• The Korean Journal of Applied Statistics
• /
• v.19 no.2
• /
• pp.339-347
• /
• 2006

In mixture experiments, performing experiments in highly constrained regions causes collinearity problems. We can use the ridge regression as a means for stabilizing the coefficient estimators in the fitted model. But there is no theory available on which to base statistical inference of ridge estimators. The bootstrap technique could be used to seek the confidence intervals for ridge estimators.

• The Korean Journal of Applied Statistics
• /
• v.10 no.1
• /
• pp.15-27
• /
• 1997

Randomized block design (RBD) with three replication is very frequently adopted in agricultural experiments of the Rural Development Administration of Korea. Even though it works well in field trials of traditional crops, it may not accomodate trial site conditions and/or experimental environment. In this research report, we deal with two such cases. The first case is for a crop experiment in green houses. In house conditions, RBD may not be appropriate since it cannot reflect two directions of the yield gradient. So, a Latin square design is suggested as an alternative. The second case is for local field experiments of the newly-inbred rice. RBD with three replications is used without doubt for decades, even though the site layout is not appropriately shaped for the design. In this case, we suggest the RBD in two blocks with multiple replicates for control varieties as an alternative. To improve the quality of statistical experimental designs in over one-thousand agricultural trials performed annually in the Rural Development Administration, we need to re-train agricultural researchers on the design and analysis of experiments and call for concerns of Korean statisticians.

• Journal of Ocean Engineering and Technology
• /
• v.21 no.3 s.76
• /
• pp.26-32
• /
• 2007

A series of large-scale experiments were carried out in order to examine wave-induced liquefaction in a loosely packed sandbed, its afterward high densification and liquefaction by oscillatory pore pressure. The experiments were conducted in a Large Hydro-Geo Flume that can nearly solve the problems of scale effects of the sandbed, and the 50% sieve diameter of sand was 0.2 mm. The generation of residual pore pressure and its afterward high densification which had observed by Takahashi et al. (1999) in a wave flume experiment using fine sand with the size of 0.08 mm. As a result, the relative density of the sandbed after high densification was increased up to 79% and liquefaction by oscillatory pore pressure was not observed.

• Journal of Korean Society for Quality Management
• /
• v.24 no.4
• /
• pp.156-167
• /
• 1996

D-optimality is used often in design augmentation of mixture experiments. Although such alphabetic criteria provide a valuable foundation for generating designs, they often fail to convey the true nature of the design's support of the fitted model in terms of prediction variance over a region of interest. Thus, a graphical method is proposed to evaluate augmented designs in mixture experiments. This method can be used to evaluate the effect of missing observation and outlier in mixture experiments.

• Transactions of the Korean Society of Machine Tool Engineers
• /
• v.13 no.4
• /
• pp.1-8
• /
• 2004

Various controllable parameters of an experiment have influence on grinding process. In order to get good products with a high quality, these parameters should be considered whether each parameter has relations to the quality. This paper describes the use of the design of experiments to minimize geometric error in surface grinding. Controllable parameters for the design of experiments were selected as spindle speed, table speed, depth of cut and grain size. From the experimental results, a degree of influence between these parameters and the geometric error was evaluated. An optimal set of grinding conditions was obtained by means of analysis of variance(ANOVA).

• Journal of the korean Society of Automotive Engineers
• /
• v.14 no.2
• /
• pp.88-98
• /
• 1992

Real automobiles have been utilized to study the occupant behaviors and the response of the structures in the crash environment. Since various parameters are involved in the automobile crashworthiness, a number of experiments must be conducted. The experiments have been contributed to increasing the cost due to the fact that the test is quite expensive. Therefore, computer simulation is adopted to reduce the number of experiments. A few computer programs have been developed specifically to solve the occupant responses in the crash environment. In this research, a software is used to study the occupant dynamic analysis. A modeling of occupant analysis is established for a passenger car and the results are verified through comparisons with real experiments. In the modeling, data are tuned very carefully so that simulated results such as HIC(Head Injury Criterion) and acceleration of each body may approximate to the experimental results. The compared experiment is a barrier test which is carried out by frontal impact. A feedback to the design process is suggested from the result of this research.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.26 no.7
• /
• pp.1348-1356
• /
• 2002

The present work concerns the optimal design for injection molding processes by using the design of experiments (DOE) and numerical analysis. The DOE approaches is planned to be able to consider two-way interaction, and have been applied progressively for both mold design and process design. Numerical analyses have been carried out as a design of experiments for mold parameters such as runner specifications and cooling channel configurations. In order to determine optimal process parameters, experiments have been performed for various process conditions with the DOE scheduling. As a result, the quality and productivity of the product have been improved, and the proposed approach can be successfully reflected on the industrial injection molding process of precision electronics parts.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.28 no.7
• /
• pp.885-895
• /
• 2004

Generally, automatic design is carried out with computer simulation and the simulation models are established by investigating the correlations between the simulation and real experiments. Therefore, the experiment results are utilized as complimentary data although they are considered to be precise. Orthogonal arrays have been adopted for discrete design. A method is proposed to directly exploit the experiment results in the design process with orthogonal arrays. Experiments are allocated to some rows of an orthogonal array and computer simulations are allocated to the others. A rule for the allocation is found to keep the orthogonality. Error analysis of the design results is performed. Mathematical examples are made to verify the validity of the proposed method. Error models are defined with the examples and the design solutions from the examples are discussed.

• Journal of the Korean Society for Precision Engineering
• /
• v.21 no.10
• /
• pp.50-56
• /
• 2004

Finishing is the final process in molds manufacturing and consumes much process time. Also, it influences on surface quality of molds. But, there are few systematic methods to control the process. In this work, basic experiments were carried out to study the machining characteristics of the finishing tools. From the experiments, critical surface roughness and wear coefficient are suggested to reduce the number of finishing steps and to plan a systematic finishing procedure. Comparison experiments were carried out between the expert's method and the new method, which is based on the results of this research. From the experiments, it is verified that the systematic method takes less time and generates less form error in the machined surface than the worker's method.

• Journal of Korean Society for Quality Management
• /
• v.34 no.2
• /
• pp.122-128
• /
• 2006

One feature of a computer simulation experiment, different from a physical experiment, is that the output is often deterministic. Moreover the codes are computationally very expensive to run. This paper deals with the design and analysis of computer experiments(DACE) which is a relatively new statistical research area. We model the response of computer experiments as the realization of a stochastic process. This approach is basically the same as using a spatial linear model. Applications to the optimal mechanical designing and model calibration problems are illustrated. Algorithms for selecting the best spatial linear model are also proposed.