• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.13 no.4
• /
• pp.141-150
• /
• 1993

For the three parameter Weibull distribution, the parameter estimation techniques are applied and the asymptotic variances of the quantile to obtain the confidence limits for a given return period are derived. Three estimation techniques are used for these purposes: the methods of moments, maximum likelihood and probability weighted moments. The three parameter Weibull distribution as a flood frequency model is applied to actual flood data.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.12 no.4
• /
• pp.790-805
• /
• 1988

Static strength and fatigue life scattering of glass fiber reinforced epoxy composite materials has been studied. Normal, lognormal, two-parameter and three-parameter Weibull distribution functions are used for strength and one-stress fatigue life distribution. The value of mean fatigue life is analysed using mean fatigue life, mean log fatigue life and expected value of 2 and 3-parameter Weibull distribution functions. Modification on non-statistical cumulative damage models is made in order to interpret the result of two-stress level fatigue life scattering. The comparison results show that 3-parameter Weibull distribution has better predictions in static strength and one-stress level fatigue life distributions. However, no advantage of 3-parameter Weibll distribution is found over 2-parameter Weibull distribution in two-stress level fatigue life predictions. It is found that two-stress level fatigue life prediction by the expanded equal rank assumption is close to the experimental data.

• Journal of Korea Water Resources Association
• /
• v.46 no.6
• /
• pp.643-653
• /
• 2013

The most important procedure in frequency analysis is to determine the appropriate probability distribution and to estimate quantiles for a given return period. To perform the frequency analysis, the goodness-of-fit tests should be carried out for judging fitness between obtained data from empirical probability distribution and assumed probability distribution. The previous goodness-of-fit could not consider enough extreme events from the recent climate change. In this study, the critical values of the modified Anderson-Darling test statistics were derived for 3-parameter Weibull distribution and power test was performed to evaluate the performance of the suggested test. Finally, this method was applied to 50 sites in South Korea. The result shows that the power of modified Anderson-Darling test has better than other existing goodness-of-fit tests. Thus, modified Anderson-Darling test will be able to act as a reference of goodness-of-fit test for 3-parameter Weibull model.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.36 no.1
• /
• pp.20-36
• /
• 2024

In this study, probabilistic models for the wave- and lifting forces were derived directly from long-term in-situ wave data embedding the Korean marine environment characteristics varying from sea to sea based on the Three-Parameter Weibull distribution. Korean marine environment characteristics varying from sea to sea carved out their presence on the probability coefficients of probabilistic models for wave- and lifting forces. Energetic wave conditions along the southern coast of Korea distinguish themselves from the others with a relatively large scale coefficient, small location coefficient, and shape coefficient around 1.3. On the other hand, mild marine environment along the western coast has a small variability, leading to small scale-coefficient, large location coefficient and shape coefficient around 2.0. In the sea off Mokpo, near the boundary between the South- and West Seas, marine environment was characterized by small scale-coefficient, large location coefficient, and shape coefficient around 1.2, implying that marine environments characteristics of the South-and West Sea coexist in the sea off Mokpo.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2006.05a
• /
• pp.697-700
• /
• 2006

저수시 하천유량(Low Streamflow)의 추정은 하천의 수질관리, 용수공급계획, 댐 방류계획등의 수자원관리에 있어서 매우 중요한 부분이다. 이러한 중요성에 따라 Vogel과 Kroll (1989)은 저수시 하천유량을 추정하기 위한 여러 가지 확률분포함수를 제안하였다. 가장 흔히 제안되어지는 이변수 확률분포(Two-Parameter Distribution)로는 Lognormal 분포와 Weibull 분포가 있으며 이와 더불어 Three-Parameter Lognormal, Three-Parameter Weibull, Log Person Type Ⅲ 분포도 널리 사용되어진다. 그러나 이러한 여러 가지 확률 분포함수 중에서 가장 적절한 확률분포의 선택은 저수시 하천유량의 물리적인 측면과는 상관없이 주로 적합도(Gooness of Fit)에 기인된 통계치에 의해서만 결정되기도 하는데 이러한 경우 잘못된 가정을 받아들이는 확률이 높아짐에 따라 추정결과의 신뢰성(Reliability)을 감소시킬 수 있다. 이러한 문제점을 극복하기 위해서 Onoz와 Bayazit (2001)는 Recession Curve를 지수함수로 가정하고 최대 갈수 기간의 길이(Maximum Dry Period Length)의 확률에 대한 이론적인 결과치들을 사용하여 Weibull 분포의 특정한 경우에 해당되어지는 Power 분포를 유도하였으며 유도된 Power 분포의 매개변수를 추정하기 위하여 L-Moment 방법을 사용하였다. 또한 Onoz와 Bayazit (2001) 작은 유출량에서 확률분포와 잘 맞지 않는 경우 작은 유출량값에 작은 가중치를 부여하여 확률분포에 대한 영향을 줄이는 방법인 LL-Moment 방법을 제안하였다. 본 연구에서는 낙동강 유역의 1번부터 5번 소유역에 대해 SSARR 모형을 이용하여 모의한 유출량을 이용하여 Weibull 분포, L-Moment방법에 의해 추정된 매개변수를 사용한 Power 분포, LL-Moment 방법에 의해 추정된 매개변수를 사용한 Power 분포를 적용하였으며 이들 분포의 적합도를 PPCC Test를 사용하여 평가해봄으로써 낙동강 유역에서의 저수시의 유출량 추정에 대한 Power 분포의 적용성을 판단해 보았다.

• Water for future
• /
• v.29 no.1
• /
• pp.129-139
• /
• 1996

It is very important to select appropriate distributions for hydrological data in planning and designing hydraulic structures. Also, it is necessary to check whether the selected distribution reproduces the statistical characteristics of the real data. In this study, the parameters of the two- and three-parameter gamma, two- and three-parameter lognormal, Gumbel, two- and three-parameter log-Gumbel, GEV, log-Pearsonn type III, two- and three-parameter Weibull, four- and five-parameter Wakeby distributions were estimated for the rainfall data of 22 sites in Korea with 7 different durations based on the methods of moments, probability weighted moments, and maximum likelihood. And the validity conditions were checked for the estimated parameters. The separation effect for each distribution was examined throught 10,000 simulations using the estimated parameters. As results, the separation effect was the smallest: log-Pearson type III for moment method, log-Pearson type III and GEV for probability weighted moment method, and GEV for maximum likelihood method. However, it is large for the two-parameter distributions.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.35 no.10
• /
• pp.1281-1289
• /
• 2011

This paper deals with several statistical distribution functions for the analysis of fatigue life data of composite laminates for small wind-turbine blades. A series of tensile tests was performed on triaxial glass/epoxy laminates for loading directions of $0^{\circ}$, $45^{\circ}$, and $90^{\circ}$. Then, fatigue tests were carried out to determine the fatigue life at the aforementioned loading directions and the fatigue stresses at four levels. Two-parameter Weibull, three-parameter Weibull, normal, and log-normal distributions were used to fit the fatigue life data of the triaxial composite laminates. The three-parameter Weibull distribution most accurately described the fatigue life data measured experimentally for all the cases considered. Furthermore, the variation of fatigue life was simultaneously affected by the loading direction and fatigue stress level.

• Journal of Korean Society of Forest Science
• /
• v.90 no.2
• /
• pp.176-183
• /
• 2001

Loblolly pine (Pinus taeda L.) is the most economically important timber producing species in the southern United States. Much attention has been given to predicting diameter distributions for the solution of multiple-product yield estimates. The three-parameter Weibull diameter distribution yield prediction systems were developed for loblolly pine plantations. A parameter recovery procedure for the Weibull distribution function based on four percentile equations was applied to develop diameter distribution yield prediction models. Four percentiles (0th, 25th, 50th, 95th) of the cumulative diameter distribution were predicted as a function of quadratic mean diameter. Individual tree height prediction equations were developed for the calculation of yields by diameter class. By using individual tree content prediction equations, expected yield by diameter class can be computed. To reduce rounding-off errors, the Weibull cumulative upper bound limit difference procedure applied in this study shows slightly better results compared with upper and lower bound procedure applied in the past studies. To evaluate this system, the predicted diameter distributions were tested against the observed diameter distributions using the Kolmogorov-Smirnov two sample test at the ${\alpha}$=0.05 level to check if any significant differences existed. Statistically, no significant differences were detected based on the data from 516 evaluation data sets. This diameter distribution yield prediction system will be useful in loblolly pine stand structure modeling, in updating forest inventories, and in evaluating investment opportunities.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.35 no.4
• /
• pp.375-381
• /
• 2011

The primary aim of this study is to investigate the probabilistic characteristics of the fatigue parameters that describe the fatigue crack growth behavior in magnesium alloy. Statistical fatigue crack propagation experiments have been performed on rolled AZ31 magnesium alloy CT specimens with different specimen thickness, load ratio, and maximum load at ambient temperature in a laboratory. Using the statistical fatigue data obtained from these experiments, the goodness-of-fit of the probability distribution of the fatigue behavior parameters is evaluated in this study by performing statistical analyses. The crack growth rate coefficient is a fatigue parameter having a very large COV(Coefficient of Variation), but the variation of a crack growth rate exponent is not substantial. It is considered that a crack growth rate exponent can be a material constant. It is also found that the best fit probability distribution of the parameters such as the crack growth rate coefficient and crack growth rate exponent for a magnesium alloy is a three-parameter Weibull distribution, and two-parameter Weibull distribution is a good distribution only for the crack growth rate coefficient.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.30 no.11 s.254
• /
• pp.1417-1424
• /
• 2006

In order to evaluate variation of fatigue life of mechanical components including engineering plastics, it is important to estimate probabilistic strain-life curves to accurately define the variation of fatigue characteristics. This paper intends to provide new assessment of P-$\varepsilon$-N (probabilistic strain-life curves) for considering the variation of fatigue characteristics in polyacetal. The fatigue strain controlled tests were conducted under constant 50% humidity and room temperature condition by a universal testing machine at strain ratio, R=0. A practical procedure is introduced to evaluate probabilistic strain-life curves. Three probabilistic distributions were used for generating P-$\varepsilon$-N curves such as normal, 2-parameter and 3-parameter Weibull. In this study, 3-parameter Weibull distribution was found to be most appropriate among assumed distributions when the probability distributions of the fatigue characteristic were examined using chi-square and Kolmogorov-Smirnov test. The more appropriate P-$\varepsilon$-N curves for these materials are generated by the proposed method considering 3-parameter Weibull distribution.