抄録
Warp sizing is an important pre-process of weaving. The sizing machine, however, has many parameters (e. g. concentration of sizing agents, viscosity of sizing agents, pressure of squeezing roller, speed of winding roller, etc.) to be set for realizing required strength, elongation, fuzz, and abrasion proof of warp. Only experts have been able to fulfill this difficult tuning.
This paper presents fuzzy modelings of the sizing machine for constructing an assisting system of the machie operation. A fuzzy neural network is used for the automatic fuzzy modeling. The obtained fuzzy models are sufficiently precise and the acquired fuzzy rules coincide well with experts' experience.
