品質工学
Online ISSN : 2189-9320
Print ISSN : 2189-633X
ISSN-L : 2189-633X
5 巻, 3 号
選択された号の論文の5件中1~5を表示しています
論説
  • 田口 玄一
    1997 年 5 巻 3 号 p. 26-31
    発行日: 1997/06/01
    公開日: 2016/12/24
    ジャーナル フリー

    The difference between the traditional robust design approach and the new one are discussed in this article. The traditional engineering way for functionality is as follows. Design a product or a process to perform its intended function at normal condition at first, then improve its functionality at various noise conditions by minimizing the mean square deviation from the intended function. The new way improves the functionality first, considering all variables happened at customer's conditions are either signal or noise where the signal should have an intended function but noises should not, then adjust the function to meet the intended function at only norrnal initial condition by least square methods.

開発と研究
  • 青木 昭夫, 大村 一世
    1997 年 5 巻 3 号 p. 41-46
    発行日: 1997/06/01
    公開日: 2016/03/24
    ジャーナル フリー
    Recently, the components with a driving control mechanism are equipped in many cars. The sensor attached to the control mechanism is called accelerometer which detects acceleration or deceleration when a car is accelerated, decelerated or steered. Since this type of detecting sensor uses electronic circuits and filling materials,its output characteristics vary a great deal due to the environmental conditions such as temperature changes. As a result, the per formance of control function is affected. In this paper, the technology development of accelerometers is introduced targeting for a stable output characteristics: robust against temperature and other environmental changes. ln the study, efforts was made to improve origin quality (generic function) rather than robust quality (objective function).
  • 橋本 晃
    1997 年 5 巻 3 号 p. 47-54
    発行日: 1997/06/01
    公開日: 2016/03/24
    ジャーナル フリー
    Bonding is an indispensable technique in automobile manufacturing which can be used to bond a great variety of component parts. The technique is important to develop new products for a better performance and a higher value. Today, brazing is not used often because of its insufficient strength and variability after brazing, although the process is very suitable for mass production. The improvement of a brazing process for a higher strength would make the process much more applicable in manufacturing. This paper introduces a 2-piece gear brazing experiment to find a robust condition for stable and high strength joints using Taguchi Methods.
  • L12直交表による網掛け実験の検討
    高橋 和仁, 阿部 憲瑞, 辻 千尋, 矢野 宏
    1997 年 5 巻 3 号 p. 55-62
    発行日: 1997/06/01
    公開日: 2016/03/24
    ジャーナル フリー
    In the study of cutting process for the improvement of precision after cutting, evaluation has been made using dimensions such as the cutting depth of tools or computer input dimensions as the signal. From such type of approach, quality may be improved to a certain extent. ln those case studies however, the effects of some energy-related control factors such as processing speed or feeding speed often show either a mountain-shape or valley-shape, indicating the existence of interactions. To avoid such problems, a suggestion has been made to use the electrical power consumption for cutting as the input signal. ln this paper, a systematic experiment was conducted mainly focusing the generic function of machine processing so that the additivity of SN ratio may be expected. Also some points related to the sue of electrical power were discussed.
事例研究
  • 日座 和典
    1997 年 5 巻 3 号 p. 63-70
    発行日: 1997/06/01
    公開日: 2016/03/24
    ジャーナル フリー
    A study with the same objective contained in this paper was reported in 1994 (Journal of Quality Engineering Forum, Vol. 2, No. 4). ln the study, parameter design was conducted using sensory-type graded data as a means of evaluation. ln this paper, Mahalanobis Distance was used to analyze the same data collected in the previous study. From this study, it was found that the trend of some factors affecting legibility were quite similar in both methods. For example, the trend of two distributions showing legibility and illegibility based on the significant factors using Mahalanobis Distance is fairly similar to the ones obtained from graded data.
feedback
Top