This research project is an application of quality engineering to the overall system design of a digital printer, as proposed by Genichi Taguchi in 2006. At the time, the companies to which the proposal was made opposed it because the number of experimental variables was too large to evaluate on a prototype printer. The idea in the present project was to carry out Taguchi's proposal by simulation and attempt to optimize the design in the conceptual design stage,before prototyping. In the development of digital printers in the past, individual optimization was the rule: first the component technologies were developed for each subunit; then they were combined and optimally adjusted, at the expenditure of much labor. In the macroscopic approach taken here, control factors were selected for the entire system so as to cover each subunit, the selected control factors were assigned to an L36 orthogonal array, and simulation was carried out. Noise factors were also assigned to an L12 orthogonal array in consideration of variability of components and assembly procedures used in the production process and environmental variations in the field. Reproducibility of the optimum conditions thus obtained was confirmed in a confirmation experiment. The economic benefit, when converted from the benefit of shortening the design study time, was estimated to be at least 218 million yen. The present project showed that it was possible to confirm overall system robustness from the conceptural design stage without a prototype printer, and to design-in quality earlier than in the past.
Control of engine cylinder bore wall temperature is an important issue:at low temperatures, there is a risk of corrosion due to sulfuric acid resulting from condensation of combustion moisture;if the temperature is too high, however, there are other risks, such as piston seizure. A study of the reduction of bore wall temperature variation was therefore made, using a simplified temperature simulation in which bore wall temperature due to combustion in the engine was the quality characteristic, with two signal factors and with time and place as indicative factors. As a result, optimal collditions were found that improved cooling efficiency while reducing temperature differences between different sites. From a detailed interaction analysis, the factors that affect bore wall temperature were clarified and an inductive analysis was carried out, showing that the cooling water flow rate by itself affected the bore wall temperature and that the effect of differences in water temperature due to driving conditions could not be changed by design strategies.