One of the major problems of injection molding is that due to uneven resin injection, the molded product does not conform to the shape of the mold, in other words, the shape lacks transferability. The conventional solution is to trim or otherwise reshape the mold, but this is an expensive undertaking. An attempt was made to improve transferability by adjusting easily adjusted molding conditions instead of by altering the mold. Since uniform injection is a basic function of injection molding, the internal and external dimensions of the molded product were used as signal factors, and the molding conditions, the molding machine itself, and the mold conditions were selected as controllable factors, and parameter design was carried out. In the transferability analysis, the molded product was divided into parts of differing thickness. Good gain was obtained for some of the unit parts, but the gain for the molded product as a whole was low. It therefore became clear that only a small improvement in transferability could be obtained by adjustment of the molding conditions, and that mold design and product design are important factors in injection molding technology.

Recently Doppler auto trace that draws the envelope of Doppler spectrum image automatically and measures the maximum flow velocity etc. in every heartbeat cycle is used in the medical division. Although the main purposes of automation are the reductions in operation time and user variation, it is also important to obtain a reliable measurement value simultaneously. For this purpose,a technique for the stable detection of the Doppler trace waveform and the heartbeat cycle is required. Since optimization of Doppler auto trace was reported last time, we will report optimization of the automatic heartbeat cycle detection based on auto trace waveform this time. We newly developed heartbeat cycle detection processing using adaptive BPF system and optimized the parameters for a few diagnostic parts. As this result heartbeat cycle detection stability could get 45.9 db as an S/N ratio of dynamic response. This value is equivalent to the hit rate of heartbeat detection 98.7% (by the Ω value, 18.9). We could reach the enough level for use beyond initial target 85%. Compared with a conventional trial and error development, we were able to shorten the time for parameter design and improve the reliability of a product performance.

In the previous report, we proposed an ideal stress-strain diagram as a new method of evaluating stress and strain, differing in concept from conventional methods of evaluating metal materials. By evaluating this diagram with the standard S/N ratio, we concluded that fatigue strength should be tuned, using as noise factors production variability and field environment instead of fatigue degradation. In this study, we applied this new evaluation philosophy to the development of ductile cast iron. As a result, we were quickly able to develop robust ductile cast iron with higher strength and elongation than the starting material used as a base. This shows that the stress-strain evaluation presented in our previous report can be used to evaluate cast iron.

An earlier report described a stability study in which a flow rate v. pressure ratio curve obtained from simulation was used as an evaluatioll index. As a result of iterative optimization a large gain was obtained from a series of smaller gains. Study of the parameter design results at each iteration also clarified such issues as level ranges and system selection. Reproducibility was obtained by resolving these issues, leading to a compressor with a higher pressure ratio and higher efficiency. Further improvements on the basis of this index, however,turned out to be difficult, and there were also tuning problems. An iterative optimization was therefore performed using the input-output energy relationship in the basic function as the evaluation index. This led to higher gain, and the tuning problem was solved through parallel use of the results of the flow rate v. pressure ratio curve. Through these studies we established a method of applying quality engineering to the design of centrifugal compressors, which we report here,

One of the major problems of injection molding is poor transferability between the mold and the product due to uneven resin injection, requiring expensive reworking of the mold. An attempt was made to improve transferability through parameter design of the molding conditions instead of reworking the mold. As discussed in a previous report, accurate measurement of the a test piece output from the mold is essential in evaluation of transferability; in the present study, thermal degradation was introduced as a noise factor, the product was divided into parts, and transferability was evaluated of each part. A modest improvement in gain was obtained as a result.

An automated blood analyzer pours blood serum extracted from an examinee's blood and a reagent into a reaction vessel, mixes them, and measures the quantity of the product generated by chemical reaction in the liquid with an absorptiometer. As the temperature of the reaction liquid must be kept at human body temperature (about 37℃) due to the nature of the measurement, a thermostatic chamber is provided inside. In the thermostatic chamber, the temperature of the reaction liquid must be raised to 37℃ reliably and quickly. In this study, the design of the thermostatic chamber was optimized by means of simulation to ensure measurement accuracy and to reduce measuring time. An evaluation by the standard S/N ratio was performed with shapes and dimensions of the parts of the thermostatic chamber treated as control factors, time elapsed after the reagent was poured into the reaction vessel as the signal factor, and temperature of the reaction liquid as the output. Results showed that with the optimal design, time required for the reaction liquid temperature to reach 37℃ was reduced to about one third and variation of the reaction liquid was reduced to one seventeenth as compared with a set of comparative conditions.