To improve the quality of veneer cutting, it is necessary to monitor the veneer cutting process. There is the possibility that during softwood veneer cutting, pressure of the bar will change because of changes of knot numbers and so on. Change of the bar pressure has an effect on the quality. In this study, a time series AR model was used to analyze the veneer thickness variations, which were obtained separately from both the "normal veneer cutting" process and the "abnormal veneer cutting" process. Relationships between the autoregressive parameter phi(1), the average residual sum of squares sigma(a)(2), and veneer cutting length are discussed. Normally, the fluctuation of the parameter phi(1) was very small (see Fig. 3) because of an increase of knots, which become more from a cutting length of 3.5 m, and the sigma(a)(2) became large from the cutting length of 4 m (see Fig. 4). In the "abnormal", from the cutting length 4 m where the pressure of the bar had been changed from the cutting length of 3.5 m, the phi(1) became gradually smaller (see Fig. 3), and because of knots and the pressure change, the sigma(a)(2) became larger than that of the "normal" (see Fig. 4).
Furthermore, based on the patterns composed of phi(1) and sigma(a)(2), an identifying function for the classification of the two veneer cutting processes was formed by means of the potential function. All the patterns were classified correctly by the identifying function.