as a feedback mechanism, which measured discharge velocity and location, as well as particle diameters to predict a spread pattern of a single disc. Simulations showed that the feed gate adaptation algorithm produced high quality patterns for any given application rate in the dual disc spreader. The performance of the feed gate control method was assessed using data collected from a Sulky spinner disc spreader. The results showed that it was always possible to find a spread pattern with an acceptable CV lower than 15%, even though the spread pattern was obtained from a rudimentary flat disc with straight radial vanes. A mathematical optimization method was used to find the initial parameter settings for a specially designed experimental spreading arrangement, which included the feed gate control system, for a given flow rate and swath width. Several experiments were carried out to investigate the relationship between the gate opening and flow rate, disc speed and particle velocity, as well as disc speed and predicted landing location of fertilizer particles. All relationships found were highly linear ($r^2$ > 0.96), which showed that the time-of-flight sensor was well suited as a feedback sensor in the rate and uniformity controlled spreading system.

• 대한의용생체공학회:학술대회논문집
• /
• 대한의용생체공학회 1996년도 추계학술대회
• /
• pp.161-164
• /
• 1996

An actively-shielded $r^{2}$-gradient coil has been developed for brain localized MRI or MRS. Spatial localization is very useful for spatial volume selection in MRI or MR Spectroscopy(MRS). The radial(or $R^{2}-$) gradient coil is useful in reducing the artifact or in improving the SNR by selecting the volume with less number of RF pulses. It is, however, difficult to implement the coil with a gradient intensity strong enough to use it for practical whole-body MRI system. For example, the smallest volume size for selection is just 6 cm in diameter with a 250 Ampere of current driving for a whole-body system (in case of 70-cm-diameter). In this study, an asymetric $r^{2}$-coil with a small diameter of 35 cm has been designed and implemented for brain localized MRI or MRS. An 8-rod high-pass-type birdcage RF coil has also been implemented. The coil set has been developed for 1.0 Tesla Medison MRI system and its performance has been verified experimentally.