本対談は，2020 年1 月11 日に品質工学会事務室で開催された鴨下隆志氏へのインタビューを基にしたもので，50 年以上の研究活動から，硬さの計測，官能検査，プラスチック成形などを通し，品質工学やMTシステムに至るまでの発展過程をヒアリングしたものである。インタビューから，品質工学が計測を元にしている部分を理解してもらえれば幸いである。 　なお内容が多岐にわたるため，前半と後半の2 部構成とした。

抄録全体を表示

With an eye toward future unmanned factories, a basic study was carried out concerning the metrology underlying the Internet of Things (IoT). The object was to confirm the operating procedure and required precision of an image measuring machine, and establish an evaluation method. During functional evaluation, the need for, and effectiveness of, experimental selection of noise factors also became evident. The noise factor selection experiment was carried out together with parameter design to improve the accuracy of measurements made by the image measuring machine. Based on the results, and on the JIS Z 9090 standard, a calibration method was established. This method was applied to three in-house measurement machines of the same type, and was found to yield adequate precision with all of them.

抄録全体を表示

Polymer products are generally manufactured according to formulas specified in the working standards for each brand. For polymers of high molecular weight, the catalyst quantities are minute, so variations in feedstocks and operating conditions have a large effect on molecular weight. In continuous polymerization, molecular weight variations are monitored by process analysis and fine adjustments are made to the method of polymerization. These adjustments are entrusted to the judgement of an experienced foreman, but since they are carried out on the basis of experience and the most recent adjustment results, there is considerable variation from person to person. To enable anyone to make appropriate adjustments, a simulation model was created from past operating data, and a viscosity adjustment system using orthogonal polynomials was constructed. This adjustment system simplified the adjustment process easy even for relatively inexperienced foremen and reduced variation in the molecular weight of products by 20 ％ in terms of variance with no increase in cost. Use of the Taguchi method (1) to gain the adjustment system on the basis of small amounts of data, but good results were not obtained in the present study.

抄録全体を表示