Screw tightening is the most common simple fastening method, with nulnerous applications in everyday life. Despite its widespread general use, however,there are many problems due to under-tightening, over-tightening or loosening of screws; examples include falling, breakage, and reduced strength of fastened objects. In manufacturing, screw tightening is an important elemental technology:countless screws are used to fasten parts. Control of this technology depends on torque control and worker training. We defined the relationship between additional tightening torque and rotation angle of a screw as a generic functon,and attempted optimization using the standard S/N ratio. As a result, we succeeded in establishing an elemental technology of screw tightening.

We have been testing two methods of efficiently evaluating multilayer thin films coated on cemented carbide molding toels:the load-friction force characteristic and load-indentation area characteristic. In a study reported prevlously, we demonstrated reproducible gain in the S/N ratio ofthe load-indentation area characteristic but discovered several problems. In the present study, we took measures to solve those problems, performed a re-evaluation, and obtained new results showing reproducible gain in the S/N ratio of the load-friction force characteristic,

One of the major properties of paper is its stiffness, which is critical to the paper feeding performance and image quality of copiers and printers. Paper stiffness is accordingly an important inspection item that rarely fails to appear in contracts with paper manufacturers. The basic method of measuring paper stiffness stipulated in the Japan Industrial Standards (JIS) is used both at the authors' company, which deals in copiers and printers, and by paper manufacturers. However, although paper stiffness is an inlportant property, at present there is no recognized national or international stiffness standard to which measured values can be traced, and JIS does not specify the measurement system in detail. The authors therefore undertook to find stable measurement conditions with minimal measurement error, which could be used to establish known true values. In this report, parameter design is used to calculate the standardized S/N ratio for a commercially available Clark stiffness tester employed at the authors' company. This led to successful optimization of the measurement conditions.

The making of musical instruments such as guitars and violins is known to involve the work of highly experienced and highly skilled craftsmen, but it is also thought to involve much loss due to trial and error. To reduce this loss, in this study we used the MT system to create a measurement scale for guitar tone, and performed parameter design to obtain a provisional ideal tone. The measurement scale is based on the one-sided T test, one of the methods employed in the MT system. Tene difference due to string diameter was used as a signal factor. The fifth string was tuned to A to provide a provisional ideal tone; then the third string was tuned to A, using a graphic equalizer to approximate the same tone. In a first L18 orthogonal array experiment using the standardized S/N ratio, the distance in the MT system approached the normal value, but the sensitivity was not reproducible. Reproducible sensitivity was then obtained through optimization by making a tone transferability evaluation of the factorial effect of the A note of the fifth string, using the distance obtained by the one-sided T test as the characteristic value of the factorial effect. We believe that application of the MT system to sound can help craftsmen in the pursuit of ideal guitars.

In our previous research, we applied the two-sided Taguchi method to condominiums in an urban area of Tokyo and showed that the effects of green landscape differ depending on the area (riverside area, uptown area, or bay area) in which the condominium is located. A problem with this analysis is that in the development of new real estate, the results cannot be applied to actual develepment cenditions without knowing the type of area in which the real estate will be classified. In this study, we tried to classify each area by using item diagnosis so that the previous research results could be used for practical purposes. With the conventional method, diagnosis took time, as it was hard to judge eiearly whether patterns of areas were alike. By calculating correlation coefficients between items and drawing a scatter chart, we were able to classify the item pattems for each area and obtain appropriate criteria that could be applied to real estate developments of unknown type. In addition, it was shown that this method can provide criteria for classification of anomalies. Although this method is still in the trial stage, it has made item diagnosis more efficient.

In recent years, optimization methods in which quality engineering is applied to simulation and no prototypes are designed have had great success in a broad range of fields. Advantages over the use of prototypes include lower-cost evaluation,quicker results, and the ability to work with characteristics that would be hard to measure with prototypes. In this study, a clamping mechanism that clamps a mold of an injection molding machine was eptimized by using general-purpose spreadsheet software to perform theoretical calculations in a parameter design evaluated by the standardized S/N ratio, and using β1 and β2 to tune the characteristic values to target values.