• 한국정밀공학회지
• /
• 제20권5호
• /
• pp.196-203
• /
• 2003

This paper demonstrates that an improved hybrid neural-genetic multimodel parameter estimation algorithm can be applied to the structured system identification of an electro-hydraulic servo system. This algorithm is consists of a recurrent incremental credit assignment (ICRA) neural network and a genetic algorithm, The ICRA neural network evaluates each member of a generation of model and the genetic algorithm produces new generation of model. We manufactured an electro-hydraulic servo system and the improved hybrid neural-genetic multimodel parameter estimation algorithm is applied to the task to find the parameter values, such as mass, damping coefficient, bulk modulus, spring coefficient and disturbance, which minimize total square error.

• 한국지반공학회논문집
• /
• 제19권5호
• /
• pp.253-261
• /
• 2003

샌드드레인공법(SD공법)이나 고치환율의 샌드콤팩션파일공법(SCP공법)과 비교하여 저치환율SCP공법에 의해 개량된 복합지반의 역학거동은 모래말뚝과 점토지반 양방의 역학적 상호작용의 영향을 보다 현저하게 받는다. 따라서 본 공법의 적용에 있어서 침하저감효과나 모래말뚝 사이의 점토의 강도증가 등을 정확히 평가하기 위해서는 복합지반 내부에서의 모래말뚝과 점토와의 역학적 상호작용을 해명하여야 할 필요성이 있다. 본 연구에서는 저치환율의 모래말뚝이 타설된 복합지반에 대한 일련의 모형실험을 통하여 압밀중에 발생되는 모래말뚝과 점토 각각의 역학거동과 복합지반 내부에 대한 응력분담거동에 대하여 고찰하였다.

• Journal of Information Processing Systems
• /
• 제14권4호
• /
• pp.941-947
• /
• 2018

Since the amplitudes of spin echo train in nuclear magnetic resonance logging (NMRL) are small and the signal to noise ratio (SNR) is also very low, this paper puts forward an improved de-noising algorithm based on wavelet transformation. The steps of this improved algorithm are designed and realized based on the characteristics of spin echo train in NMRL. To test this improved de-noising algorithm, a 32 points forward model of big porosity is build, the signal of spin echo sequence with adjustable SNR are generated by this forward model in an experiment, then the median filtering, wavelet hard threshold de-noising, wavelet soft threshold de-noising and the improved de-noising algorithm are compared to de-noising these signals, the filtering effects of these four algorithms are analyzed while the SNR and the root mean square error (RMSE) are also calculated out. The results of this experiment show that the improved de-noising algorithm can improve SNR from 10 to 27.57, which is very useful to enhance signal and de-nosing noise for spin echo train in NMRL.

• Journal of applied mathematics & informatics
• /
• 제30권3_4호
• /
• pp.489-498
• /
• 2012

Most of the software reliability models are based on black box approach and these models consider the entire software system as a single unit. Present day software development process has changed a lot. In present scenario these models may not give better results. To overcome this problem an improved additive model has been proposed in this paper, to estimate the reliability of software with modular structure. Also the concept of imperfect debugging has been also considered. A maximum likelihood estimation technique has been used for estimating the model parameters. Comparison has been made with an existing model. ${\chi}^2$ goodness of fit has been used for model fitting. The proposed model has been validated using real data.

• 한국전산구조공학회:학술대회논문집
• /
• 한국전산구조공학회 2000년도 봄 학술발표회논문집
• /
• pp.69-76
• /
• 2000

In this study, an improved analysis model for the more efficient and accurate structural analysis of cable-stayed bridges is presented. In this model, beam elements, of which stability functions are stabilized by the use of Taylor's series expansions, are used to model space frame structures, and truss elements, of which equivalent elastic moduli are evaluated on the assumption that the deflected shape of a cable has a catenary function, are used to model cables. By using the proposed analysis model, nonlinear static analysis and natural vibration analysis of 2-dimensional and 3-dimensional cable-stayed bridges are carried out and are compared with the analysis results reported by other researchers.

• 한국주거학회논문집
• /
• 제16권3호
• /
• pp.37-47
• /
• 2005

The Chosun Exhibition was held in 1929 and the Architectural Association of Chosun(AAC) displayed model houses fur the fair. These model houses represented the new housing trend of the Chosun society, so this research was necessary to find out the facts. The main materials used for this research were issues from 'Chosun and Architecture'. The objects of this research was to find out the activity of the AAC and to classify the displayed model houses according to the structure, plan and interior. These three model houses depicted the improved housing designs at that time. Improved housing designs also depicted the convenience of everyday life, hygiene, improvement of facilities and the application of modem lift to that era. Therefore, these three houses would be revaluated as new models of houses for the Chosun society under education activity by the AAC.

• 한국정밀공학회지
• /
• 제18권9호
• /
• pp.204-209
• /
• 2001

The finite element model for automotive muffler analysis is improved by modifying the boundary condition of outlet and the mesh of finite element model. The model minimizes the difference between transmission loss from analysis and that from experiment. Four different boundary conditions and the four types of finite element model are tested to find out the best one of those. From the case study it is verified that the bevel of transmission loss can be changed by the variation of radiation impedance value. Also the resonance or anti-resonance frequencies of transmission loss can be shifted by the variation of finite element mesh. An improved finite element model of muffler is proposed in consideration with the accuracy and the computing time of analysis.

• Journal of Power Electronics
• /
• 제16권1호
• /
• pp.374-387
• /
• 2016

This paper derives an improved analytical model to estimate switching loss and analyze the effects of parasitic elements on the switching performance of SiC MOSFETs. The proposed analytical model considers the parasitic inductances, the nonlinearity of the junction capacitances and the nonlinearity of the trans-conductance. The turn-on process and the turn-off process are illustrated in detail, and equivalent circuits are derived and solved for each switching transition. The proposed analytical model is more accurate and matches better with experimental results than other analytical models. Note that switching losses calculated based on experiments are imprecise, because the energy of the junction capacitances is not properly disposed. Finally, the proposed analytical model is utilized to account for the effects of parasitic elements on the switching performance of a SiC MOSFET, and the circuit design rules for high frequency circuits are given.

• 응용통계연구
• /
• 제11권2호
• /
• pp.415-421
• /
• 1998

Mangat(1994)는 Mangat-Singh(1990)이 제안한 2단계 관련질문모형의 사용 절차를 좀 더 단순화시킨 개선된 관련질문모형을 제안하였다. 본 논문에서는 Mangat(1994)의 개선된 관련질문모형을 무관질문모형으로 확장하고자 한다. 또한, 제안한 무관질문모형이 Mangat의 개선된 관련질문모형과 Greengerg dt al.(1969)의 무관질문모형보다 효율적임을 보였고, 김종호 외(1992) 2인이 제안한 2단계 무관질문모형보다 효율적이 되는 조건을 제시하였다.

• Communications for Statistical Applications and Methods
• /
• 제30권5호
• /
• pp.501-516
• /
• 2023

Knowledge distillation (KD) refers to extracting knowledge from a large and complex model (teacher) and transferring it to a relatively small model (student). This can be done by training the teacher model to obtain the activation function values of the hidden or the output layers and then retraining the student model using the same training data with the obtained values. Recently, relational KD (RKD) has been proposed to extract knowledge about relative differences in training data. This method improved the performance of the student model compared to conventional KDs. In this paper, we propose a new method for RKD by introducing a new loss function for RKD. The proposed loss function is defined using the area difference between the teacher model and the student model in a specific hidden layer, and it is shown that the model can be successfully compressed, and the generalization performance of the model can be improved. We demonstrate that the accuracy of the model applying the method proposed in the study of model compression of audio data is up to 1.8% higher than that of the existing method. For the study of model generalization, we demonstrate that the model has up to 0.5% better performance in accuracy when introducing the RKD method to self-KD using image data.