Optimization of injection molding was studied both with actual machinery and by simulation, focusing on dimensional transferability, in an attempt to verify the principle that optimal molding occurs when every part shrinks equally relative to the mold. Data obtained in an optimization study conducted by simulation,paired with experimental data obtained in a past transferability study, were reanalyzed on a localized basis. As a result, it was confirmed that the effects of control factors differ significantly from one place to another in the mold. Control factors taking these local variations into consideration were selected, and the gain resulting from the optimization by simulation showed some degree of reproducibility, but large local interaction effects with the control factors were found, demonstrating the limits of optimization of injection molding.

With the rise of the BRICs in recent years, it has become impossible to maintain competitive prices just by followillg conventional manufacturing methods. If Japan is to compete, we must break out of the conventional pattern of step-by-step processes such as lathing, machining, electrodischarge, and polishing,and employ innovative processes that eliminate intermediate steps. For mold manufacture, this type of innovative process is possible with a milling-combined laser metal sintering machine, which alternately builds layers by laser sintering of a metal powder and machines the layers, eliminating the conventional electrodischarge step. This paper reports on a significant reduction in development time achieved by the application of quality engineering to optimize the laser metal sintering part of the process, discusses the effect of measurement data, especially missing data, on the calculation results, stresses the importance of measurement methods, and reports the results of activities in which personnel in charge of technical development took the initiative in promoting quality engineering within the firm.

In overseas factories where labor costs are lower than in Japan, assembly work and other such production line operations rely heavily on manpower, so the quality and quantity of the products produced depend on the capabilities of the workers. To produce conforming articles speedily, workers have to be able to do repetitive work evenly on the production line. However, work abilities differ greatly from individual to individual. To hire the workers best suited for the tasks to be done and place them in the right jobs, a way of accurately testing their aptitudes, a paper-and-pencil test, for example, is necessary. In this study,to establish an effective autonomous way to find out if a person is suited for production work in advance in production sites throughout the world, we studied the use of the RT method to judge workers' working capabilities from the results of the Kraepelin general-purpose aptitude test. As a result, it was confirmed that worker capability can be determined at a level that is effective enough for practical use.

In the design of an automobile seat frame, there are strength requirements that must be met by law. Among those requirements, seat anchorage strength,which is tested by applying a load to the seat belt, is an important consideration in the basic structural design. Design methods that consider energy balance in order to reduce variation in buckling position can be used to design a structure that can meet the strength requirement. This report describes a method that applies quality engineering techniques to balance strain energy.

To evaluate image quality of hard copy using the MT method, it is necessary to collect a large number of samples from the field to construct a unit space. In this study, we propose a method for constructing a uniform unit space efficiently using an orthogonal array, In this method, factors that cause image quality variations were assigned to an L27 orthogonal array, and purposely varied image patterns were created and printed to create a group of image samples. In the selection of images for constructing a unit space, to obtain obective evaluation, observer conditions, evaluation conditions, and other conditions were allocated to an L27 orthogonal array and a sensory evaluation experiment was carried out. In this experiment, since the number of extracted samples was smaller than the number of measurement items, random dummy data were inserted into the actual measurement data to avoid decreased computational preeision due to insufficient data. The experiment confirmed that the MD values obtained by this method corresponded to human evaluation.

Resin conerete is made by mixing resin and inorganic filler at a construction site and is applied to the floor with a trowel. Resin concrete has various features,including chemical resistance and mechanical strength, that derive from features of its reactive resin component, and is used for various purposes as a durable floor material. However, surface irregularities caused during troweling bend to remain, so improved leveling of the finished surface is required. We estabhshed a quantitative evaluation method by using a leveling tester to evaluate surface irregularities caused by troweling and applied the parameter design techniques of quality engineering. As a resulL, a 14 to 17 db improvement in leveling was obtained and troweling efficiency was also improved. The improvement in troweling efficiency permitted the use of only one grade of filler instead of the two grades that had been used selectively before. The fact that these results were obtained in a short period of time, which would have been difficult with conventional methods, shows that quality engineering can be efficiently applied to formulation design for materials.

In order to realize a higher efficiency in the secondary refinement of molten steel, a new method was proposed to evaluate the degassing and mixing performance of RH degassing vessels. Experiments were carried out using a water model on a reduced scale of the real vessel. The optimum injection method of inert gas, which is supplied to make the molten steel circulate in the vessel, was examined on the basis of measured velocities for water at the outlet of the down-leg. Two types of the S/N ratio were used to evaluate the effects of control factors on the velocity fluctuations. The one was a linear equation, and the other was larger-the-best characteristics. The optimum condition depended on the S/N ratio type. The confirmation tests revealed that the optimum condition obtained by the S/N ratio of the linear equation had higher reproducibility.