• 한국정보처리학회논문지
• /
• 제7권3호
• /
• pp.852-861
• /
• 2000

Software reliability growth models are used in testing stages of software development to model the error content and time intervals betwewn software failures. In this paper, using priors for the number of fault with the negative binomial distribution nd the error rate with gamma distribution, Bayesian inference and model selection method for Jelinski-Moranda and Goel-Okumoto and Schick-Wolverton models in software reliability. For model selection, we explored the sum of the relative error, Braun statistic and median variation. In Bayesian computation process, we could avoid the multiple integration by the use of Gibbs sampling, which is a kind of Markov Chain Monte Carolo method to compute the posterior distribution. Using simulated data, Bayesian inference and model selection is studied.

• Structural Engineering and Mechanics
• /
• 제76권4호
• /
• pp.505-512
• /
• 2020

This paper presents an error estimation technique for 2-D crack analysis by an enriched natural element (more exactly, enriched Petrov-Galerkin NEM). A bare solution was approximated by PG-NEM using Laplace interpolation functions. Meanwhile, an accurate quasi-exact solution was obtained by a combined use of enriched PG-NEM and the global patch recovery. The Laplace interpolation functions are enriched with the near-tip singular fields, and the approximate solution obtained by enriched PG-NEM was enhanced by the global patch recovery. The quantitative error amount is measured in terms of the energy norm, and the accuracy (i.e., the effective index) of the proposed method was evaluated using the errors which obtained by FEM using a very fine mesh. The error distribution was investigated by calculating the local element-wise errors, from which it has been found that the relative high errors occurs in the vicinity of crack tip. The differences between the enriched and non-enriched PG-NEMs have been investigated from the effective index, the error distribution, and the convergence rate. From the comparison, it has been justified that the enriched PG-NEM provides much more accurate error information than the non-enriched PG-NEM.

• 대한공간정보학회지
• /
• 제13권2호
• /
• pp.79-84
• /
• 2005

현재 위성영상 센서모형화의 정확도에 관한 연구는 자료의 전체적인 오차크기를 추정하는데 초점이 맞추어져 있다. 이러한 유형의 오차평가는 오차의 정량적인 크기만을 고려한 것으로 오차의 분포 특성을 해석할 수 없다는 한계점을 안고 있다. 위성영상 센서모형화 결과의 수치적 평가는 오차의 분포 특성의 부재로 위성영상 모형화 결과의 신뢰성이 떨어지게 된다. 본 연구를 통해 오차의 크기를 잘 표현할 수 있는 RMSE와 더불어 오차의 방향성 계수를 산정함으로써 오차의 분포 특성을 확인할 수 있었다. 또한 방향성 계수가 1에 가까울수록 특정 계통오차가 있으며 0에 가까울수록 우연오차가 있음을 확인 할 수 있어 오차의 방향성 분석을 통하여 오차를 감소시킬 수 있는 방안을 제시할 수 있었다.

• Journal of Electrical Engineering and Technology
• /
• 제13권4호
• /
• pp.1504-1514
• /
• 2018

The objective of this study is to compare the suitability of a non-parametric and 3 parametric distributions in the characterization of prediction intervals of photovoltaic power forecasts with high confidence levels. The prediction intervals of the forecasts are calculated using a method based on recent past data similar to the target forecast input data, and on a distribution assumption for the forecast error. To compare the suitability of the distributions, prediction intervals were calculated using the proposed method and each of the 4 distributions. The calculations were done for one year of day-ahead forecasts of hourly power generation of 432 PV systems. The systems have different sizes and specifications, and are installed in different locations in Japan. The results show that, in general, the non-parametric distribution assumption for the forecast error yielded the best prediction intervals. For example, with a confidence level of 85% the use of the non-parametric distribution assumption yielded a median annual forecast error coverage of 86.9%. This result was close to the one obtained with the Laplacian distribution assumption (87.8% of coverage for the same confidence level). Contrasting with that, using a Gaussian and Hyperbolic distributions yielded median annual forecast error coverage of 89.5% and 90.5%.

• 조명전기설비학회논문지
• /
• 제21권9호
• /
• pp.81-86
• /
• 2007

Electric utility companies have the responsibility of providing good electricity for their customers. They have introduced the DAS(Distribution Automation System) to automate the power distribution networks. DAS engineers require state-of-the-art applications, such as a way to actively manage the distribution system and gain economic benefits from a flexible DAS architectural design. The existing DAS is not capable of handling these needs. It requires operator intervention whenever feeder overloading is detected while operator error could cause the feeder overload area to be extended. It also utilizes a closed architecture and it is therefore difficult to meet the system migration and future enhancement requirements. This paper represents a web based, platform-independent, flexible DAS architectural design and active database application. Recent advanced Internet technologies are fully utilized in this new DAS architecture allowing it to meet the system migration and future enhancement requirements. By using an active database, the DAS can minimize the feeder overloading area in the distribution system without operator intervention, thereby minimizing mistakes due to operator error.

• 산업공학
• /
• 제22권3호
• /
• pp.214-222
• /
• 2009

In the early design stage, system reliability must be estimated from life testing data at the component level. Previously, a point estimate of system reliability was obtained from the unbiased estimate of the component reliability after assuming that the number of failed components for a given time followed a binomial distribution. For deriving the confidence interval of system reliability, either the lognormal distribution or the normal approximation of the binomial distribution was assumed for the estimator of system reliability. In this paper, a new estimator is used for the component level reliability, which is biased but has a smaller mean square error than the previous one. We propose to use the beta distribution rather than the lognormal or approximated normal distribution for developing the confidence interval of the system reliability. A numerical example based on Monte Carlo simulation illustrates advantages of the proposed approach over the previous approach.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2001년도 춘계학술대회논문집A
• /
• pp.262-267
• /
• 2001

In this study, the Reliability of degraded steam turbine blade was evaluated using the limited fatigue data. The statistical estimation of limited fatigue data implies that some unknown uncertainties which may be involved in fatigue reliability analysis. Therefore, an appropriate distribution in the fatigue strength was determined by the characteristic distribution - linear correlation coefficient, fatigue physics, error parameter. 3-parameter Weibull distribution is the most appropriate distribution to assume for infinite region. The load applied on the blade is mainly tensile. The maximum Von-Mises stress is 219.4 MPa at the steady state service condition. The failure probability($F_p$) derived from the strength-stress interference model using Monte carlo simulation under variable service condition is 0.25% at the 99.99% confidence level.

• Journal of the Korean Data and Information Science Society
• /
• 제27권5호
• /
• pp.1399-1412
• /
• 2016

We compare many normality tests consisting of different sources of information extracted from the given data: Anderson-Darling test, Kolmogorov-Smirnov test, Cramervon Mises test, Shapiro-Wilk test, Shaprio-Francia test, Lilliefors, Jarque-Bera test, D'Agostino' D, Doornik-Hansen test, Energy test and Martinzez-Iglewicz test. For the purpose of comparison, those tests are applied to the various types of data generated from skewed distribution, unsymmetric distribution, and distribution with different length of support. We then summarize comparison results in terms of two things: type I error control and power. The selection of the best test depends on the shape of the distribution of the data, implying that there is no test which is the most powerful for all distributions.

• 터널과지하공간
• /
• 제26권3호
• /
• pp.201-211
• /
• 2016

본 연구에서는 3차원 원통형 조사창에서의 양끝내포선 길이 분포를 이용하여 절리 직경 분포 추정을 수행하였다. 추정 결과의 수치적인 오차를 줄이기 위하여, 보조변수를 도입한 개선된 방안을 제시하였다. 몬테-카를로 시뮬레이션으로 검증한 결과, 보조변수의 변화에 따라 추정 분포의 진동이 줄어들어 오차가 크게 감소한 것을 확인하였다. 또한 절리 직경의 최적 분포 및 적정한 보조변수 값을 찾기 위한 알고리즘을 제안하였다.

• Journal of the Korean Data Analysis Society
• /
• 제20권6호
• /
• pp.2747-2757
• /
• 2018

We consider the MLE (maximum likelihood estimate) and Bayesian estimates of three-parameter bathtub-shaped lifetime distribution based on the progressive type II censoring with binomial removal. Jung, Chung (2018) proposed the three-parameter bathtub-shaped distribution which is the extension of the two-parameter bathtub-shaped distribution given by Zhang (2004). Jung, Chung (2018) investigated its properties and estimations. The maximum likelihood estimates are computed using Newton-Raphson algorithm. Also, Bayesian estimates are obtained under the balanced loss function using MCMC (Markov chain Monte Carlo) method. In particular, BSEL (balanced squared error loss) function is considered as a special form of balanced loss function given by Zellner (1994). For comparing theirs MLEs with the corresponding Bayes estimates, some simulations are performed. It shows that Bayes estimates is better than MLEs in terms of risks. Finally, concluding remarks are mentioned.