• 응용통계연구
• /
• 제17권3호
• /
• pp.499-505
• /
• 2004

2수준 직교배열표를 이용한 실험자료에 대한 파레토 그림에 의한 분석 방법은 실무에서 많이 활용되고 있는데, 그 이유는 유의한 요인을 선별하기 위해서 분산분석법을 사용하지 않고 시각적이고 간결한 방법에 의하여 실험자료를 분석하기 때문이다. $L_{18}(2 \times 3^7)$ 직교배열표를 이용한 실험자료에 대한 분석방법으로 Park(1996)은 효과의 크기를 각 효과의 평균제곱으로 정의하고 파레토 원칙을 사용한 Pareto ANOVA를 제안하였다. 이 논문에서는 $L_{18}(2 \times 3^7)$ 실험자료에 대한 새로운 파레토 그림에 의한 분석 방법이 제시된다. 주요 요점은 3수준 효과의 크기를 일차와 이차 직교대비의 크기에 의해 분할하는 것이다.

• 한국품질경영학회:학술대회논문집
• /
• 한국품질경영학회 2006년도 춘계학술대회
• /
• pp.80-85
• /
• 2006

HSS (high strength steel) is widely applied to reduce the weight but improve the strength in automobiles. This research has been peformed to secure a methodology to accurately predict the springback of HSS for successful tool and process designs in sheet stamping operations. We first peformed U-draw bending test to evaluate the springback characteristic. We then evaluated forming and springback processes using the 1-row model of the finite element method. Based on the peformance measure and parameters selected, extensive analyses of the factor effects on the springback have been made using experimental design concepts. We specifically selected Taguchi's orthogonal array, $L_{18}(2^1{\times}3^7)$, and the optimal level combination of the factors have been drawn from the analysis.

• 한국자동차공학회논문집
• /
• 제13권2호
• /
• pp.72-82
• /
• 2005

The objective of this paper is to describe the optimization of design parameters in a large-sized commercial bus heater system by using CFD(computational fluid dynamics) analysis and Taguchi method. In order to obtain the best combination of each control factor which results in a desired performance of heater system, the parameter design of the Taguchi method is adopted for the robust design considering the dynamic characteristic. The research activity may be divided into four phases. The first one is analyzing the problem, i.e., ascertaining the influential factors. In the second phase the levels were set in such a way that their variation would significantly influence the response. In the third phase the experimental runs were designed. In the final phase the planned runs were carried out numerically to evaluate the optimal combination of factors which is able to provide the best response. In this study, eight factors were considered for the analysis: one with two level and seven with three level combinations comprising the $L_{18}(2^1{\times}3^7)$ orthogonal array. The results of this study can be summarized as follows ; (i)The optimum condition of control factor is a set of <$A_2\;B_1\;C_3\;D_3\;E_1\;F_2\;G_3\;H_2$> where A is shape of the outer fin, B is pitch of the outer fin, C is height of the outer fin, D is the inner fin number, E is the inner fin height, F is length of the flame guide, G is diameter of the heating element and H is clearance between air guide and heating element. (ii)The heat capacity of heated discharge air under the optimum condition satisfies the equation y=0.6M w here M is a signal factor. (iii)The warm-up rate improves about three times, more largely as com pared with the current condition, which results in about 9.2minutes reduction.

• 한국자동차공학회논문집
• /
• 제10권3호
• /
• pp.111-122
• /
• 2002

European offset impact analysis of an electric vehicle was done using the robust design method. Section properties of some of the vehicle structure components which influence the european offset impact characteristics were chosen as the main factors fur the design. Eight factors were considered for the analysis: one with two level and seven with three level combinations comprising the L$\_$18/(2$^1$$\times$3$\^$7/) orthogonal array. It is shown that the sensitivities of the factors and the best combination of the factors can be obtained. One of the noteworthy results is that the design with the lower stiffness of the front components of engine room than the rear components can absorb more crash energy. Also it is shown that there exist interactions between some of factors considered.

• 설비공학논문집
• /
• 제16권11호
• /
• pp.1077-1083
• /
• 2004

This study is to systematically analyze the effect of various kinds of design factors on the performance of stratified thermal storage tank. Taguchi method, known as a very reasonable tool in the parametric study, is employed in the present work. Three dimensional unsteady numerical experiment is conducted for 4 design parameters of stratified thermal storage tank: inlet Reynolds number, Froude number, diffuser size d with 3 levels (Re=400, 800, 1200, Fr=0.5, 1.0, 2.0 and d=150 mm, 200mm, 300 mm) and diffuser shape with 2 levels. Orthogonal array $L_{18}(2{\times}3^7)$ is adopted for the analysis of variance. The result gives quantitative estimation of the various design parameters affecting the performance and helps to select the main factors for the optimum design of stratified thermal storage tank. Reynolds number is found to be the most dominant parameter and the diffuser shape plays significant role on the performance of stratified thermal storage tank. Based on this finding, the prior questions on the contribution of the diffuser shape proposed by the authors become clear. The optimum condition for the performance is a set of d=300mm, Re=800, and radial regulated plate diffuser. Conformation test shows the repeatability in the analysis and $1.3\%$ difference between the estimated thermocline thickness and that of numerical result.