A new estimation algorithm of physical properties based on a group contribution and support vector machine

En Sup Yoon,이창준,이기백,Won So 한국화학공학회 2008 Korean Journal of Chemical Engineering Vol.25 No.3

There are two ways to evaluate the properties of unknown chemical compounds. One is by traditional approaches, which measure the desired data from the experiments and the other is by predicting them in the theoretical approaches using a kind of prediction model. The latter are considered to be more effective because they are less time consuming and cost efficient, and there is less risk in conducting the experiments. Besides, it is inconvenient to conduct experiments to obtain experimental data, especially for new materials or high molecular substances. Several methods using regression model and neural network for predicting the physical properties have been suggested so far. However, the existing methods have many problems in terms of accuracy and applicability. Therefore, an improved method for predicting the properties is needed. A new method for predicting the physical property was proposed to predict 15 physical properties for the chemicals which consist of C, H, N, O, S and Halogens. This method was based on the group contribution method that was oriented from the assumption that each fragment of a molecule contributes a certain amount to the value of its physical property. In order to improve the accuracy of the prediction of the physical properties and the applicability, we extended the database, significantly modifying the existing group contribution methods, and then established a new method for predicting the physical properties using support vector machine (SVM) which is a statistical theory that has never been used for predicting the physical properties. The SVM-based approach can develop nonlinear structure property correlations more accurately and easily in comparison with other conventional approaches. The results from the new estimation method are found to be more reliable, accurate and applicable. The newly proposed method can play a crucial role in the estimation of new compounds in terms of the expense and time.

Inverse estimation of properties for charring material using a hybrid genetic algorithm

Chang, Hee-Chul,Park, Won-Hee,Yoon, Kyung-Beom,Kim, Tae-Kuk,Lee, Duck-Hee,Jung, Woo-Sung 대한기계학회 2011 JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY Vol.25 No.6

Fire characteristics can be analyzed more realistically by using more accurate material properties related to the tire dynamics and one way to acquire these fire properties is to use one of the inverse property estimation techniques. In this study an optimization algorithm which is frequently applied for the inverse heat transfer problems is selected to demonstrate the procedure of obtaining fire properties of a solid charring material with relatively simple chemical structure. Thermal decomposition is occurred at the surface of the test plate by receiving the radiative energy from external heat sources and in this process the heat transfer through the test plate can be simplified by an unsteady one dimensional problem. The input parameters for the analyses are the surface temperature and mass loss rate of the char plate which are determined from the actual experiment of from the unsteady one-dimensional analysis with a given set of eight properties. The performance of hybrid genetic algorithm (HGA) is compare with a basic genetic algorithm (GA) in order to examine its performance. This comparison is carried out for the inverse property problem of estimating the fire properties related to the reaction pyrolysis of some relatively simple materials; redwood and red oak. Results show that the hybrid genetic algorithm has better performance in estimating the eight pyrolysis properties than the genetic algorithm.

On-Orbit Estimation of Dynamic Properties for STSAT3

Sungwook Yang,Dong-Ik Cheon,Sangchul Lee,Hwa-Suk Oh 제어로봇시스템학회 2008 제어로봇시스템학회 국제학술대회 논문집 Vol.2008 No.10

Reliable and efficient attitude control requires the accurate dynamic properties. On orbit, the mass and inertia values are to be changed by some activities such as the propellant use, the solar panel deployment, etc. To obtain the accurate mass properties, the estimation should be performed. For the initial on-orbit estimation of the mass properties, the gyro-based attitude data is applied for the batch method. The gyro-based attitude data includes the noise, which affects the estimation accuracy. We suggested several filtering methods including the extended Kalman filtering algorithm to obtain the well-conditioned attitude data. Applying the proper filters, we can obtain the reliable estimation result of the mass properties of STSAT3.

A Combined Robust Estimator Between the Least Squares Estimator and a t-type Regression Estimator

정강모 한국자료분석학회 2011 Journal of the Korean Data Analysis Society Vol.13 No.5

When the distribution of the errors in linear regression follows a normal distribution the least squares estimator is most efficient. However, if it follows a heavy-tailed distribution such as t distribution, then the least squares estimator is no longer efficient. We propose a combined estimator between the least squares estimator and a t-type regression estimator which is efficient even if the errors have a heavy-tailed or thin-tailed distributions. We calculated the asymptotic property of the proposed estimator. The results of simulation showed that the proposed estimator is robust and effective in many situations.

Indirect Estimation Based Active Disturbance Rejection Control for Solar Sail Station-Keeping

Jia Huang,Yuliang Bai,James D. Biggs,Naigang Cui 한국항공우주학회 2021 International Journal of Aeronautical and Space Sc Vol.22 No.6

This paper presents an indirect estimation based active disturbance rejection control (ADRC) for solar sail station-keeping. In contrast to previous applications of ADRC to spacecraft station-keeping the proposed approach estimates the deviation of the orbit state from the reference orbit, instead of estimating the orbit state directly. Since the deviation of the orbit state varies more slowly than the state itself, the indirect estimation significantly improves the estimation accuracy and hence the station-keeping performance. Moreover, this approach enables a large control time step which is two orders of magnitude larger than those in previous applications. This is beneficial to the implementation of attitude control and power consumption in practice. In addition, errors in the optical properties of the solar sail are taken into account which is a major challenge for solar sail station-keeping that has not been considered in previous station-keeping control design. The simulation results show that the presented ADRC station-keeping is robust to errors in these optical properties.

Influence of Depth-dependent Rock Properties on Stress Field Determination in KURT Site

Chae-Soon Choi,Yong-Ki Lee,Kyung-Woo Park 한국방사성폐기물학회 2023 한국방사성폐기물학회 학술논문요약집 Vol.21 No.1

In KAERI, a site descriptive model for stress field estimation had already been constructed by using integrated field data within KURT site scale. A sub-divided rock block domain containing major fracture zones has spatial rock mass and fault properties. The properties were decided based on the rock classification results of several borehole investigations. Modeling for maximum and minimum horizontal stress field estimation was performed and compared with the in-situ data. As a result, a depth-dependent stress ratio was adopted to obtain numerical results closer to actual in-situ data. Although the results were suitable at a relatively low depth (~500 m), there is still some deviation trend at a deep depth. This study aims to improve these modeling results by incorporating not only depth-dependent stress ratio but also changes in rock mass properties along the depth. The deep borehole of DB2 in the KURT site indicated fracture distribution corresponding to the property changes. Natural fractures are typically randomly oriented, and the fracture frequency decreases with increasing depth. The increase in P-wave velocity log data accompanies these features. A discrete fracture network (DFN) model can be used to simulate fractured rock explicitly, but DFN modeling is not feasible for site scale analysis because of its numerical efficiency. Therefore, as a preliminary model in this study, the effect of fracture distribution was considered by substituting the influence for the depth-dependent property. The properties were estimated from the fracture frequency and P-wave velocity log data. The influence of elastic modulus and density on the site stress field was dominant, with decreasing the deviation trend between modeling and in-situ data at a deep depth. Considering that the depth of the repository construction is within about 500 m, it may not be necessary to consider the change of rock properties with depth. However, it was determined that the rock property effect might need to be considered when the loading conditions change due to subsidence in the long-term evolution scenario. Continuously, this site descriptive modeling will be interdependently conducted with a representative DFN block model for deriving equivalent properties in fractured rock.

A Follow-Up Study of Psychometric Properties of IRT Proficiency Estimates

김성훈(Seonghoon Kim) 한국교육평가학회 2012 교육평가연구 Vol.25 No.4

본 연구는 Kolen과 Tong(2010)의 추수 연구(follow-up study)로서, 네 가지 문항반응이론(IRT) 능력 추정량(proficiency estimators)의 측정학적 특성을 보다 심도 있게 탐구하였다. 관심 네 가지 추정량은 최대우도(ML), 검사특성곡선(TCF), 베이지언 사후기대(EAP), 합산점수(summed score)에 기초한 사후기대(sEAP) 추정량이었다. 네 능력 추정량의 조건부 편향(bias), 오차분산, 평균제곱오차(MSE) 및 주변적 신뢰도(marginal reliability)에 있어서의 상대적 기능을 다양한 유형의 선다형 문항 검사 조건에서 검토하였다. 전반적으로, ML 및 TCF 추정량의 기능은 EAP 및 sEAP 추정량의 기능과 뚜렷한 대조를 보였다. ML 및 TCF 추정량의 편향은 능력모수와 정적 상관을 보인 반면, EAP 및 sEAP 추정량의 편향은 능력모수와 부적 상관을 보였다. 모든 검사 조건에서 ML 추정량은 EAP 추정량보다 더 작은 절대편향을, 더 큰 오차분산을, 그리고 더 낮은 신뢰도를 산출하였다. EAP 추정량의 조건부 MSE와 비교하여, ML 추정량은 능력모수의 중간 지점들에서 더 큰 MSE를 산출한 반면 양극단 지점들에서는 더 작은 MSE를 산출하였다. 조건부 및 주변적 측정의 정확성에 있어, 문항반응패턴에 기초한 ML 및 EAP 추정량은 합산점수에 기초하고 있는 TCF 및 sEAP 추정량보다 근소하게 우수한 기능을 보였다. As a follow-up study of Kolen and Tong (2010), this paper investigates more deeply the psychological properties of the four item response theory (IRT) proficiency estimators, the maximum likelihood (ML), test characteristic function (TCF), Bayesian expected a posteriori (EAP), and summed score based EAP (sEAP) estimators. Relative performances in conditional bias, error variance, and mean squared error (MSE) and marginal reliability of the four estimators were examined with various multiple-choice item tests through computer simulations. Overall, the performances of the ML and TCF estimators were distinct from those of the EAP and sEAP estimators. The biases of the ML and TCF estimators each were positively correlated with the true proficiency Ɵ, while the biases of the EAP and sEAP estimators each were negatively correlated with Ɵ. With all the tests considered, the ML estimator had smaller (absolute) bias, larger error variance, and lower reliability than the EAP estimator. The conditional MSE of the ML estimator was larger at most middle points of Ɵ but smaller at few extreme points than that of the EAP estimator. The ML and EAP estimators based on pattern-scoring slightly outperformed, in conditional and marginal measurement precision, their respective summed-scoring counterparts, the TCF and sEAP estimators.

지식자산위험을 고려한 기술가치평가 할인율 적산모형에 관한 연구

성웅현(Sung Oong-Hyun) 한국기술혁신학회 2008 기술혁신학회지 Vol.11 No.2

기술가치평가에서 적절한 할인율의 적용은 가치평가의 신뢰성을 확보하는데 중요한 요소이다. 개별 기술가치평가 할인율은 개별 지식자산에 내재된 위험과 기회를 반영하는 것이기 때문에, 기업가치평가 표준 할인율인 WACC과 상당히 다르다고 판단된다. 본 연구의 목적은 기술가치평가 할인율 적산모형의 위험구조와 위험프리미엄 추정방법을 제안하는 것이다. 적산모형의 성분은 무위험이자율, 전반적 시장위험프리미엄과 베타, 지식자산위험프리미엄 등 세 가지로 구성하였다. 특히, 본 연구에서 할인율 구성의 핵심인 지식자산위험 수준을 평가할 수 있는 10개 항목을 제안하였고, 위험수준 결과를 위험프리미엄으로 변환하기 위한 추정함수인 선형함수, 자연로그함수, 지수함수 등을 적용하였다. 상기 논리와 결과는 기술가치평가 할인율 추정의 객관성을 개선할 수 있는 실무적 대안이 될 수 있을 것이다. Within any income approach, a discount rate is used to convert some projected free cash flow to its presented value. In case of valuing companies, the most frequently used discount rate is the weighted average cost of capital(WACC) at the aggregate level. But technology valuation is different to discounting aggregate corporate cash flow since it is concerned about individual Intellectual property. Therefore, blindly applying standard discount rate such as WACC in technology valuation is unlikely to lead to the right result. The primary focus of this paper is to establish the structure of discount rate for technology valuation and to suggest the method of estimation. To determine an appropriate discount rate for technology valuation, the level of technology risk, market risk and competitive risk should be included in the structure of discount rate. This paper suggests the build-up model which consists of three components as a expansion of the CAPM. It includes (1) a risk-free rate of return, (2) general market risk premium and beta and (3) intellectual property risk premium related to technology risk and specific target market risk. However, there is no specific check list for examining the intellectual property risk until now and no specific method for quantifying its risk into risk premium. This paper developed the 10 element to determine the level of the intellectual property risk and applied estimation function such as linear function, natural log function and exponential function to transform the level of risk into risk premium. The limitation of this paper is that the range of intellectual property risk premium is inferred based on the information of foreign and domestic valuation agency. Finally, this paper explored the development of an intellectual property discount rate for technology valuation and presented the method in order to quantify the intellectual property risk premium.

FDM 방식 3D 프린터에 대한 딥러닝 기반 물성 추정

임주현(Joo Hyeon Im),김원곤(Wongon Kim),안성훈(Sung-Hoon Ahn),윤병동(Byeng Dong Youn) 대한기계학회 2022 대한기계학회 춘추학술대회 Vol.2022 No.11

Additive manufacturing (AM) is playing a major role in Industry 4.0. AM can simplify the fabrication of complex shapes while minimizing manufacturing time and cost. However, the poor surface quality and weak mechanical properties of the output hinder the broad application of AM. In this study, we propose a deep learning-based method for estimating the mechanical property in a fused deposition modeling-type 3D printer. Acceleration signals were acquired in the process of outputting a tensile specimen through construction of a test-bed. In addition, a tensile test was performed using the specimen to collect information on mechanical properties. After that, features were extracted from the divided signals to analyze the correlation with the mechanical properties. Finally, the quality of highly correlated property was estimated through deep neural network. We confirmed that deep learning-based method is good at estimating property through mean square error loss and root mean square error.

함수특성에 근거한 국내 풍화토의 불포화 투수곡선 추정

김윤기,최경림,이성진,이승래,권형석 한국지반공학회 2010 한국지반공학회논문집 Vol.26 No.10

Unsaturated permeability function is an important factor in the design and analysis of various unsaturated soil structures. Generally the permeability characteristics decrease as the matric suction increases and the trend is similar to water retention characteristics of a soil. The permeability of unsaturated soils can be obtained directly by laboratory tests or indirectly by estimation methods from other soil properties. For unsaturated soils sampled from 7 areas in KOREA, SWCCs and unsaturated permeability functions were obtained by experimental tests. The unsaturated permeability results were also compared with the unsaturated permeability functions derived from the SWCCs theoretically. However, the current estimation models of unsaturated permeability function did not express the unsaturated permeability characteristics. Therefore, the FXK-M permeability function was modified to predict more accurate permeability functions for Korean weathered soils using a correction factor that can be calculated from the air-entry value of SWCC. The new estimation model resulted in good agreements for all tested soils.